Documents - MX - QT Study Pack 2006 Wip PDF

ii
Acknowledgement
Acknowledgement
This study kits success is attributed first and foremost to Saint Jose Maria Escriva
the founder of opus dei and the inspiration to the formation of Strathmore University
and its values.
Much appreciation is also extended to Mr. Randhir Ahluwalia of Strathmore
University for his contribution, Mr. Paul Maloba for his contributon and
compilation of the book.
Acknowledgement is given to the books that were used in research for compilation of
this book; Applied Mathematics for Business Economics and Social Sciences, and
Quantitative Techniques 6th edition.
We gratefully acknowledge permission to quote from the past examination

papers of the following bodies: Kenya Accountants and Secretaries National
Examination Board (KASNEB); Chartered Institute of Management
Accountants (CIMA); Association of Chartered Certified Accountants
(ACCA).
STRATHMORE UNIVERSITY STUDY PACK
Instructions for Students
iii
Instructions for Students
Quantitative techniques paper in section 4 of the CPA examination is an eight

question paper devided intotwo sections.
Section I has 5 questions and only 3 questions are to be attempted, section II has 3
questions and only 2 are to be attempted. All questions have a total of 20marks.
Section I contains mainly questions on calculus, probability, matrices, test of
hypothesis, regression analysis, sampling and estimation and descriptive statistics.
Section II contains question from operations research and decision theory.
from past papers statitics the topicss frequently feature in exam are network analysis,
matrices, test of hypothesis and calculus. Other topics that are also favorites of the
examiner are regression analysis, time series, decision making techniques, linear
programming, probability and sets theory. It is very important to clearly understand
these topics through and through, emphasize more on working out examples and
assignments from this book.
The course demands rigorous practice of of questions to internalize concepts, thats
the only way to go about it. Work out reinforcing questions at the end of each study
session and compare them with answers given in study session 9. comprehensive
assignments are to be handed in to DLC department for marking.
QUANTITATIVE TECHNIQUES
iv
Contents
Contents
Acknowledgement.............................................................................................................. ii
Instructions for Students ................................................................................................... iii
Contents............................................................................................................................. iv
Course Description............................................................................................................. v
Index.................................................................................................................................. vi
LESSON ONE....................................................................................................................1
1. Linear Algebra and Matrices.....................................................................................1

LESSON TWO..................................................................................................................45
2. Sets Theory and Calculus........................................................................................45

LESSON THREE.............................................................................................................81
3. Descriptive Statistics and Index Numbers ..............................................................81

LESSON FOUR.............................................................................................................. 118
4. Measures of Relationships and Forecasting..........................................................118

LESSON FIVE................................................................................................................ 157
5. Probability.............................................................................................................157
LESSON SIX................................................................................................................... 185
6. Sampling and Estimation ......................................................................................185

LESSON SEVEN............................................................................................................226
7. Decision Theory....................................................................................................226
LESSON EIGHT............................................................................................................248
8. Operation Research ...............................................................................................248

LESSON NINE ..............................................................................................................303
9. Revision Aid .........................................................................................................303
Course Description
Course Description
Quantitative techniques is a mathematical paper and fundamental to many proffessionalcourses

it forms the basis of finance and acconting. The course dwells more on practical applicationtype
questions and sometimes even requires one to make inference and decisions based on accurate
analysis of information .
This paper also formas a basis for section 5 & 6 papers, management accounting and Financial
management.
The book has been tailored to make you comfortable in a mathematical environment so as to
excel in the accounting profession.
vi
Index
Index
Lesson One
LESSON ONE
Linear Algebra and Matrices
Contents
- Functions and graphs
- Linear equations, higher order equations, inequalities and simultaneous equations
- Matrix algebra
- Application of matrix algebra to input-output analysis and elementary Markovian
process.
1.1
Functions and graphs
A function is a mathematical relationship in which the value of a single dependent variable are
determined from the values of one or more independent variables. The following is an example
of a function in which y is said to be a function of x.
y = a + bx
In the above example, both x and y are variables this is because they may assume different values
throughout the analysis of the function. On the other hand, a and b are referred to as constants
because they assume fixed values.
The variable y is a dependant variable in the sense that its values are generated from an
independent variable x.
The collection of all the values of the independent variable for which the function is defined is
referred to as the domain of the function corresponding to this we have the range of the
function, which is the collection of all the values of the dependent variable defined by the
function
The fact that it is a function of x can also be denoted by the following general form
y = f(x)
Functions of a single independent variable may either be linear or non linear.
Linear functions can be represented by:
y = a + bx
Whereas non linear functions can be represented by functions such as:
i.
ii.
iii.
iv.
v.
vi.
y = 0 + 13 x + 2x3
y2 = 3x + 18
y = 2x2 + 5x + 7
ax2 + bx + cy + d = 0
xy = k
y = ax
Where , a, b, c, d, k = constants
Graph of a function
A graph is a visual method of illustrating the behaviour of a particular function. It is easy to see
from a graph how as x changes, the value of the f(x) is changing.
The graph is thus much easier to understand and interpret than a table of values. For example
by looking at a graph we can tell whether f(x) is increasing or decreasing as x increases or
decreases.
We can also tell whether the rate of change is slow or fast. Maximum and minimum values of
the function can be seen at a glance. For particular values of x, it is easy to read the values of f(x)
and vice versa i.e. graphs can be used for estimation purposes
Different functions create different shaped graphs and it is useful knowing the shapes of some
of the most commonly encountered functions. Various types of equations such as linear,
quadratic, trigonometric, exponential equations can be solved using graphical methods.
Equations
An equation is an expression with an equal sign (=)
Equations are classified into two main groups linear equations and non linear equations.
Examples of linear equations are
x + 13 = 15
7x + 6 = 0
Non linear equations in the variable x are equations in which x appears in the second or higher
degrees. They include quadratic and cubic equations amongst others. For example
Lesson One
5x2 + 3x + 7 = 0 (quadratic equation)

2x3 + 4x2 + 3x + 8 = 0 (cubic equation)
The solution of equations or the values of the variables for which the equations hold is called the
roots of the equation or the solution set.
Solution of Linear Equation

Supposing M, N, and P are expressions that may or may not involve variables, then the
following constitute some rules which will be useful in the solution of linear equations
Rule 1: Additional rule
If M = N then M + P = N + P
Rule 2: Subtraction rule
If M = N, Then M P = N P
Rule 3: multiplication rule
If M = N and P O then M x P = N x P
Rule 4: Division rule
If P x M = N and P O
And N/P = Q Q being a raterial number then
M = N/P
Example
i.
Solve 3x + 4 = - 8
ii.
Solve
i.
3x + 4 = 8
y
=-4
3
Solutions
3x + 4 4 = 8 4 (by subtraction rule)
3x = 12
(simplifying)
3x
12
=
3
3
(by division rule)
x=4
ii.
y
3
(simplifying)
= 4 3
y = 12
(simplifying)
Solution of quadratic equations

Suppose that we have an equation given as follows
ax2 + bx + c = 0
Where a, b and c are constants, and a 0. such an equation is referred to as the general quadratic
equation in x. if b = 0, then we have
ax2 + c = 0
Which is a pure quadratic equation

There are 3 general methods for solving quadratic equations; solution by factorization, solution
by completing the square and solution by the quadratic formula.
Solution by Factorization
The following are the general steps commonly used in solving quadratic equations by
factorization
(i) Set the given quadratic equation to zero
(ii) Transform it into the product of two linear factors
(iii) Set each of the two linear factors equal to zero
(iv) Find the roots of the resulting two linear equations
Example
Solve the following equation by factorization
i.
6x2 = 18x
ii.
15x2 + 16x = 15
Solutions
i.
6x2 = 18x
6x2 18x = 0..........................................................(step 1)
6x(x 3) = 0 ..........................................................(step 2)
6x = 0 ......................................................................(step 3)
and x 3 = 0
x = 0 or x = 3 ..................................................(by step 4)
ii.
15x2 + 16x = 15
15x2 + 16x 15 = 0 .............................................. (step 1)
(5x 3) (3x +5) = 0 ..............................................(step 2)
(5x 3) = 0} Step 3
{3x + 5 = 0}
x=-5
or + 3 ............................................(step 4)
Solution by Completing the Square

The process of completing the square involves the construction of a perfect square from the
members of the equation which contains the variable of the equation.
Consider the equation 9x2 bx = 0
The method of completing the square will involve the following steps
Make the coefficient of x2 unity
i.
ii.
Add the square of the coefficient of x to both sides of the equal sign. The
left hand side is now a perfect square
Lesson One
iii.
Factorize the perfect square on the left hand side.
iv.
Find the square root of both sides
v.
Solve for x
Example
Solve by completing the square.
i.
3x2 = 9x
ii.
2x2 + 3x + 1 = 0
Solutions
i.
3x2 = 9x
or
(3x2 - 9x = 0)
x2 - 3x = 0 ...................................................................... (Step 1)
2
3 3
x 2 3x + =
2 2 .....................................(Step 2)
2
3 9
x =
2
4 ..............................................................(Step 3)
9
4 ...............................................................(Step 4)
x3 =
x=
3 3
2 2
3+3 3 3
or =
2
2 2
(= 3 or 0)
ii.
2x2 + 3x + 1 = 0
x2 +
or
(2x2 + 3x = -1)
3x
1
= - ..... (Step 1)
2
2
2
3x 3 3 1
+ =
X +
(Step 2)
2
2 4 4
2
3
1
.. (Step 3)
x + =
4
16
x+
3
4
x = 34 =
34 +
1
4
1
16
1
4
or - 34 -
1
4
x = 12 or x = 1
Solution by Quadratic Formula
Consider the general quadratic equation
ax 2 + bx + c = 0 where a 0
The roots of the equation are obtained by the following formula:
x=
b b 2 4ac
2a
Example
Solve for x by formula
5x2 + 2x 3 = 0
Solution
a = 5, b = 2, c = - 3
x=
b b 2 4ac
2a
2 2 2 4(5)(3)
x=
2(5)
3
x = or 1
5
Inequalities
An inequality or inequation is an expression involving an inequality sign (i.e. >, <, , , i.e.
greater than, less than, less or equal to, greater or equal to) The following are some examples of
inequations in variable x.
3x + 3 > 5
x2 2x 12 < 0
The first is an example of linear inequation and the second is an example of a quadratic in
equation.
Lesson One
Solutions of inequations
The solutions sets of inequations frequently contain many elements. In a number of cases they
contain infinite elements.
Example
Solve and graph the following inequation
x 2 > 2 ; x w (where x is a subset of w)
Solution
x 2 > 2 so
x2+2>2+2
Thus, x>4
The solution set is infinite, being all the elements in w greater than 4
10
11
Example
Solve and graph
3x 7 < - 13;
Solution
3x - 7 < -13
3x - 7 + 7 < -13 + 7
3x < -6
3x
-6
<
3
3
x < -2
.. R Line
-4
-3
-2
-1
Rules for solving linear inequations

Suppose M, M1, N, N1 and P are expressions that may or may not involve variables, then the
corresponding rules for solving inequations will be:
Rule 1: Addition rule
If M > N and M1> N1
Then
M + P > N + P and
M1 + P >N1+ P
Rule 2: Subtraction Rule

If M < N and M1 N1
Then
M P < N P and
M1 P N1 P
Rule 3: Multiplication rule

If M N and M1 > N1 and P 0
Then MP NP; M1P > N1P
M(-P) N(-P) and M1(-P) < N1(-P)
Rule 4: Division
If M > N and M1< N1 and P 0
Then
M/P > N/P: M1/P < N1/P
M/(-P) < N/(-P) : and M1/(-P) > N1/(-P)

Rule 5: Inversion Rule
If M/P N/Q where P, Q 0
M1/P > N1/Q
Then P/M Q/N and P/M1 < Q/N1
Note: The rules for solving equations are the same as those for solving equations with one
exception; when both sides of an equation is multiplied or divided by a negative number, the
inequality symbol must be reversed (see rule 3 & Rule 4 above).
Example
Solve and graph the following:
i.
7 2x > - 11 ;
ii.
5x + 4 2x 10 ;
iii.
3 2x + 1 < 7 ;
Solutions
i.
7 - 2x > -11
-2x > -18 (subtraction rule)
-2x
-18
<
(bydivision rule)
-2
-2
x<9
Lesson One
line
Q
-3 -2 -1 0
9 10 11
-5x + 4 2x - 10
ii.
-7x + 4 -10
(by subtraction rule)
-7x -14
(by subtraction rule)
x2
(b y division rule)
Q
line -4
-3
-2
-1
-3 2x + 1 < 7
iii.
-4 2x < 6
(by substraction rule)
-2 x < 3
( by division rule)
Q
line -4
-3
-2
-1
Linear inequation in two variables: relations

An expression of the form
y 2x 1
is technically called a relation. It corresponds to a function, but different from it in that,
corresponding to each value of the independent variable x, there is more than one value of the
dependent variable y
Relations can be successfully presented graphically and are of major importance in linear
programming.
1.2
Linear simultaneous equations:
Two or more equations will form a system of linear simultaneous equations if such equations be
linear in the same two or more variables.
10
For instance, the following systems of the two equations is simultaneous in the two variables x
and y.
2x + 6y = 23
4x + 7y = 10
The solution of a system of linear simultaneous equations is a set of values of the variables
which simultaneously satisfy all the equations of the system.
Solution techniques
a) The graphical technique
The graphical technique of solving a system of linear equations consists of drawing the graphs of
the equations of the system on the same rectangular coordinate system. The coordinates of the
point of intersection of the equations of the system would then be the solution.
Example
10
9
.
8
.
7
.
6
.
(2,4)
4
.
3
.
x + 2y = 10
2x + y = 8
-1
10 11 12
13
The above figure illustrates:

Solution by graphical method of two equations
2x + y = 8
x + 2y = 10
The system has a unique solution (2, 4) represented by the point of intersection of the two
equations.
b) The elimination technique

This method requires that each variable be eliminated in turn by making the absolute value of its
coefficients equal in the equations of the system and then adding or subtracting the equations.
Making the absolute values of the coefficients equal necessitates the multiplication of each
equation by an appropriate numerical factor.
Consider the system of two equations (i) and (ii) below
Lesson One
11
2x 3y = 8 .. ........................................................(i).
3x + 4y = -5 .........................................................(ii).
Step 1
Multiply (i) by 3
6x 9y = 24 .......................................................................(iii).
Multiply (ii) By 2
6x + 8y = - 10 ...................................................................(iv).
Subtract iii from iv.
17y = -34 .............................................................................(v).
y = -2
Step 2
Multiply (i) by 4
8x 12y = 32 . ...................................................................(vi)
Multiply (ii) by 3
9x + 12y = -15 .. .....................................................................(vii)
Add vi to vii
17x = 17 .. ............................................................................(viii)
x=1
Thus x = 1, y = -2
i.e. {1,-2}
c) The substitution technique

To illustrate this technique, consider the system of two equations (i). and (ii) reproduced below
....... 2x 3y = 8 ..
(i).
....... 3x + 4y = -5
(ii).
The solution of this system can be obtained by

a) Solving one of the equations for one variable in terms of the other variable;
b) Substituting this value into the other equation(s) thereby obtaining an equation with one
unknown only
c) Solving this equation for its single variable finally
d) Substituting this value into any one of the two original equations so as to obtain the
value of the second variable
Step 1
Solve equation (i) for variable x in terms of y
2x 3y = 8
x= 4 + 3/2 y
(iii)
Step 2
12
Substitute this value of x into equation (ii). And obtain an equation in y only
3x + 4y = -5
3 (4 + 3/2 y) + 4y = -5
8 y = - 17 .
(iv)
Step 3
Solve the equation (iv). For y
8y = -17
y = -2
Step 4
Substitute this value of y into equation (i) or (iii) and obtain the value of x
2x 3y = 8
2x 3(-2) = 8
x=1
Example
Solve the following by substitution method
2x + y = 8
3x 2y = -2
Solution
Solve the first equation for y
y = 8 2x
Substitute this value of y into the second equation and solve for x
3x 2y = -2
3x 2 (8-2x) = -2
x=2
Substitute this value of x into either the first or the second original equation and solve for y
2x + y = 8
(2) (2) + y = 8
y=4
1.3
MATRICES
A matrix is a rectangular array of items or numbers. These items or numbers are arranged in
rows and columns to represent some information.
The position of an element in one matrix is very important as well be seen later; therefore an
element is located by the number of the row and column which it occupies.
The size of a matrix is defined by the number of its rows (m) and column (n).
Lesson One
13
a b
For example =
and B =
c d
a b c
d e f
g h i
are (2 x 2) and (3 x 3) matrices since A has 2 rows and 2 columns and B has 3 rows and 3
columns.
A matrix A with three rows and four columns is given by one of:
a11 a12
A= a 21 a 22
a
31 a 32
a13
a 23
a 33
a14
a 24
a 34
or
A = ( a ij ) i = 1, 2, 3
j = 1, 2, 3, 4
where i represents the row number whereas j represents the column number
Properties of matrices
Equal Matrices
Two matrices A and B are said to be equal, that is
A=B
or
(a ) = (b )
ij
ij
If and only if they are identical if they both have the same number of rows and columns and the
elements in the corresponding locations in the two matrices should be the same, that is, aij = bij
for all i. And j.
Example
3 4 0 3 4 0
The following matrices are equal 2 2 3 = 2 2 3

5 1 1 5 1 1
Column Matrix or column vector

A column matrix, also referred to as column vector is a matrix consisting of a single column.
For example x =
x1

x2
.

.
.

xn
14
Row matrix or row vector

It is a matrix with a single row
For example y = ( y1 , y 2 , y3 .........y n )
Transpose of a Matrix
The transpose of an mxn matrix A is the nxm matrix AT obtained by interchanging the rows and
columns of A.
A
=
aij
The transpose of A i.e. AT is given by
AT =
aij
aji
mxn
nxm
Example
Find the transposes of the following matrices
1 5 7
A= 2 1 4
0 9 3
B= ( b1 , b 2 , b3 , b 4 )
x1

C= x 2
x
3
Solution
T
i. A T
1 5 7
1 2 0
= 2 1 4 = 5 1 9
0 9 3
7 4 3
ii. BT = ( b1 , b 2 , b3 , b 4 )
b1

b
= 2
b3

b4
Lesson One
15
T
x1

T
iii. C = x 2 =
x
3
( x1 x 2 x 3 )
Square Matrix
A matrix A is said to be square when it has the same number of rows as columns
e.g.
2 5
A=
3 7
is a square matrix of order 2
B = n n is a square matrix of the order n
Diagonal matrices
It is a square matrix with zeros everywhere in the matrix except on the principal diagonal
e.g.
3 0 0
A = 0 1 0 , B =
0 0 7
9 0 0
0 0 0
0 0 0
An identity of unity matrix

It is a diagonal matrix in which each of the diagonal elements is a positive one (1)
e.g.
1 0
I2 =
0 1
1 0 0
and I 3 = 0 1 0
0 0 1
2 2 unit matrix
3 3 unit matrix
A null or zero matrix

A null or zero matrix is a matrix whose elements are all equal to zero.
Sub matrix
The sub matrix of the matrix A is another matrix obtained from A by deleting selected row(s)
and/or column(s) of the matrix A.
16
7 9 8
e.g, if A = 2 3 6
1 5 0
2 3 6
7 9
then A1 =
and A 2 =
1 5 0
1 5
are both sub matrices of A
OPERATION ON MATRICES
Matrix addition and subtraction
We can add any number of matrices (or subtract one matrix from another) if they have the same
sizes. Addition is carried out by adding together corresponding elements in the matrices.
Similarly subtraction is carried out by subtracting the corresponding elements of two matrices as
shown in the following example
Example: Given A and B, calculate A + B and A B
A=
6 1 10 5
4
2 5
3
9 13 6 0
12 4 7 3
B = 0 4 10 4
7 3 7
9
3
3 8
6 1 10 5 12 4 7 3 18
A+B= 3
4
2 5 + 0 4 10 4 = 3
0 12 9
9 13 6 0 7 3 7 9 2 16 1 9
2
6 1 10 5 12 4 7 3 6 5 17
A-B= 3
4
2 5 - 0 4 10 4 = 3
8
8 1
9 13 6 0 7 3 7 9 16 10 13 9
If it is assumed that A, B, C are of the same order, the following properties are fulfilled:
a) Commutative law:
b) Associative law:
A+B =B+A
(A + B) + C
= A + (B + C) = A + B + C
Multiplying a matrix by a number

In this case each element of the matrix is multiplied by that number
Lesson One
17
Example
If A =
6 1 10 5
3 4 2 5
9 13 6 0
60 10 100 50
then (10)A = 30
40 20 50
90 130 60 0
Matrix Multiplication
a) Multiplication of two vectors
Let row vector A represent the selling price in shillings of one unit of commodity P, Q, R
respectively and let column vector B represent the number of units of commodities P, Q, R sold
respectively. Then the vector product A B will be equal to the total sales value
i. e.
A B =Total sales value
100
Let A = ( 4 5 6 ) and B = 200

300
100
then ( 4 5 6 ) 200 = 400 + 1,000 + 1,800 = Shs 3,200

300
Rules of multiplication
i.
ii.
iii.
The row vector must have the same number of elements as the column vector
The first vector is a row vector and the second is a column vector
The corresponding elements in each vector are multiplied together and the results
obtained are added. This addition is always a single number
Going back to the example given before
100
A B = ( 4 5 6 ) 200 = 4 100 + 5 200 + 6 300=Shs3,200, a single number

300
18
b) Multiplication of two matrices

Rules
Multiplication is only possible if the first matrix has the same number of columns as
the second matrix has rows. That is if A is the order ab, then B has to be of the
order bc. If the AB = D, then D must be of the order ac.
ii.
The general method of multiplication is that the elements in row m of the first
matrix are multiplied by the corresponding elements in columns n of the second
matrix and the products obtained are then added giving a single number.
We can express this rule as follows
i.
a11 a12
a 21 a 22
Let A =
b12
b
and b = 11
b 21 b 22
d11 d12
d 21 d 22
Then A B = D =
A = 2 x 2 matrix
b13
b 23
d13
d 23
B = 2 x 3 matrix D = 2 x 3 matrix
Where
d11 = a11 b11 + a12 b21
d12 = a11 b12 + a12 b22
Example I
6 1
3 0 2
6 3 + 1 4 6 0 + 1 5 6 2 + 1 8
2 3
4 5 8
2 3 + 3 4 2 0 + 3 5 2 2 + 3 8
22 5 20
=
18 15 28
Example II
Matrix X gives the details of component parts used in the make up of two products P1 and P2
matrix Y gives details of products made on each day of the week as follows:
Matrix Y
Products
Matrix X
P1
Parts
A B C
P 3 4 2
Products 1
P2 2 5 3
Tues 2
Wed 3
Thur 2
Fri 1
Mon
P2
2
3
2
2
Lesson One
19
Use matrix multiplication to find the number of component parts used on each day of the week.
Solution:
After careful consideration, it will be easy to decide that the correct order of multiplication is
YXX (Note the order of multiplication). This multiplication is compatible and also it gives the
desired answer.
2
Y X = 3
2
1
3
2
2
1
3 4 2
2 5 3
5 x 2 matrix
A B
Mon
Tues
Wed
Thur
Fri
12
13
10
5
2 x 3 matrix
13+22
23+32
33+22
23+22
13+12
14+25 12+25
24+35 22+33
34+25 32+23
24+25 22+23
14+15 12+13
5 x 3 matrix
14
23 13
22 12
18 10
9 5
Interpretation
On Monday, number of component parts A used is 7, B is 14 and C is 8. in the same way, the
number of component parts used for other days can be interpreted.
The determinant of a square matrix

The determinant of a square matrix det (A) or A is a number associated to that matrix. If the
determinant of a matrix is equal to zero, the matrix is called singular matrix otherwise it is called
non-singular matrix. The determinant of a non square matrix is not defined.
Determination of a 2 x 2 matrix
(A)
a b
=
c d
= ad - cb
ii. Determinant of a 3 x 3 matrix
a b c
A = d e f
g h i
e f
= a
h i
d f
b
g i
a ( ei - fh ) - b ( di - gf ) +c ( dh - eg )
simplify
d e
+c
g h
20
iii.
A=
Determinant of a 4 x 4 matrix
a b c d
e f g h
i j k l
m n o p
f g h
A = a j k l
n o p
e g h
b i k l
n o p
e f h
+c i j l
m o p
e f g
d i j k
m n o
Simplify 3 x 3 determinants as in ii and then evaluate the 4 x 4 determinants.
Inverse of a matrix
If for an n ( n square matrix A, there is another n ( n square matrix B

such that there product is the identity of the order n X n, In, that is A X B
= B X A = I, then B is said to be inverse of A. Inverse if generally
written as A-1
Hence AA-1 = I
Note: Only non singular matrices have an inverse and therefore the
inverse of a singular matrix is non defined.
General method for finding inverse of a matrix
In order to introduce the rule to calculate the determinant as well as the

inverse of a matrix, we should introduce the concept of minor and
cofactor.
The minor of an element
Given a matrix A = (aij), the minor of an element aij in row i and column
j (call it mij), is the value of the determinant formed by deleting row i
and column j in matrix A.
Example
Let A = J EMBED Equation.DSMT4 JJJ

The minors are,
Lesson One
m11 =
m12 =
6 1
3 0
21
= 60 31 = 3
5 1
2 0
= 50 12 = 2
Similarly
m13 =
5 6
2 3
=15 12 = 3
m31 =
m 21 =
2 3
m 22 =
3 0
=0 9 = 9
2 3
6 1
= 2 -18 = -16
m 32 =
4 3
m 23 =
2 0
= 06 =6
4 3
5 1
m33 =
= 4 -15 = -11
4 2
2 3
= 12 4 = 8
4 2
5 6
= 24 -10 = 14
The cofactor of an element

The cofactor of any element aij (known as cij) is the signed minor associated with that element.
The sign is not changed if (i+j) is even and it is changed if (i+j)is odd. Thus the sign alternated
whether vertically or horizontally, beginning with a plus in the upper left hand corner.
+ +
i.e. 3 x 3 signed matrix will have signs +

+ +
Hence the cofactor of element a11 is m11 = -3, cofactor of a12 is m12 = +2 the cofactor of
element a13 is +m13 = 3 and so on.
Matrix of cofactors of A =
in general for a matrix M
3 2
9 6
16 11
=
d
g
8
14
b c
e f
h i
Cofactor of a is written as A, cofactor of b is written as B and so on.

Hence matrix of cofactors of M is written as
22
A B C
= D E F
G H I
The determinant of a nn matrix

The determinant of a nn matrix can be calculated by adding the products of the element in any
row (or column) multiplied by their cofactors. If we use the symbol for determinant.
Then = aA + bB + cC
or
= dD + eE + fF e.t.c
Note: Usually for calculation purposes we take = aA + bB + cC

Hence in the example under discussion
= (4 3) + (2 2) + (3 3) = 1
The ad joint of a matrix
A B C
The ad joint of matrix D E F is written as

G H I
A D G
B E H i.e. change rows into columns and columns into rows (transpose)
C F I
a b c
The inverse of the matrix

d e f
g h i
is written as
i.e. A
-1
1
(adjoint of the matrix of cofactors)
determinant
A D G
1
=
B E H

C F I
Where = aA + bB + cC
Lesson One
23
4 2 3
Hence inverse of 5 6 1
2 3 0
is found as follows
= (4 3) + (2
2) + (3 ( 3) = 1
A = -3
B=2
C=3
D=9
E = -6
F = -8
G = -16
H = 11
I = 14
J EMBED Equation.DSMT4 JJJ

(note: Check if A ( A-1 = A-1 A = 1)
Solution of simultaneous equations
In order to determine the solutions of simultaneous equations, we may use either of the
following 2 methods
The cofactor method
i.
Cramers rule
ii.
The cofactor method

This method requires that we obtain
a) The minors and cofactors
b) The adjoint of the matrix
c) The inverse of the matrix
d) Multiply the original by the inverse on both sides of the matrix equation
Example
Solve the following
a) 4x1 + x2 5x3 = 8
-2x1 + 3x2 + x3 = 12
3x1 x2 + 4x 3 = 5
b) 4x1 + 3x3 + 5x3 = 27
x1 + 6x2 + 2x3 = 19
3x1 + x2 + 3x3 = 15
c) 4x1 + 2x2 + 6x3 = 28
3x1 + x2 + 2x3 = 20
24

10x1 + 5x2 + 15x3 = 70
d) 2x1 + 4x2 3x3 = 12
3x1 5x2 + 2x3 = 13
-x1 + 3x2 + 2x3 = 17
Solution
a) From a, we have
4 1 -5 x1
8

-2 3 1 x 2 = 12
3 -1 4 x
5
3

A
We need to determine the minors and the cofactors for the above matrix
Definition
A minor is a determinant of a sub matrix obtained when other elements are detected as shown
below.
A cofactor is the product of (-1) i + j and a minor where
i = Ith row i = 1, 2, 3 .
j = Jth row j = 1, 2, 3 .
Cofactor of 4 (a11) = (-1) 1+1
Cofactor of -2 (a21) = (-1) 2+1

Cofactor of 3 (a31) = (-1) 3+1
Cofactor of 1 (a12) = (-1) 1+2
Cofactor of 3 (a22) = (-1) 2+2
Cofactor of -1 (a23) = (-1) 2+3

Cofactor of -5 (a13) = (-1) 1+3
1 4
1
1 5
3 1
2 1
3
4 5
3
2 1
2 3
3
= 13
= 1
= 16
= 11
= 31
= 6
= 7
Lesson One
25
Cofactor of +1 (a23) = (-1) 2+3
3 1
4
Cofactor of 4 (a33) = (-1) 3+3
2 3
= 7
= 14
The matrix of C of cofactors is
13 11 7
1 31 7
16 6 14
13 1 16
CT = 11 31 6
7 7 14
= Adjoint of the original matrix of coefficients
The original matrix of coefficients
4 1 5
= 2 3
1
3 1 4
Therefore
(48 + 3 10) (-45 4 8)

=
41 + 57
98
The inverse of the matrix of coefficients, see (*) will be
13 1 16
1
=
11 31 6
98
-7 7 14
by multiplying the inverse on both sides of * we have,
26
13 1 16
1
11 31 6
98
-7 7 14
4 1 5
2 3 1
3 1 4
13 1 16
1
11 31 6
=
98
-7 7 14
8

12

5
98 0 0
1
=
0 98 0
98
0 0 98
1 0 0
= 0 1 0
0 0 1
x1

x2

x3
x1

x2
x
3
196
1
=
490
98
98
x1
2

x 2 = 5

x3
1
x1
2

= x2 = 5
x

3
1
X1 = 2, X2 = 5, X3 = 1
c) 4x1 + 2x2 + 6x3 = 28
3x1 + x2 + 2x3 = 20
10x1 + 5x2 + 15x3 = 70
4 2 6
3 1 2
10 5 15
4 2 6
3 1 2
10 5 15
x1
28

x 2 = 20
x

3
70
4
=
3 1 2 3 1
10 5 15 10 5
= (60 + 40 + 90) (60 + 40 + 90)

=0
Lesson One
27
Hence the solutions of x1, x2, and x3 do no exist. The equations are independent
Now work out part (b) on your own.
Cramers Rule in Solving Simultaneous Equations

Consider the following system of two linear simultaneous equations in two variables.
a11 x1 + a12 x2 = b1 (i)
a21 x1 + a22 x2 = b2 (ii)
after solving the equations you obtain
x1 =
b1a 22 b 2 a12
=
a11a 22 a12 a 21
b1 a12
b 2 a 22
a11 a12
a 21 a 22
and
x2 =
a11 b1
a 21 b 2
a11 a12
a21 a22
a11b 2 - a 21b1
=
a11a 22 - a12 a 21
Solutions of x1 and x2 obtained this way are said to have been derived using Cramers rule,
practice this method over and over to internalize it. It is advisable for exam situation since it is
shorter.
Example
Solve the following systems of linear simultaneous equations by Cramers rule:
2x1 5x2 = 7
i)
x1 + 6x2 = 9
ii)
x1 + 2x2 + 4x3 = 4
2x1 + x3 = 3
3x2 + x3 = 2
Solutions
i.
2x1 5x2 = 7
x1 + 6x2 = 9
can be expressed in matrix form as
2 5 x 1
7
1 6 x = 9
2

A
28
and applying cramers rule
7 -5
x1 =
9 6
=
2 -5
1
87
2
= 5
17
17
2 7
x2 =
1 9
=
2 -5
1
11
17
(ii) can be expressed in matrix form as
1 2 4
2 0 1
0 3 1
x1

x2

x3
4

3

2
and by cramers rule
4 2 4
3 0 1
x1 =
2 3 1
1 2 4
22
17
7
17
2 0 1
0 3 1
1 2 4
2 0 3
x3 =
0 3 2
1 2 4
2 0 1
0 3 1
1 4 4
2 3 1
x2 =
0 2 1
1 2 4
9
17
2 0 1
0 3 1
Lesson One
29
Solving simultaneous Equations using matrix algebra

i.
Solve the equations

2x + 3y = 13
3x + 2y = 12
in matrix format these equations can be written as
2 3 x
13
=
3 2 y
12
pre multiply both sides by the inverse of the matrix
2 3
3 2
= 5
and inverse of the matrix is
2
5
1 2 3

=
5 3 2
3
3
5
2

5
Pre multiplication by inverse gives
2
5
3
5
2

5
2
5
=
3
2 3
3 2
Therefore x = 2
3
2
5 13
=
2
12
3

5
y=3
Solve the equations
ii.
4x + 2y + 3z = 4
5x + 6y + 1z = 2
2x + 3y = -1
Solution:
Writing these equations in matrix format, we get
A
4 2 3
5 6 1
2 3 0
BX =
x
4

y = 2
z
-1

Pre-multiply both sides by the inverse
30
3 9 16
= 2 6 11
3 8 14
the inverse of A as found before is A-1
3 9 16
2 6 11
3 8 14
hence x = 22
4 3 2
5 6 1
2 3 0
x
3 9 16

y = 2 6 11
z
3 8 14

4
22

2 = -15
-1
-18

y = -15 z = -18
(Note: under examination conditions it may be advisable to check the solution by substituting
the value of x, y, z into any of the three original equations)
MARKONIKOV CHAIN
Probability Transition Matrices (Brand switching)

These are matrices in which the individual elements are in the form of probabilities.
The probabilities represent the probability of one event following another event i.e. the
probability of transition from one event to the next
The probabilities of the various changes applied to the initial state by matrix multiplication, give
a forecast of the succeeding state.
Normally a transition matrix is defined with its columns adding upto one and state vectors as
column vectors.
In this case the succeeding state is found by pre-multiplying the transition matrix by the
proceeding state (column) vector
If the transition matrices are given with their rows adding up to one, then the succeeding state is
found by post multiplying the transition matrix by the preceding state (row) vector.
Example 1
The probability transition matrix of the switching probabilities, consider that two brands G and
X share the market in the ratio of 60% to 40% respectively of customers. If in every week 70%
of Gs customers retain the brand but 30% switch to product x where as 80% of Xs customers
retain brand but 20% percent switch to brand G. Analyse the exchange in share market per
week.
G
X
G 0.7
0.2
State the system next week
X 0.3
0.8
Column vector for initial market share
G 60
X 40
0.7 0.2 60
50
=
0.3 0.8 40
50
Share next week
Lesson One
31
0.7 0.2 50
45
=
0.3 0.8 50
55
Share week after
and so on
This process can continue till equilibrium is reached.
Let the market share be
0.7 0.2
0.3 0.8
G
G
=
X
X
0.7G + 0.2 X = G
0.2X = 0.3G
i.e.
G
X
or
or
0.3
=
0.2
0.3G + 0.8X = X
0.3X = 0.2X
3
2
Hence Gs share is 60% and Xs share is 40%
Example 2
A marketing division toothpaste manufacturing company has worked out the following
transition probability matrices concerning the behaviors of customers before and after an
advertising campaign.
Transition probability matrix
(before advertising campaign)
TO
FROM
Our brand (State I)
Another Brand (sate II)
Our Brand
(State I)
0.8
0.4
Another Brand
(Sate II)
0.2
0.6
Transition probability matrix

(After advertisement)
TO
FROM
Our brand (State I)
Another Brand (sate II)
Our Brand
(State I)
0.9
0.5
Another Brand
(Sate II)
0.1
0.5
If the advertising campaign costs Shs 20,000 per year, would it be worthwhile for the company
to undertake the campaign?
32
You may suppose there are 60,000 buyers of toothpaste in the market and for each customer
average annual profit of the company is Shs 2.50
Solution
Let P1 be the fraction share of our brand and P2 be the fraction share of another brand
Before Advertising
( P1
After Advertising
0.8 0.2
P2 )
= ( P1
0.4 0.6
( 0.8P1 + 0.4 P2
P2 )
0.2 P1 + 0.6 P2 ) = ( P1
( P1
P2 )
0.8 P1 + 0.4 P2 = P1 and 0.2 P1 + 0.6 P = P2
0.4 P2 = 0.2 P1 and 0.2 P1 = 0.4 P2
0.9 0.1
P2 )
= ( P1
0.5 0.5
( 0.9 P1 + 0.5P2
P2 )
0.1P1 + 0.5 P2 ) = ( P1
P2 )
0.5 P2 = 0.1P1 and 0.5 P2 = 0.1P1

Thus:
Thus:
P1 = 2 3 and P2 =
P1 = 5 6 and P2 =
If there are 60,000 buyers
Before Advertising
P1 = 2 3
After Advertising
P1 = 5 6
this implies that,

2
3 60, 000 = 40, 000 customers will buy our
this implies that,
brand
contribution = 40, 000 2.5

= Sh.100,000
60, 000 = 50, 000 customers
will
buy our brand
contribution = (50, 000 2.5) 20, 000

= Sh.105,000
the difference between advertising and not advertising is

105,000 100,000 = Sh.5,000 in favour of advertising,
Thus the advertising campaign is worthwhile.
INPUT OUTPUT ANALYSIS

The input output analysis is a topic which requires application of matrices
The technique analyses the flow of inputs from one sector of the economy to the other sectors
thus the technique is quite useful in studying the interdependence of sectors within a single
economy.
The input output analysis was first developed by Prof Leontief hence the Leontief matrix has
been developed. See the following example
Lesson One
33
Example (B)
INPUT OUTPUT TABLE
TO
FROM
Agric
Industry
Service
Primary inputs
Total inputs
Final
Total
Agric
Industry
Service
Demand
Demand (output)
300
360
320
1080
2060
450
470
410
800
2130
610
500
520
270
1900
700
800
650
2060
2130
1900
-
NB: In the above table, one should be able to interpret the table e.g. of the total demand of
2060 metric tones from the agriculturalsector; 300 is produced for the agricultural sector, 360 for
industrial sector, 320 for the service sector and 1080 metric tones makes up the final demand.
The final demand is the additional demand besides the sectoral demand which is normally made
by other users e.g. government, foreign countries, other manufacturers not included in the other
sectors.
For production if items besides the inputs from other sectors namely labour capital e.t.c
Technical coefficients : (to sectors)
300
Agriculture 300 =
450 =
610 =
Industry
360 =
470 =
500 =
Service
320 =
410 =
520 =
2060
450
2060
610
2060
360
2130
470
2130
500
2130
320
1900
410
1900
520
1900
0.14
0.22
0.30
0.7
0.22
0.23
0.17
0.22
0.27
The matrix of technical coefficients is:

TO
FROM
Agric
Agric
0.14
Industry
0.7
Service
0.17
Final
Total
Demand
1080 (y1)
Demand (output)
2060
34
Industry
Service
Primary inputs
0.22
0.30
x
2060(x1)
0.22
0.23
x
2130(x2)
0.22
0.27
x
1900(x3)
800 (y2)
270(y3)
-
2130
1900
-
From the above table, we may develop the following equations

0.14x1 + 0.7x2 + 0.17x3 + y1 = x1
0.22x1 + 0.22x2 + 0.22x3 + y2 = x2
0.30x1 + 0.23x2 + 0.27x3 + y3 = x3
0.14 0.17 0.17 x1

y1
x1

0.22 0.22 0.22 x 2 + y 2 = x 2 ...................(*)
0.30 0.23 0.27 x

3
y3
x3
A
Let the coefficient matrix be represented by
a11 a12
A = a 21 a 22
a
31 a 32
a13
a 23
a 33
y1

y = y2
y
3
x1

x = x2
x
3
Equation (*) may be written as

AX + Y = X
Y = X AX
Y = X (I-A)
(I A)1 Y = X
The matrix I A is known as Leontief Matrix
Technical Coefficients
These show the units required from each sector to make up one complete product in a given
sector e.g. in the above matrix of coefficients it may be said that one complete product from the
agricultural sector requires 0.14 units from the agricultural sector itself, 0.22 from the industrial
sector and 0.30 from the service sector
NB: The primary inputs are sometimes known as value added
Example 1
Determine the total demand (x) for the industry 1, 2, 3 given the matrix of technical coefficients
(A), Capital and the final demand vector B.
1
A = 2
3
0.3 0.4 0.1
0.5 0.2 0.6

0.1 0.3 0.1
20

B = 10

30
34
Lesson One
35
From the input output analysis
X = (I A)1 B, Where X =
X1

X 2 is the demand vector
x
3
1 0 0
0.3 0.4 0.1
I - A = 0 1 0 - 0.5 0.2 0.6

0 0 1
0.1 0.3 0.1
0.7 -0.4 -0.1

= -0.5 0.8 -0.6
-0.1 -0.3 0.9
(I - A)-1
0.7 -0.4 -0.1
= -0.5 0.8 -0.6

-0.1 -0.3 0.9
The matrix of cofactors of ( I - A ) is

=
0.54 0.51 0.23
0.39 0.62 0.25

0.32 0.47 0.36
The transpose (adjoint) of the above matrix is
0.54 0.39 0.32
0.51 0.62 0.47

0.23 0.25 0.36
of (I - A)
= 0.495 (0.008 + 0.126 + 0.18

= 0.809
0.54 0.39 0.32

1
0.51 0.62 0.47

(I - A)-1 =
0.809
0.23 0.25 0.36
0.66 0.48 0.40
0.63 0.77 0.58

0.23 0.31 0.44
36
0.66 0.48 0.40 20

X1

0.63 0.77 0.58 10 = X 2
0.23 0.31 0.44 30

X3
Therefore X =
30
X = 37.7
21.9
The total demand from the three industries 1, 2 and 3 is 30 from 1, 37.7 from 2 and 21.9 from 3.
Example 2
Three clients of Disrup, Ltd P, Q and Rare direct competitors in the retail business. In the first
week of the year P had 300 customers Q had 250 customers and R had 200 customers. During
the second week, 60 of the original customers of P transferred to Q and 30 of the original
customers of P transferred to R. similarly in the second week 50 of the original customers of Q
transferred to P with no transfers to R and 20 of the original customers of R transferred to P
with no transfers to Q.
Required
a) Display in a matrix the pattern of retention and transfers of customers from the first to the
second week
(4 marks)
b) Re-expres the matrix that you have obtained in part (a) showing the elements as decimal
fractions of the original numbers of customers of P, Q and R (2 marks) Refer to this re
expressed matrix as B
c) Multiply matrix B by itself to determine the proportions of the original customers that have
been retained or transferred to P, Q and R from the second to the third week. (4 marks)
d) Solve the matrix equation (xyz)B = (xyz) given that x + y + z = 1
(8 marks)
e) Interpret the result that you obtain in part (d) in relation to the movement of customers
between P, Q and R
(2marks)
(Total 20 marks)
Solution
a). Think of each row element as being the point from which the customer originated and each
column element as being the destination e.g. 210 customers move from P to P, 60 move
from P to Q and 30 move from P to R. The sum of the elements of the first row totalling
300, that is the number of customers originally with P.
Hence required matrix is
P 210 60 30
From Q 50 200 0
R 20
0 180
To
row total 300
row total 250

row total 200
b). The requirement of this part is to express each element as a decimal fraction of its
corresponding row total. The second row, first element is therefore 50/250, that is 0.2 and
the second element is therefore 200/250 that is 0.8.
36
Lesson One
37
0.7 0.2 0.1
0.2 0.8 0
0.1 0 0.9
Hence B =
0.7 0.2 0.1
0.2 0.8 0
0.1 0 0.9
c).
0.7 0.2 0.1
0.2 0.8 0
0.1 0 0.9
0.54 0.30 0.16
= 0.30 0.68 0.02

0.16 0.02 0.82
The result can be checked by the normal rules of matrix multiplication.
d).
(x
0.7 0.2 0.1
y x ) X 0.2 0.8 0 = ( x
0.1 0 0.9
y z)
This produces from the first row

0.7x + 0.2y + 0.1z = x
Which reduces to
0.2y + 0.1z = 0.3x
Or
2y + z = 3x ...(i)
Or
The second row produces,
0.2x + 0.8y = y
Reducing to
0.2x = 0.2 y
X = y ..(ii)
Or
The third row produces
0.1x + 0.9z = z
Reducing to
0.1x = 0.1z
X = z .
(iii)
At this point you will notice that condition h (ii) and condition (iii) produce 2x + x = 3x when
substituted into condition (i), we therefore need extra condition x + y + z = 1 to solve the
problem.
Thus
x+x+x=1
Or
That is x = 1
Leading to
3x = 1
3
x= 1 ,
3
y= 1 ,
3
z= 1
38
e). In proportion terms this solution means that P, Q, and R will in the long term each have one
third of the total customers
Example 4
There are three types of breakfast meal available in supermarkets known as brand BM1, brand
BM2 and Brand BM3. In order to assess the market, a survey was carried out by one of the
manufacturers. After the first month, the survey revealed that 20% of the customers purchasing
brand BM1 switched to BM2 and 10% of the customers purchasing brand BM1 switched to
BM3. similarly, after the first month of the customers purchasing brand BM2, 25% switched to
BM1 and 10% switched to BM3 and of the customers purchasing brand BM3 0.05% switched to
BM1 and 15% switched to BM2
Required
i.
Display in a matrix S, the patterns of retention and transfers of customers from the first
to the second month, expressing percentage in decimal form.
(2marks)
Multiply matrix S by itself (that is form S2)
(5 Marks)
Interpret the results you obtain in part ii with regard to customer brand loyalty (3 marks)
ii.
iii.
Solution
The objective of the first part of the question was to test the candidates ability to formulate and
manipulate a matrix, then interpret the result of such manipulation.
a. i. The matrix showing the pattern of retention and transfer from the first to the second
month is
BM1 BM2 BM3

0.70
S = 0.25
0.05
0.20
0.65
0.15
0.10 BM1
0.10 BM2
0.80 BM3
(The second element in the first row shows the 20% movement from BM1 to BM2 and so on)
The product of matrix S with itself is demonstrated as follows
i.
0 .7 0
0 .2 5
0 .0 5
0 .2 0
0 .6 5
0 .1 5
0 .1 0
0 .1 0
0 .8 0
0 .7 0
0 .2 5
0 .0 5
0 .2 0
0 .6 5
0 .1 5
0 .1 0
0 .1 0
0 .8 0
0 .5 4 5 0
0 .3 4 2 5
0 .1 1 2 5
0 .2 8 5 0
0 .4 8 7 5
0 .2 2 7 5
0 .1 7 0 0
0 .1 7 0 0
0 .6 6 0 0
Where for example second element in the first row, that is 0.2850 is the result of multiplying the
corresponding elements of the first row of S by the second column of S and summing the
product.
0.2850 =
ii.
0.70 0.20 + 0.20 0.65 + 0.10 0.15

=
0.14 + 0.13 + 0.015 e.t.c.
The resulting matrix may be interpreted in the following way

Of the original customers who buy BM1, 54.5% will remain loyal to the brand in
month three, 28.5% will have switched to BM2 and 17% will have switched to
BM3.
38
Lesson One
39
Of the original customers who buy BM2, 48.7% will remain loyal to the brand in
month three 34.25% will have switched to BM1 and 17% will have switched to
BM3
Of the original customers who buy BM3, 66% will remain loyal to the brand in
month three 11.25% will have switched to BM1 and 22.75% will have switched to
BM2
Alternatively
In month three 54.5% of the customers buying BM1 are original customers. 34.5%
came from BM2 originally and remaining 11.25% have switched from BM3 and so
on.
MARKOV CHAINS/PROCESSES
The Markov processes are defined as a set of trials which follow a certain sequence which
depend on a given set of probabilities known as transition probabilities. These probabilities
indicate how a particular activity or product moves from one state to another.
Applications of Markov Chains in Business

The Markov processes or chains are frequently applied as follows:-
1. Brand Switching
By using the transitional probabilities we can be able to express the manner in which consumers
switch their tastes from one product to another.
2. Insurance industry
Markov analysis may be used to study the claims made by the insured persons and also decide
the level of premiums to be paid in future.
3. Movement of urban population

By formulating a transition matrix for the current population in the urban areas, one can be able
to determine what the population will be in say 5 years.
4. Movement of customers from one bank to another.

It is a fact that customers tend to look for efficient banks. Therefore at a certain time when a
given bank installs such machinery as computers it will tend to attract a number of customers
who will move from certain banks to efficient ones.
PROPERTIES OF MARKOV CHAINS

1. Each outcome in a markov process belongs to a state space or transition matrix. E.g.
S1
S2
S3
S1 P11
S2 P21
S3 P31
P12
P22
P32
P13
P23
P33
Where S1, S2, S3 are states and P11 P12 e.t.c are probabilities
2. The outcome of each trial depends on the immediate preceding activities but not on the
previous activities
BASIC TERMS IN MARKOV CHAINS

40

a) Probability Vector
This is a row matrix whose elements are non-negative and also they add up to 1 e.g. u = 0.2,
0.1, 0.2, 0.5)
Example
Consider
u =( , 0, - , 1/2 ) Not because is negative

v = ( , , 0, )
Not because the sum of the element <1
w= ( , , 0, )
Adds up to 1, each element is non negative.

Therefore its a prob, vector
State the ones which are probability vectors
b) Stochastic matrix
A matrix whose row elements are all non negative and also add up to 1.
Example (i)
0.1
0.0
=
0.5
0.3
0.2 0.3 0.4
0.7 0.1 0.2

0.1 0.1 0.3
0.4 0.2 0.1
Example ii) = Consider the following matrices
A=
13
3
4
1
3
0
1
2
1
3
14
1
3
2
3
14
B= 1
3
0 1 0
1 1 1
C= 2 6 3
1 2 0
3 3
1
3
3
4
A is not stochastic matrix because the element in the 2nd row and 3rd column is negative.
B is not Stochastic matrix because the elements in the second row do not add up to 1
C is stochastic matrix because each element is non negative and they add up to 1 in each row.
c) Regular stochastic matrix

m
A matrix P is said to be regular stochastic matrix if all the elements in P are all positive,
where m is a power, m = 1, 2, 3 e.t.c
Let A =
A2 =
A3 =
0 1
1 1
2 2
0 1
1 1
2 2
Where A is a Stochastic Matrix
12
0 1
1 1 1
2 2
4
0 1 12
1 1 = 1
2 2 4
14
= 3
3
4
8
1
2
3
4
5
8
1
2
3
4
40
Lesson One
41
Since the elements in A2 and A3 are all positive then A is regular Stochastic matrix.
ABSORBING STATES
A state Si (I = 1, 2, 3 ) of a markov chain is called absorbing if the system remains in the state,
Si once it enters there. Thus a state, Si is absorbing if and only if the ith row of the transition
matrix p has a 1 on the main diagonal and zeroes every where else. See the following example.
The following matrix, P is a transition matrix of the markov chain.
S1 S2
P=
1
S1 4
S2 0
S3 1
2
S4
0
S5
0
S3 S4
1
4
0
4
0
4
0
4
0
S5
1
4
0
0
1
The States S2 and S5 are absorbing states since the 2nd and 4th rows have 1 on the main diagonal.
42
REINFORCING QUESTIONS
Work out these questions and compare with answers given in lesson 9
QUESTION ONE
Determine
a) f(1)
b) f(-2)
for the following functions
1. f(x) = 5x + 2
2. f(x) = x2 + 3x+10
3. f(t) = 10 t + t3
4. f(u) =
QUESTION TWO
Solve
a)
2 x + 5 y = 20
the following simultaneous equations.
4x + y = 4
b)
2x y = 9
x + 3 y = 6
c)
12 x 4 y = 18
4 x + y = 6
QUESTION THREE
Because of inreasing cost increasing cost energy, the population within Maueni district
seem to be shifting from the north to the south the transition matrix S describes the
migration behaviour observed between the regions.
to north to south
0.90 0.10
S =
0.05 0.95
from north
from south
determine whether the populations will attain an equillibrium condition and if so, the
population of the two regions.
QUESTION FOUR
A simple hypothetical economy of three industries A, B and C is represented in the following
table (data in millions of shillings).
42
Lesson One
User
Producer
A
B
C
43
Final
Demand
Total
Output
80
80
80
100
200
100
100
60
100
40
60
20
320
400
300
Determine the output vector for the economy if the final demand changes to 60 for A, 60 for B
and 60 for C
QUESTION FIVE
A tea blender uses two types of tea, T1, and T2, to produce two blends, B1 and B2 for sale. B1
uses 40% of available T1 and 60% of the available T2 whilst B2 uses 50% of the available T1 and
25% of the available T2.
Required:
a) Given that t1 kilos of T1 and t2 of T2 are made available to produce b1 kilos of B1 and b2
kilos of B2. Express the blending operation in the matrix format.
b) If 400 kilos of T1 and 700 kilos of T2 were made available for blending, what quantities of B1
and B2 would be produced?
c) If 600 kilos of B1 and 450 kilos of B2 were produced, use a matrix method to determine
what quantities of T1 and T2 would be used to produce the blends.
QUESTION SIX
2 2
3 -3
Let A =
a) Find A2 and A3
b) If F(x) = x3 3x2 2x + 4I
Find F (A)
c) Find the inverse of matrix A
QUESTION SEVEN
A childs toy is marketed in three sizes standard size contains 10 squares (S), 15 triangles (T) and
6 hexagons (H). The deluxe set contains 15 S, 20 T, 6 H and 4 octagons (O). The super set
contains 24 T, 8 H, 16 H, 16 S and 6 (O). Squares cost 12 pence to produce, triangles cost 8p,
hexagons cost 18p and octagons 22p.
The standard set is sold at 6, the deluxe set for 10 and super for 15. The manufacturer
produces 100 standard sets, 80 deluxe sets and 50 super sets per week.
Use matrix form and matrix multiplications to find:
The cost of producing each set.
The number of each shape required each week
44
Total expenditure on shapes each week
QUESTION EIGHT
Matrix N below shows the number of items of type A, B, and C in warehouses Y and W. Matrix
p shows the cost in pence per day of storing (S) and maintaining (M) one item each of A, B and
C
B C
Y 10 12 50
N=
W 60 0 20
S M
A 2 0.5
P = B 3 1.5
C 2 0.5
a) Evaluate the matrix (NP) and say what it represents.

b) Stock movement occurs as follows:
At the start of the day 1:
Withdrawal of 2 type B from warehouse Y, 20 of type A from warehouse W.
At the start of day 2:
Delivery of 7 type B and 10 of type C to warehouse Y and 15 of type B to warehouse
W.
Evaluate the total cost of storage and maintenance for days 1 and 2.
c) Write down without evaluating a matrix expression which could be used to evaluate the
storage and maintenance cost of items A, B and C for the period from day 1 to 4.
Allow for the stock movements on days 1 and 2, as described in part (b). There were
no stock movements on days 3 and 4.
44
Lesson Two
45
LESSON TWO
Sets Theory and Calculus
Sets and set theory

Calculus
Differentiation and integration of polynomial, exponential and logarithmic functions
Application of calculus to economic models
2.1
Sets and set theory
A set is a collection of distinct objects. We may consider all the ocean in the world to be a set
with the objects being whales, sea plants, sharks, octopus etc, similarly all the fresh water lakes in
Africa can form a set. Supposing A to be a set
A = {4, 6, 8, 13}
The objects in the set, that is, the integers 4, 6, 8 and 13 are referred to as the members or
elements of the set. The elements of a set can be listed in any order. For example,
A = {4, 6, 8, 13} = {8, 4, 13, 6}
Sets are always precisely defined. Each element occurs once and only once in a set.
The notion is used to indicate membership of a set. represents non membership. However,
in order to represent the fact that one set is a subject of another set, we use the notion . A set
S is a subject of another set T if every element in S is a member of T
Example
If A = {4, 6, 8, 13} then
i)
4 {4, 6, 8, 13} or 4 A; 16 A
ii)
{4, 8} A; {5, 7} A; A A
Methods of set representation

Capital letters are normally used to represent sets. However, there are two different methods for
representing members of a set:
i.
The descriptive method and
ii.
The enumerative method
The descriptive method involves the description of members of the set in such a way that one
can determine the elements of the set without difficulty.
The enumerative method requires that one writes out all the members of the set within the curly
brackets.
For example, the set of numbers 0, 1, 2, 3, 4, 5, 6 and 7 can be represented ass follows
P = {0, 1, 2, 3, 4, 5, 6, 7} ,
enumerative method
P = {X/x = 0, 1, 27}
descriptive method
Or
P = {x/0 x 7} where x is an interger.
46
Finite and infinite sets

A set can be classified as a finite or infinite set, depending upon the number of elements it has.
A finite set has a finite number of elements whereas an infinite set has an infinite number of
elements.
For example, set P below has ten elements and is therefore a finite set. Set S, on the other hand,
is an infinite set since it has an infinite number of elements.
P = {2, 4, 620}
S = {1, 3, 5}
Universal set
The term refers to the set that contains all the elements that an analyst wishes to study.
The notation U or is generally used to denote universal sets
The null set or empty set
This is a set which contains no elements. It is normally designated by a
Greek letter , or { }.
The sets and { } are not the same thing since the former has no elements in it, while the
later has one element in it, namely zero
Equal or equivalent sets
Two sets C and D are said to be equal if every member of set C belongs to D and every member
of set D also belongs to C
Complement of a set
The complement of set A is written as A. This set contains all those elements of universal set
which are not in A
Intersection and union
B C denotes the intersection of B and C. it is the set containing all those elements, which
belong to both B and C
If B = {5, 8, 11, 20, 25} and C = {1, 3, 5, 7, 9, 11, 13}
Then B C = {5, 11}
B C = {1, 3, 5, 7, 8, 9, 11, 13, 20 25}

SET OPERATIONS AND SOME LAWS OF SET THEORY
THE VENN DIAGRAM
A simple way of representing sets and relations between sets is by means of the Venn diagram.
Venn diagram consists of a rectangle that represents the universal set. Subjects of the universal
set are represented by circles drawn within the rectangle, or the universe.
Suppose that the universal set is designated by U and the sets A, B and C are subject of U.The
Venn diagram below can be used to illustrate the sets as follows
46
Lesson Two
47
U
B
A
Venn diagram below representing the intersection of set A and B or A B = C is illustrated as

follows
Intersection of sets
Example:
You are given the universal set
T = {1, 2, 3, 4, 5, 6, 7, 8}
And the following subjects of the universal set:
A = {3, 4, 5, 6,}
B = {1, 3, 4, 7, 8}
Determine the intersection of A and B
Solution
The intersection of A and B is the subject of T, containing elements that belong to both A and B
A B = {3, 4, 5, 6,} {1, 3, 4, 7, 8}

= {3, 4}
or
T
U
A
5
6
3
4
1
7
8
48
Example
Consider the following universal set T and its subjects C, D and E
T = {0, 2, 4, 6, 8, 10, 12}
C = {4, 8,}
D = {10, 2, 0}
E = {0}
Find
ii)
DE
CDE
iii)
Solution
ii)
D E = {10, 2, 0} {0} = {0}

T
D
D E = Shaded area
ii) C D E = {4, 8} {10, 2, 0} {0} = { } =
T
D
C
4; 8
2; 10
Mutually exclusive or disjointed sets

Two sets are said to be disjointed or mutually exclusive if they have no elements in common.
Sets P and R below are disjointed
48
Lesson Two
49
Disjointed sets are represented by a null set in this case

PR =
The union of sets

Venn diagram representing the union of sets A and B or A B = Shaded area is illustrated
below;U
A B = Shaded area
Example
Consider the universal set T and its subsets A, B and C below:
T = {a, b, c, d e, f}
A = {a, d}
B = {b, c, f}
C = {a, c, e, f}
Find
ii)
iii)
iv)
v)
Solution
i)
ii)
iii)
iv)
A B
AC
B C
A B C
A B = {a, d} {b, c, f} = {a, b, c, d, f}
A C = {a, d} {a, c, e, f} = {a, c, d, e, f}
B C = {b, c, f} {a, c, e, f} = {a, b, c, e, f}
A B C = {a, d} {b, c, f } {a, c, e, f} = {a, b, c, d, e, f} = T
Complement of a set
Venn diagram representing the complement of a set say A represented by A1 is illustrated below.
50
A1
A1= shaded area

Example
For the universal set T = {1, 2, 3, 4, 5} and its subset A ={2, 3} and B ={5, }
Find
ii)
iii)
iv)
A1
(A1)1
(B1)1
Solution
i)
ii)
iii)
A1 ={2, 3}1 = {1, 4, 5}

(A1)1 =({2, 3}1)1 = {1, 4, 5}1={2, 3} = A
(B1)1=({5}1)1 = {1, 2, 3, 4}1={5} = B
Some Laws of Set Algebra

From the following Venn diagram where T is the universal set and A its subset, we can deduce a
number of laws.
T
i)
ii)
iii)
iv)
v)
vi)
vii)
viii)
A = A
AT = T
AA = A
A A= A
A T= A
A A1 = T
A A1=
(A1)1 = A
50
Lesson Two
51
Solving Problems Using Venn Diagrams

Example 1
Of the 20 girls in a form, 16 play hockey 12 play tennis and 4 play basketball. Every girl plays at
least one game and two play all the three. How many play two and only two games.
Solution
N() = 20
n(T) = 12
n(H) = 16
y
12 y z 2
16 x y 2
2
x
4 x z 2
n(B) = 4
H = Hockey
T = Tennis
B = Basketball
Those who play two and only two games = x + y + z
Using the diagram

(14 x y) + (10 y z) +(2 x z ) + 2 + x + y+ z = 20
28 x y z = 20
x+y+z=8
Example 2
250 members of a certain society have voted to elect a new chairman. Each member may vote
for either one or two candidates. The candidate elected is the one who polls most votes
Three candidates x, y z stood for election and when the votes were counted, it was found that
- 59 voted for y only, 37 voted for z only
- 12 voted for x and y, 14 voted for x and z
- 147 voted for either x or y or both x and y but not for z
- 102 voted for y or z or both but not for x
Required
i)
How may voters did not vote
How many voters voted for x only
ii)
Who won the elections
iii)
52
Solution
N() = 250
Y
Tail end
P
14
Q
37
P + 12 + 59 = 147 giving P = 76
Q + 59 + 37 = 102 giving Q = 6
i)
Those who did not vote

= 250 ( 76 + 12 + 14 + 59 + 6 + 37)
= 250 204 = 46
ii)
x = 76 + 12 + 14 = 102
y = 12 + 59 + 6 = 77
z = 37 + 14 + 6 = 57
iii)
X won the election
2.2
CALCULUS
Calculus is a branch of mathematics which explains how one variable changes in relationship to
another variable. It enables us to find the rate of change of one variable with respect to another
variable.
Example
i.
ii.
iii.
iv.
The rate at which business revenue is increasing at a particular stage when volume
of sales is increasing.
The rate at which costs are changing at a particular stage when volume of sales is
given
The evaluation of rate of change can help us to identify when the change in one
variable reaches a maximum or minimum.
Calculus may be used in production management when the production manager
wants to know
a) How much is to be manufactured in order to maximize the profits, revenues
e.t.c
52
Lesson Two
53
b) How much is to be produced in order to minimize the production costs

Calculus is divided into two sections namely:
Differentiation and integration

Differentiation deals with the determination of the rates of change of business activities or simply
the process of finding the derivative of a function.
Integration deals with the summation or totality of items produced over a given period of time or
simply the reverse of differentiation
The derivative and differentiation
The process of obtaining the derivative of a function or slope or gradient is referred to as
derivation or differentiation.
dy
or f(x) and is given by dividing the change in y variable by the
The derivative is denoted by
dx
change in x variable.
The derivative or slope or gradient of a line AB connecting points (x,y) and (x+dx, y + dy) is
given by
y
x
Change in y
Change in x
( y + dy ) y = dy
(x + dx ) x dx
Where dy is a small change in y and dx is a small change in x variables.

Illustration
Line AB
(y + dy)
B = (x + dx, y +dy)
dy
(x,y) = A
dx
(x + dx)
Rules of Differentiation
1. The constant function rule
If given a function y = k where k is a constant then
Example
Find the derivative of (i) y = 5
dy
dx
= 0
54
Solution
i. y = 5 dy =
dx
Illustration
y=5
slope =
dy 5 0
=
=0
dx
0
dy
derivative of a constant function
2. Power function rule

Given a function y = x r
Then
dy
= rx r 1
dx
Example
Find
dy for;
dx
(i).
y = x7
(ii).
(iii).
(iv).
y = x2
y = x-3
y=x
Solution
i.
y = x7
dy = 7x 7-1 = 7x6
dx
ii.
y = x2
dy = 2 x(2 - 1)
54
Lesson Two
55
dx
y = x-3
iii.
dy = -3x 3-1 = -3x-4

dx
iv.
y=x
dy = 1x 1-1 = 1.x0= 1 (since x0=1)
dx
3. Power function multiplied by a constant

If given y = Axr, then
dy = rAxr-1
dx
4. The sum rule

The derivative of the sum of two or more functions equals the sum of the derivatives of the
functions.
For instance
If H(x) = h(x) + g(x)
Then
dy or H(x) = h(x) + g(x)

dx
5. The difference rule

The derivative of the difference of two or more functions equals the difference of the derivatives
of the functions
If H () = h(x) g(x)
Then H() = h(x) g(x)
Examples.
Find the derivatives of
i.
y = 3x2 + 5x + 7
ii.
y = 4x2 2xb
Solution
i.
y = 3x2 + 5x + 7
2
dy d ( 3 x ) d ( 5 x ) d ( 7 )
=
+
+
dx
dx
dx
dx
= 6x + 5 + 0
= 6x + 5
ii.
y = 4x2 2xb
56
2
b
dy d ( 4 x ) d ( 2 x )
=
dx
dx
dx
= 8 x 2bx b 1
6. The product rule both factors are functions
The derivative of the product of two functions equals the derivative of the first function
multiplied by the second function PLUS the derivative of the second function multiplied by the
first function.
given that H ( x ) = h ( x ) .g ( x )
Then H ( x ) = h ( x ) .g ( x ) + h ( x ) .g ( x )
Example
Find
dy for
dx
i.
y = x2(x)
ii.
y = (x2+ 3) (2x3+ x2- 3)
Solution
i.
y = x2(x)
d (x )
d ( x)
dy
= x.
+ x2 .
dx
dx
dx
2
= x.2 x + x 2 .1
= 2x2 + x2
= 3x 2
Note that y = x2(x) = x3. Directly differentiating this we get 3x2.
ii.
y = (x2+ 3) (2x3+ x2- 3)

2
d ( 2 x3 + x 2 3)
dy d ( x + 3)
3
2
2
=
. ( 2 x + x 3) + ( x + 3) .
dx
dx
dx
2 x. ( 2 x3 + x 2 3) + ( x 2 + 3) . ( 6 x 2 + 2 x )
10 x 4 + 4 x3 + 18 x 2
56
Lesson Two
57
7. Quotient Rule
The derivative of the quotient of two functions equals the derivative of the numerator times the
denominator MINUS the derivative of the denominator times the numerator, all which are
divided by the square of the denominator
h (x )
If given H (x) =
g (x )
h ( x ) . g ( x ) h ( x ) . g ( x )
then H ( x ) =
g ( x )
For example
Find
i.
ii.
dy for
dx
x
3 + x2
x
3x + 7
Solutions
i.
x
3 + x2
d (3 + x2 )
d ( x)
2
.(3 + x )
.x
dy
dx
dx
=
2
dx
(3 + x2 )
(3 + x ) (2 x ).x
(3 + x )
3 + x 2 2x 2
3 x2
=
3 + x2
3 + x2
ii.
y=
x3
(3x + 7)
dy
dx
(3x 2 )(3x + 7) (3)(x 3 ) = 6x 3 + 21x 2

(3x + 7 )2
(3x + 7 )2
Example
A farmer of a large farm of poultry announced that egg production per month follows the
equation;
58

w=
3m3 m2
m2 + 10
Where w Total no of eggs produced per month

m amount in kilograms of layers mash feed.
Required
Determine the rate of change of w with respect to m (i.e. the rate at which the number of eggs
per month increase or decrease depending on the rate at which the kilos of layers marsh are
increased).
Solution
Let u = 3m3 m2
du = 9m2 2m
dm
Let v = m2 + 10
dv = 2m
dm
2
2
3
2
dw ( m + 10 )( 9m + 2m ) ( 3m m ) 2m
=
2
dm
( m2 + 10 )
9m 4 + 90m 2 2m3 20m ( 6m 4 + 2m3 )
(m
+ 10 )
3m 4 + 90m 2 20m
(m
+ 10 )
8. Chain Rule
This rule is generally applied in the determination of the derivatives of composite functions,
which can be defined as a function in which another function can be considered to have taken
the place of the independent variable. The composite function is also referred to as a function of
a function.
It is normally of the form y = (2x2 + 3)3. If we let u = (2x2 + 3), then y = u3.
In order to differentiate such an equation we use the formula
dy dy du
=
dx du dx
Solution
y = (2x2 + 3)3
Let u = 2x2 + 3
du = 2x
dx
58
Lesson Two
59
Let y = u3
dy
du
= 3u2
dy =
dx
dy . du = 3u2 x 4x = 12xu2
du dx
= 12x(2x2 + 3)2
Example
Consider the function
y = (x2 + 16x + 5)2
which can be decomposed into
y = u2
and u = x2 + 16x + 5. in this case y is a function of (x2 + 16x + 5)
Hence y = f(u) and u = g(x)

dy =
dx
dy . du
du dx
= (2u) (2x +16)

= 2 (x2 + 16x + 5) (2x + 16)
9. The derivative of a function raised to power r; the composite function

rule.
The derivative of a function raised to power r equals to the power r times the function which is
raised by power (r-1), all of which is multiplied by the derivative of the function
If y = [g(x)]r
Then dy = r[g(x)]r-1 . g(x)
dx
For example
Find
5
dy
given y = ( 3 x 2 + 4 x )
dx
Solution
4
dy
= 5 ( 3x 2 + 4 x ) . ( 6 x + 4 )
dx
Differentiation of an implicit function

An Implicit function is one of the y = x2 y + 3x2 + 50. it is a function in which the dependent
variable (y) appears also on the right hand side.
To differentiate the above equation we use the differentiation method for a product, quotient or
function of a function.
60
Solution
y = x2 y + 3x2 + 50
2
2
dy d ( x y ) d ( 3 x ) d ( 50 )
=
+
+
dx
dx
dx
dx
dy
dy
= y ( 2x ) + x2 + 6 x + 0
dx
dx
0 = 2 xy + x 2
dy dy
+ 6x
dx dx
0 = 2 xy + ( x 2 1)
( x 2 1)
dy
+ 6x
dx
dy
= 2 xy + 6 x
dx
dy ( 2 xy + 6 x )
=
dx
( x2 1)
Partial derivatives
These derivatives are used when we want to investigate the effect of one independent variable
on the dependent variable.
For example, the revenues of a farmer may depend on two variables namely; the amount of
fertilizer applied and also the type of the natural soil.
Let J = 30x2y + y2 + 50x + 60y
Where J = annual revenue in 000
x = type of soil
y = amount of fertilizer applied
Required
Determine the rate of change of the J with respect to x and y
Solution
J = 30x2y + y2 + 50x + 60y
Differentiating J with respect to x keeping y constant we have
dJ = 60xy + 50
dx
Differentiating Jwith respect to y keeping x constant we have
dJ = 30x2 + 2y + 60
dy
60
Lesson Two
61
Maxima, minima and points of inflexion

a) Test for relative maximum
Consider the following function of x whose graph is represented by the figure below
y = f(x)
dy = f(x)
dx
dy
=0
C dx
dy
>0 B
dx
dy
<0
dx
x1
x2
x3
y = f ( x)
Relative maximum point

The graph of the function slopes upwards to the right between points A and C and hence has a
positive slope between these two points. The function has a negative slope between points C
and E. At point C, the slope of the function is Zero.
Between points X1 and X2
and between X2 and X3
dy
> 0 Where X1 X < X2
dx
dy
< 0 Where X2 < X X3.
dx
Thus the first test of the maximum points require that the first derivative of a function equals
zero or
dy
= f ( x) < 0
dx
The second text of a maximum point requires that the second derivative of a function is negative
or
d2y
= f ( x ) < 0
dx 2
62
Example
Determine the critical value for the following functions and find out the critical value that
constitutes a maximum
y = x3 12x2 + 36x + 8
Solution
y = x3 12x2 + 36x + 8
then
dy
= 3x2 24x + 36 +0
dx
The critical values for the function are obtained by equating the first derivative of
iii.
the function to zero, that is:
= 0 or 3x2 24x + 36 = 0
dy
dx
Hence (x-2) (x-6) = 0
And x = 2 or 6
The critical values for x are x = 2 or 6 and critical values for the function are y = 40 or 8
ii. To ascertain whether these critical values of x will give rise to a maximum, we apply the
second text, that is
d2y
<0
d2x
= 3x2 24x + 36 and
dy
dx
= 6x - 24
d2y
d2x
a) When x = 2
Then d2y = -12 <0
d2x
b) When x = 6
Then
d2y = +2 > 0
d2x
Hence a maximum occurs when x = 2, since this value of x satisfies the second condition. X = 6
does not give rise to a local maximum
b) Tests for relative minimum
There are two tests for a relative minimum point
i.
The first derivative, that is
dy = f(x) = 0
dx
The second derivative, that is
ii.
d2y = f(x) > 0
dx2
Example:
For the function
h(x) = 1/3 x3 + x2 35x + 10
Determine the critical values and find out whether these critical values are maxima or minima.
Determine the extreme values of the function
Solution
i.
Critical values
62
Lesson Two
63
h(x) = 1/3 x3 + x2 35x + 10 and

h(x) = x2 + 2x 35
by first text,
then h(x) = x2 + 2x 35 = 0
or (x-5) (x+7) = 0
Hence x = 5 or x = -7
ii.
The determinant of the maximum and the minimum points requires that we test the
value x = 5 and 7 by the second text
H(x) = 2x + 2
a) When x = -7 h(x) = -12 <0
b) When x = 5 h(x) = 12>0
There x = -7 gives a maximum point and x = 5 gives a minimum point.
iii.
Extreme values of the function
h(x) = 1/3 x3 + x2 35x + 10
when x = -7, h(x) = 189 2/3
when x = 5, h(x) = -98 1/3
The extreme values of the function are h(x) = 189 2/3 which is a relative maximum
and h(x) = -98 1/3 , a relative minimum
c) Points of inflexion
Given the following two graphs, points of inflexion can be determined at points P and Q as
follows:
y=g(x)
P
k1
64
Diagram (i)
y
y =f(x)
Q
k2
The points of inflexion will occur at point P when

g(x) = 0
at
x = k1
at
x < k1
g (x) < 0
g (x) > 0
at
x > k1
and at point Q when
at
x = k1
f(x) = 0
at
x < k1
f(x) > 0
f (x) < 0
at
x > k1
Example
Find the points of inflexion on the curve of the function
y = x3
Solution
The only possible inflexion points will occur where
d2y
=0
dx 2
From the function given
dy
d2y
= 3 x 2 and
= 6x
dx
dx 2
Equating the second derivative to zero, we have
6x = 0 or x = 0
We test whether the point at which x = 0 is an inflexion point as follows
d2y
< 0 which means a downward concavity
dx 2
d2y
When x is slightly larger than 0,
> 0 which means an upward concavity
dx 2
When x is slightly less than 0,
64
Lesson Two
65
Therefore we have a point of inflexion at point x = 0 because the concavity of the curve changes
as we pass from the left to the right of x = 0
Illustration
y=x3
point of
inflexion
0
Example
1. The weekly revenue Sh. R of a small company is given by
R = 14 + 81x
x3
Where x is the number of units produced.
12
Required
i.
Determine the number of units that maximize the revenue
Determine the maximum revenue
ii.
Determine the price per unit that will maximize revenue
iii.
Solution
i.
To find maximum or minimum value we use differential calculus as follows
R = 14 + 18 x
x3
12
1
dR
= 81 .3 x 2
12
dx
d 2R
x
1
= 0 .3.2 x =
2
12
2
dx
66
put
dR
=0
dx
i.e. 81
1 2
x =0
4
which gives x = 18 or x = 18
d 2R
x
=
2
dx
2
thus when x = 18;
d 2R
= 9 which is negative indicating a maximum value
dx 2
Therefore at x = 18, the value of R is a maximum. Similarly at x = -18, the value of R is a

minimum. Therefore, the number of units that maximize the revenue = 18 units
ii. The maximum revenue is given by
R = 14 + 81 + 18 (18)3
12
= Shs. 986
ii.
The price per unit to maximize the revenue is

986 = 54.78 or Shs.54.78
18
2.3
INTEGRATION
It is the reversal of differentiation
An integral can either be indefinite (when it has no numerical value) or definite (have specific
numerical values)
It is represented by the sign f(x)dx.
Rules of integration
i.
The integral of a constant
adx = ax +c
where a = constant
Example
Find the following
a) 23dx
b) 2dx. (where is a variable independent of x, thus it is treated as a constant).
Solution
i.
23dx = 23x + c
ii.
2dx. = 2 x + c
ii. The integral of x raised to the power n
x dx = n + 1 x
n
n +1
+c
66
Lesson Two
67
Example
Find the following integrals
a) x2dx
b) x-5/2 dx
Solution
i)
ii )
x dx = x + c
x dx = x
2
1
3
52
32
2
3
+c
iii). Integral of a constant times a function
af ( x )dx = a f ( x ) dx
Example
Determine the following integrals
i.
ax3dx
ii.
20x5dx
Solution
a)
ax dx = a x dx
3
= a4 x 4 + c
b)
20 x dx = 20 20 x dx
5
= 103 x 6 + c
iv). Integral of sum of two or more functions
{f(x) + g(x)} dx = f(x)dx + g(x) dx

{f(x) + g(x) + h(x)}dx = f(x)dx + g(x)dx + h(x)dx
Example
Find the following
i.
(4x2 + x-3) dx
ii.
(x3/4 + 3/7 x- + x5)
Solution
i)
( 4x
+ 12 x 3 dx = 4 x 2 dx + 12 x 3dx
= 43 x3 14 x 2 + c
ii )
( x
3
4
+ 73 x
7
12
+ x5 dx = x 4 dx + 73 x 2 dx + x 5 dx
1
= 74 x 4 + 76 x 2 + 16 x 6 + c
68
5. Integral of a difference
{f(x) - g(x)} dx = f(x)dx - g(x) dx

Definite integration
Definite integrals involve integration between specified limits, say a and b
b
The integral
f ( x ) dx Is a definite integral in which the limits of integration are a and b

a
The integrals is evaluated as follows

1.
2.
3.
4.
Compute the indefinite integral f(x)dx. Supposing it is F(x) + c

Attach the limits of integration
Substitute b(the upper limit) and then substitute a (the lower limit) for x.
Take the difference and the result is the numerical value for the definite
integral.
Applying these steps to the definite integral

b
f ( x ) dx = F ( x ) + c a
b
= F ( b ) + c F ( a ) + c
= F (b ) F ( a )
Example
Evaluate
i.
(3x 2 + 3)dx
(x + 15)dx
ii.
Solution
a.
(3x 2 + 3)dx = [(x 3 + 3x + c)]
= (27 + 9 + c) (1 + 3 + c)
= 32
b.
(x + 15)dx = [( x2 + 15x + c)] 50
= (12 + 75 + c) (0 + 0 + c)
= 87
The numerical value of the definite integral
f(x)dx can be interpreted as the area bounded by
the function f(x), the horizontal axis, and x=a and x=b see figure below
68
Lesson Two
69
y = f(x)
f(x)
area under curve

Therefore
f(x)dx = A or area under the curve
Example
1. You are given the following marginal revenue function
MR = a + a1q
Find the corresponding total revenue function
Solution
Total revenue = MR.dq =
( a + a q )dq
1
= aq a1q + c
1
2
Example 2
A firm has the following marginal cost function
MC = a a1q + a2 q 2
Find its total cost function.
Solution
The total cost C is given by
C
= MC.dq
= (a + a1q + a2q2).dq
= aq +
a1
2
q2 +
a2
3
q3 + c
70
Note: Exams focus: Note the difference between marginal function and total function. You
differentiate total function to attain marginal function, this is common in exams,
total profit = total revenue total cost.
Example 3.
Your company manufacturers large scale units. It has been shown that the marginal (or variable)
cost, which is the gradient of the total cost curve, is (92 2x) Shs. thousands, where x is the
number of units of output per annum. The fixed costs are Shs. 800,000 per annum. It has also
been shown that the marginal revenue which is the gradient of the total revenue is (112 2x)
Shs. thousands.
Required
i.
ii.
iii.
iv.
Solution
i.
Establish by integration the equation of the total cost curve

Establish by integration the equation of the total revenue curve
Establish the break even situation for your company
Determine the number of units of output that would
a) Maximize the total revenue and
b) Maximize the total costs, together with the maximum total revenue and total
costs
First find the indefinite integral limit points of the marginal cost as the first step to
obtaining the total cost curve
Thus (92 2x) dx = 92x x2 + c
Where c is constant
Since the total costs are the sum of variable costs and fixed costs, the constant term in the
integral represents the fixed costs, thus if Tc are the total costs then,
Tc = 92x x2 + 800
or Tc = 800 + 92x - x2
ii.
As in the above case, the first step in determining the total revenue is to form the
indefinite integral of the marginal revenue
Thus (112 - 2x) dx = 112x x2 + c
Where c is a constant
The total revenue is zero if no items are sold, thus the constant is zero and if Tr represents the
total revenue, then
Tr = 112x x2
iii.
At break even the total revenue is equal to the total costs

Thus 112x x2 = 800 + 92x - x2
20x = 800
x = 40 units per annum
iv.
a) Tr = 112x x2
d (Tr )
= 112 x 2 x
dx
70
Lesson Two
71
d 2 (Tr )
= 2
dx 2
at the maximum point
d 2 (Tr )
=0
dx 2
that is 112 2 x = 0

Since
d 2 (Tr )
= 2 this confirms the maximum
dx 2
The maximum total revenue is Shs. (112 x 56 56 x 56) x 1000

= Shs. 3,136,000
ii. Tc = 800 + 92 x x2
d (Tc )
= 92 2 x
dx
d 2 (Tc )
= 2 x
dx 2
At this maximum point
d (Tc )
=0
dx
92 2x = 0
92 = 2x
since
d 2 (Tc )
= 2 x this confirms the maximum
dx 2
the maximum costs are Shs. (800 + 92 x 46 - 46 x 46) x 1000
= Shs. 2,916,000
72
REINFORCING QUESTIONS
QUESTION ONE
Find the derivative of
a)
y = 6x x
1
x2
b)
y=
c)
y = 1 + 2x
d)
y=
1
x
QUESTION TWO
A cost function is
Ksh.(c) = Q2 30Q + 200
Where Q = quantity of units produced
Find the point of minimum cost.
QUESTION THREE
250 members of a certain society have voted to elect a new chairman. Each member may vote
for either one or two candidates. The candidate elected is the one who polls most votes.
Three candidates x, y, z stood for election and when the votes were counted, it was found that,
59 voted for y only, 37 voted for z only
12 voted for x and y, 14 voted for x and z
147 voted for either x or y or both x and y but not for z
102 voted for y or z or both but not for x.
Required:
i)
How many voters did not vote?
How many voters voted for x only?
ii)
Who won the election?
iii)
QUESTION FOUR
The weekly revenue Ksh.R of a small company is given by:
R = 14 + 81x x3 where x is the number of units produced
12
Required:
a) Determine the number of units that maximize the revenue.
b) Determine the maximum revenue.
c) Determine the price per unit that will maximize the revenue
72
Lesson Two
73
QUESTION FIVE
A furniture firm has two operating departments; Production and sales. The firmss operating
costs are split between these two departments with the resultant period of fixed costs of
Shs.20,000 and Shs.6,000 respectively. The production department has a basic variable cost per
unit of Shs.6 plus additional variable cost per unit of Shs.0.0002 which relates to all the
manufactured items during the period. The sales department has a variable cost per unit of
Shs.2. The sales department receives the finished goods from the production department and
pay the basic variable cost per unit plus 80% of the same.
NB:
Demand Q is given by the following function:

Q = 40,000 2,000P, where P is the selling price of the sales department.
Required:
a) Calculate the quantity that maximizes the profits of the production department.
b) Calculate the selling price that maximizes the profits of the sales department.
c) Determine the firms profit as a result of adopting the quantity and selling prices
in i and ii.
d) Determine the quantity and selling price that maximize the ships profit. What is the
amount of this profit?
QUESTION SIX
a) Describe how quadratic equations can be used in decision making.
b) The demand for a commodity is given by p = 400 q. The average total cost of producing
the commodity is given by
ATC =
1000
+ 100 5q + q 2
q
Required
i) What does
1000
q
where p is the price in shillings and q is the quantity in

kilograms.
in the ATC equation represent economically?
ii) Determine the output that leads to maximum profit and the profit at the
level of output.
(1 mark)
(9 marks)
c) Alpha industries sells two products, X and Y, in related markets, with demand functions
given by:
Px 13 + 2X + Y = 0
Py 13 + X + 2Y = 0
The total cost, in shillings, is given by:
TC = X + Y
Required:
Determine the price and the output for each good which will maximize profits.
(7 marks)
(Total: 20 marks)
74
QUESTION SEVEN
a) The following table shows the Fixed Cost (F) and the variable cost (V) of producing 1 unit
of X and 1 unit of Y:
Cost F
Cost V
Product
Y
8
12
X
5
4
(Shs 000)
When x units of X and y units of Y are produced, the total fixed cost is Shs.640,000 and total
variable cost is Shs.820,000. Express this information as a matrix equation and hence find the
quantities of x and y produced using matrix algebra.
(10 marks)
The marginal productivity of an industrial operation (the production of electric furnaces) is given
by:
f ( x) =
60
x2
+ 10
Where x is capitalization in millions of shillings. Given that, when the capitalization is Shs.
Million they can produce 62 of the furnaces per week.
Required:
a) How many furnaces they will be able to produce if their capitalization increased to Shs 10
million.
(10 marks)
b) What does the term marginal of productivity mean?
(Total: 20 marks)
Compare your solutions with those given in lesson 9
74
Lesson Two
75
COMPREHESIVE ASSIGNMENT ONE

Work out these question for three hours (exam condition) then hand them in to DLC for marking
Instructions:
Answer any THREE questions from SECTION I and TWO questions from SECTION II.
Marks allocated to each question are shown at the end of the question. Show all your workings
SECTION I
QUESTION ONE
a) Explain the importance of set theory in business.
(4 marks)
(8 marks)
b) By use of matrix algebra, develop the Leontief inverse matrix.
c) Digital Ltd. Manufactures and sells floppy disks at Nairobi Industrial Area.
The average total cost (ATC) and Average Revenue (AR) (in thousands of shillings)
of producing x floppy disks are given by the following functions:
ATC =
1 2 5
500
x x + 50 +
2
2
x
And
AR = 800 2x2
Where: x is the number of floppy disks produced
Required:
i) The profit function
ii) The number of floppy disks required to maximize profit
iii) The maximum profit
(3 marks)
(3 marks)
(2 marks)
(Total: 20 marks)
QUESTION TWO
Define the following terms as used in Markov analysis:
Markov process
Equilibrium or steady state
Absorbing state
Closed state
(2 marks)
(2 marks)
(2 marks)
(2 marks)
The manufacturer of Tamu Soft drinks has been facing stiff competition on its main brand
Tamu-cola soda. The management is considering an extensive advertising and rebranding
campaign for Tamu-Cola soda. If the current branding remains, the transition matrix of
consumer between Tamu-Cola and other brands will be as follows:
From
Tamu-Cola
Others
To
Tamu-Cola
Others
0.85
0.15
0.25
0.75
The advertising and rebranding campaign is expected to cost Sh.20 million each year.
76
There are 40 million consumers of soft drinks in the market and for each consumer the average
profitability is Sh.5 annually.
Required:
The equilibrium state proportion of consumers using Tamu-Cola before the advertising
campaign.
(4 marks)
The equilibrium state proportion of consumers using Tamu-Cola after the advertising
ii)
campaign.
(4 marks)
iii)
The expected annual profit increase or decrease after the advertising campaign. Would you
recommend the advertising campaign?
(4 marks)
(Total: 20 marks)
i)
QUESTION THREE
a) A market researcher investigating consumers preference for three brands of beverages
namely: coffee, tea and cocoa, in Ongata town gathered the following information:
From a sample of 800 consumers, 230 took coffee, 245 took tea and 325 took cocoa, 30
took all the three beverages, 70 took coffee and cocoa, 110 took coffee only, 185 took cocoa
only.
Required:
i)
Present the above information in a Venn Diagram.
ii)
The number of customers who took tea only.
The number of customers who took coffee and tea only.
iii)
The number of customers who took tea and cocoa only.
iv)
The number of customers who took none of the beverages.
v)
(4 marks)
(2 marks)
(2 marks)
(2 marks)
(2 marks)
b)
i)
ii)
Explain the importance of the Chi-square significance test

(2 marks)
The number of books borrowed from Millennium town library during a particular
week was recorded as shown below:
Days of the week

Number of books
borrowed
Monday
132
Tuesday
110
Wednesday
128
Thursday
105
Required:
Test the hypothesis that the number of books borrowed does not depend
on the day of the week at the 1% significance level.
Friday
150
Total
625
(6 marks)
(Total: 20 marks)
QUESTION FOUR
a) The general multiple linear regression equation is expressed as:
n
Yi = o + i X i + i
i =1
Where
Yi is the response variable
Xi are the explanatory variables
o is the constant
76
Lesson Two
77
1 are the parameters, and
i is the error term
Required:
Express the above multiple linear regression equation in a matrix form. Clearly indicate the
size of each vector column and the matrix.
(10 marks)
b) Mambo Company Ltd. Manufactures five products V, W, X Y and Z. The company has
divided its sales team into three regions; A, B and C. The Matrix Q below represent the
expected sales quantities in thousands for each product in each sales region for the coming
year.
Q=
Region
B
C
20
35
30
10
42.5
5
15
35
17.5
22.5
A
50
40
25
10
25
V
W
X
Y
Z
products
Each product is manufactured using combinations of four standard components. The

matrix T below indicates the number of units of each component used in producing each
product.
T=
Components
2
1
1
2
0
2
1
1
0
3
0
1
3
2
1
1
3
3
4
0
2
1
1
1
V
W
X
Y
Z
products
The manufacture of each component requires the use of certain resources. The matrix M
below indicates the quantities of the three standard parts and the number of production
labour hours and assembly labour hours used to produce one unit of each component.
M=
Part 1
Part 2
Part 3
2
0
1
2
1
3
2
5
0
2
1
4
Productions
Labour hours
2
4
1
1
Assembly
hours
1
3
6
2
1
2
3 components
4
The costs of the resources in matrix M are Part 1 Sh.20, Part 2 Sh.10, Part 3 Sh.30 while
each labour hour in the production and assembly departments cost Sh.15 and Sh.5
respectively.
Required:
i) The total expected demand for each product.
ii) The quantities of each component needed in the production process.
iii) The quantities of each resource required in the production.
(2 marks)
(3 marks)
(3 marks)
78

iv) The total cost of producing the required units of each product.
(2 marks)
(Total: 20 marks)
QUESTION FIVE
The Young Childrens Fund (YCF) is planning its annual fund-raising campaign for its
December school holiday camp for disadvantaged children. Campaign expenditures will be
incurred at a rate of Sh.10,000 per day. From past experience, it is known that contributions will
be high during the early stages of the campaign and will tend to fall off as the campaign
continues. The function describing the rate at which contributions are received is:
C (t) = 100t2 + 200,000
Where t = days of the campaign
C (t)
= rate at which contributions are received in shillings per day
The fund wants to maximize the net precedes from the campaign.
Required:
i) The number of days the campaign should be conducted to maximize the
net proceeds.
ii) The total campaign expenditure
iii) The total contributions expected to be collected
iv) Net proceeds from the campaign
(3 marks)
(2 marks)
(5 marks)
(1 mark)
The national office of a car rental company is planning its maintenance for the next year. The
companys management are interested in determining the companys needs for certain repair
parts. The company rents saloon cars, station wagons and double cab pick-ups. The matrix N
shown below indicates the number of each type of vehicle available for renting in the four
regions of the country.
Saloons
N=
160
150
100
120
Station
wagons
400
300
100
400
Double
cabs
500
200
150
300
Coast
Central
Western
Highlands
Four repair parts of particular interest, because of their cost and frequency of replacement, are
fan belts, spark plugs, batteries and tyres. On the basis of studies of maintenance records in
different parts of the country, the management have determined the average number of repair
parts needed per car during a year.
These are summarized in matrix R below:
Saloons
N=
17
12
9
4
Station
wagons
16
8
7
7
Double
cabs
15
5
5
6
Required:
i) The total demand for each type of car.
Fan belts
Plugs
Batteries
Tyres
(3 marks)
78
Lesson Two
79
ii) The total number of each repair part required for the fleet.
(3 marks)
iii) If matrix C below contains the cost per unit in shillings for fan belts, spark plugs, batteries
and tyres, calculate the total cost s for all repair parts. C = (1250,800,6500,8000). (3 marks)
(Total: 20 marks)
SECTION II
QUESTION SIX
a) Define the following terms as used in input-output analysis:
i) Transactions table.
ii) Primary inputs.
iii) Technical coefficients.
(2 marks)
(2 marks)
(2 marks)
b) Briefly explain the importance of input-output analysis.
(4 marks)
c) A small economy has three main industries which are steel, motor vehicles and construction.
The industries are interdependent. Each unit of steel output requires 0.2 units from steel,
0.3 units from motor vehicles and 0.4 units from construction. A unit of motor vehicles
output requires 0.2 units from steel, 0.4 units from motor vehicles and 0.2 units from
construction. A unit of construction output requires 0.3 units from steel, 0.4 units from
motor vehicles and 0.1 units from construction. The final demand is 20 million units from
steel. 50 million units from motor vehicles and 30 million units from construction.
Required:
i) The technical coefficient matrix.
ii) Total output of each industry, given that the Lentiefs inverse matrix is
1__
0.192
0.46
0.43
0.30
0.24
0.60
0.24
(2 marks)
0.26
0.41
0.42
(3 marks)
iii) If the final demand from steel drops by 2 million units, and that from motor vehicles
increases by 10 million units, but there is no change in the final demand from construction,
what would be the change in the total output of constructions?
(5 marks)
(Total: 20 marks)
QUESTION SEVEN
a) Explain the purpose of Venn diagram
(3 marks)
b) A market study taken at a local sporting goods store, Maua Wahome Stores showed that of
the 200 people interviewed, 60 owned tents, 100 owned sleeping bags, 80 owned camping
stoves, and 40 owned both tents and camping stoves and 40 owned both sleeping bags and
camping stoves.
Required:
If 20 people interviewed owned a tent, a sleeping bag and a camping stove, determent how
many people owned only a camping stove. In this case, is it possible for 30 people to own
both a tent and a sleeping bag, but not a campaign stoves?
(6 marks)
80
Under One Thousand Shillings Corner Store is planning to open a new store on the
corner of Main and Crescent Streets. It has asked the Tomorrows Marketing company to
do a market study of randomly selected families within a five kilometers radius of the
store,.the questions it wishes Tomorrows Marketing Company to ask each home-owner
are:
i)
Family income
Family size
ii)
iii)
Distance from home to the store site
Whether or not the family owns a car or uses public transport
iv)
Required:
For each of the four questions, develop a random variable of interest to Under One Thousand
Shillings Corner Store. Denote which of these are discrete and which are continuous random
variables.
(11 marks)
(Total: 20 marks)
QUESTION EIGHT
Two CPA students were discussing the relationship between average cost and total cost. One
student said that since average cost is obtained by dividing the cost function by the number of
units Q, it follows that the derivative of the average cost is the same as marginal cost, since the
derivative of Q is 1.
Required:
Comment on this analysis.
(4 marks)
Gatheru and Karibu Certified Public Accountants have recently started to give business advice
to their clients. Acting as consultants, they have estimated the demand curve of a clients firm to
be;
AR = 200 Q
Where AR is average revenue in millions of shillings and Q is the output in units.
Investigations of the clients firms cost profile shows that marginal cost (MC) is given by:
MC = Q2 28Q + 211 (in millions of shillings)
Further investigations have shown that the firms cost when not producing output is Sh.10
million.
Required:
i) The equation of total cost.
ii) The equation of total revenue
iii) An expression for profit
iv) The level of output that maximizes profit.
v) The equation of marginal revenue.
(5 marks)
(2 marks)
(2 marks)
(5 marks)
(2 marks)
(Total: 20 marks)
80
Lesson Three
81
LESSON THREE
Descriptive Statistics and Index Numbers
Contents
-
i.
ii.
iii.
iv.
v.
vi.
Application of statistics
Measures of centra tendency
i.
Arithmetic mean
Median
ii.
Mode
iii.
Geometric mean
iv.
Harmonic mean
v.
Measures of dispersion
Simple range
Quartile deviation
Mean deviation
Standard deviation
Coefficient of mean deviation
Coefficient of quartile deviation
Skew ness and Kurtosis
Indices
82
3.1
Descriptive Statistics
a) Statistics
Definition: Statistics viewed as a subject is a process of collecting, tabulating and analyzing
numerical data upon which significant conclusions are drawn.
Statistics may also be defined as numerical data, which has been, collected from a given source
and for a particular purpose e.g. population statistics from the ministry of planning, Agricultural
statistics from the ministry of Agriculture
Statistics may also refer to the values, which have been obtained from statistical calculations e.g.
the mean, mode, range e.t.c.
b) Application of statistics
1. Quality Control
Usually there is a quality control departments in every industry which is charged with the
responsibility of ensuring that the products made do meet the customers standards e.g. the
Kenya bureau of standards (KeBS) is one of the national institutions which on behalf of the
government inspects the various products to ensure that they do meet the customers
specification.
The KeBS together with other control department have developed quality control charts. They
use these charts to check whether the products are up to standards or not.
2. Statistics may be used in making or ordering economic order quantities (EOQ). It is important
for a business manager to realize that it is an economic cost if one orders a large quantity of
items which have to be stored for too long before they are sold. This is because the large stock
holds a lot of capital which could otherwise be used in buying other items for sale.
It is also important to realize that the longer the items are stored in the stores the more will be
the storage costs
On the other hand if one orders a few items for sale he will incur relatively low storage expenses
but may not be able to satisfy all the clients. These may lose their customers if the goods are out
of stock. Therefore it is advisable to work out the EOQ which will be sufficient for the clients in
a certain period before delivery.
The EOQ will also ensure that minimal costs are incurred in terms of storage
3. Forecasting
Statistics is very important for business managers when predicting the future of a business for
example if a given business situation involves a dependent and independent variables one can
develop an equation which can be used to predict the output under certain given conditions.
4. Human resource management
Statistics may be used in efficient use of human resources for example we may give
questionnaires to workers to find out where the management is weak
By compiling the statistics of those who were signing it may be found useful to analyze such data
to establish the causes of resignation thus whether it is due to frustration or by choice.
3.2
Measures of Central Tendency
These are statistical values which tend to occur at the centre of any well ordered set of data.
Whenever these measures occur they do not indicate the centre of that data. These measures are
as follows:
The arithmetic mean
i.
The mode
ii.
iii.
The median
iv.
The geometric mean
v.
The harmonic mean
82
Lesson Three
1.
83
The arithmetic mean
This is commonly known as average or mean it is obtained by first of all summing up the
values given and by dividing the total value by the total no. of observations.
X
i.e. mean
=
n
Where x = no. of values
= summation
n = no of observations
Example
The mean of 60, 80, 90, 120
60 + 80 + 90 +120
4
350
4
= 87.5
The arithmetic mean is very useful because it represents the values of most observations in the
population.
The mean therefore describes the population quite well in terms of the magnitudes attained by
most of the members of the population
Computation of the mean from grouped Data i.e. in classes.

The following data was obtained from the manufacturers of electronic cells. A sample of
electronic cells was taken and the life spans were recorded as shown in the following table.
Life span hrs
No. of cells (f)
Class MP(x)
XA=d
fd
1600 1799
25
1699.5
-600
-15000
1800 1999
32
1899.5
-400
-12800
2000-2199
46
2099.5
-200
-9200
2200 2399
58
2299.5(A)
2400 2599
40
2499.5
200
8000
2600 2799
30
2699.5
400
12000
2800 2999
2899.5
600
4200
A = Assumed mean, this is an arbitrary number selected from the data, MP = mid point
Arithmetic mean
assumed mean +
2299.5 +-53.78
fd
f
= 2299.5 +
12800
238
84
= 2245.72 hours
Example 2 (use of the coded method)

The following data was obtained from students who were registered in a certain college.
The table shows the age distribution
Age (yrs)
No. of Students (f)
mid points (x)
x-a = d
D/c = u
fu
15 19
21
17
-15
-3
-63
20 24
35
22
-10
-2
-70
25 29
38
27
-5
-1
-38
30 34
49
32(A)
0
0
0
35 39
31
37
+5
+
31
40 44
19
42
+10
+2
38
193
-102
Required calculate the mean age of the students using the coded method
Actual mean =
A(assumed mean) +
32 +
fu
f
102
5
193
= 29.36 years
NB. The following statistical terms are commonly used in statistical calculations. They must
therefore be clearly understood.
i) Class limits
These are numerical values which limits uq extended of a given class i.e. all the observations in a
given class are expected to fall within the interval which is bounded by the class limits e.g. 15 &
19 are class limits as in the table of the example above.
ii) Class boundaries
These are statistical boundaries, which separate one class from the other. They are usually
determined by adding the lower class limit to the next upper class limit and dividing by 2 e.g. in
19 + 20
.
the above table the class boundary between 19 and 20 is 19.5 which is =
2
iii) Class Mid points

This are very important values which mark the center of a given class. They are obtained by
adding together the two limits of a given class and dividing the result by 2.
iv) Class interval/width

This is the difference between an upper class boundary and lower class boundary. The value
usually measures the length of a given class.
84
Lesson Three
2.
85
The mode
This is one of the measures of central tendency. The mode is defined as a value within a
frequency distribution which has the highest frequency. Sometimes a single value may not
exist as such in which case we may refer to the class with the highest frequency. Such a class
is known as a modal class
The mode is a very important statistical value in business activities quite often
business firms tend to stock specific items which are heavily on demand e.g. footwear,
clothes, construction materials (beams, wires, iron sheets e.t.c.
The mode can easily be determined form ungrouped data by arranging the figures
given and determining the one with the highest frequency.
When determining the values of the mode from the grouped data we may use the
following methods;i. The graphical method which involves use of the histogram
ii. The computation method which involves use of formula
Example
In a social survey in which the main purpose was to establish the intelligence quotient (IQ) of
resident in a given area, the following results were obtained as tabulated below:
IQ
No. of residents
1 20
21 40
41 60
61 80
81 100
101 120
121 140
Upper class bound
6
18
32 fo
48 f1
27 f2
13
2
20
40
60
80
100
120
140
Required
Calculate the modal value of the IQs tabulated above using
i.
The graphical method and
ii.
Formular
CF
6
24
56
104
131
144
146
86
Graphical method
50
40
30
20
10
20
40
60
80
100
120
140
Value of the mode
Computation method
f1 f 0
Mode = L +
c
2f1 f 0 f 2
Where L = Lower class boundary of the class containing the mode
f0 = Frequency of the class below the modal class
f1 = Frequency of the class containing the mode
f2 = frequency of the class above the modal class
c = Class interval
( 48 - 32 ) 20
2 ( 48 ) - 32 - 27
Therefore Mode = 60.50 +
= 69.14
3.
-
The median
This is a statistical value which is normally located at the center of a given set of data which
has been organized in the order of magnitude or size e.g. consider the set 14, 17, 9, 8, 20, 32,
18, 14.5, 13. When the data is ordered it will be 8, 9, 13, 14, 14.5, 17, 18, 20, 32
The middle number/median is 14.5
The importance of the median lies in the fact that it divides the data into 2 equal halves. The
no. of observations below and above the median are equal.
86
Lesson Three
-
87
In order to determine the value of the median from grouped data. When data is grouped the
median may be determined by using the following methods
i.
Graphical method using the cumulative frequency curve (ogive)
The formula
ii.
Example
Referring to the table in 105, determine the median using the methods above
The graphical method

IQ
0 20
20 40
40 60
60 80
80 100
100 120
120 140
No of resid
6
18
32
48
27
13
2
146
UCB
20
40
60
80
100
120
140
Cumulative Frequency
6
24
56
104
131
144
146
xv
160
140
120
100
80
60
40
20
20
40
60
80
100
120
Value of the median
The position of the median =
n+1 146+1
=
2
2
ii Computation
The formula used is
140
160
88
n +1 Cfbm
Median = L + 2
c
cfmc
Where L = Lower class boundary of the class containing the median

N = No of observations
Cfo= cfbm = Cumulative frequency of the class before that containing
the median
F1 fmc = Frequency of the class containing the median
Therefore median = L +
= 60 +
73.5 - 56
48
n +1
2
- Cfbm
cfmc
20
= 60 + 7.29
= 67.29
4.
Geometric mean
This is a measure of central tendency normally used to measure industrial growth rates.
It is defined as the nth root of the product of n observations or values
i.e. GM =
x1 x 2 ... x n
Example
In 1995 five firms registered the following economic growth rates; 26%. 32% 41% 18% and
36%.
Required
Calculate the GM for the above values
GM =
26 32 41 18 36
= 15 [ Log 26 + Log 32 + Log 41 + Log18 + Log 26]

No. Log
26
32
41
18
36
Therefore Log of GM
1.4150
1.5052
1.6128
1.2533
1.5563
7.3446
= 1/5 x 7.3446 = 1.46892
So GM = Antilog of 1.46892
= 29.43
88
Lesson Three
5.
89
Harmonic mean
This is a measure of central tendency which is used to determine the average growth rates for
natural economies. It is defined as the reciprocal of the average of the reciprocals of all the
values given by HM.
HM =
x1
x2
+ ...
x3
Example
The economic growth rates of five countries were given as 20%, 15%, 25%, 18% and 5%
Calculate the harmonic mean
1
The HM =
1 (1 +1 +1 +1 +1
5 20
15
25
10
5
1
0.2(0.05 + 0.07 + 0.04 + 0.10 + 0.2)
1
0.092
10.86%
6.
-
Weighted mean
This is the mean which uses arbitrarily given weights
It is a useful measure especially where assessment is being done yet the conditions prevailing
are not the same. This is particularly true when assessment of students is being done given
that the subjects being taken have different levels of difficulties.
Examples
The following table shows that marks scored by a student doing section 3 and 4 of CPA
Subject
STAD
BF
FA2
LAW
QT
FA3
Scores (x)
65
63
62
80
69
55
Weight (w)
50
40
45
35
55
60
w = 285
Weighted mean
wx
3250
2520
2340
2800
3795
3300
wx = 18005
90
Ewx
Ew
=
18005
285
= 63.17%
Merits and demerits of the measures of central tendency
The arithmetic mean (a.m)

Merits
i.
ii.
iii.
iv.
It utilizes all the observations given

It is a very useful statistic in terms of applications. It has several applications in
business management e.g. hypothesis testing, quality control e.t.c.
It is the best representative of a given set of data if such data was obtained from a
normal population
The a.m. can be determined accurately using mathematical formulas
Demerits of the a.m.
i.
ii.
If the data is not drawn from a normal population, then the a.m. may give a wrong
impression about the population
In some situations, the a.m. may give unrealistic values especially when dealing with
discrete variables e.g. when working out the average no. of children in a no. of
families. It may be found that the average is 4.4 which is unrealistic in human beings
The mode
Merits
i.
ii.
iii.
iv.
It can be determined from incomplete data provided the observations with the
highest frequency are already known
The mode has several applications in business
The mode can be easily defined
It can be determined easily from a graph
Demerits
i.
ii.
iii.
If the data is quite large and ungrouped, determination of the mode can be quite
cumbersome
Use of the formula to calculate the mode is unfamiliar to most business people
The mode may sometimes be non existent or there may be two modes for a given
set of data. In such a case therefore a single mode may not exist
The median
Merits
i.
ii.
It shows the centre of a given set of data

Knowledge of the determination of the median may be extended to determine the
quartiles
90
Lesson Three
iii.
iv.
v.
91
The median can easily be defined

It can be obtained easily from the cumulative frequency curve
It can be used in determining the degrees of skew ness (see later)
Demerits
i.
ii.
iii.
In some situations where the no. of observations is even, the value of the median
obtained is usually imaginary
The computation of the median using the formulas is not well understood by most
businessmen
In business environment the median has got very few applications
The geometric mean

Merits
i.
ii.
It makes use of all the values given (except when x = 0 or negative)

It is the best measure for industrial growth rates
Demerits
i.
ii.
The determination of the GM by using logarithms is not familiar process to all

those expected to use it e.g managers
If the data contains zeros or ve values, the GM ceases to exist
The harmonic mean and weighted mean

Merits same as the arithmetic mean
Demerits same as the arithmetic mean
3.3
-
Measures of Dispersion
The measures of dispersion are very useful in statistical work because they indicate
whether the rest of the data are scattered around the mean or away from the mean.
If the data is approximately dispersed around the mean then the measure of dispersion
obtained will be small therefore indicating that the mean is a good representative of the
sample data. But on the other hand, if the figures are not closely located to the mean
then the measures of dispersion obtained will be relatively big indicating that the mean
does not represent the data sufficiently
The commonly used measures of dispersion are
a) The range
b) The absolute mean deviation
c) The standard deviation
d) The semi interquartile and quartile deviation
e) The 10th and 90th percentile range
f) Variance
a) The range
-
The range is defined as the difference between the highest and the smallest values in a
frequency distribution. This measure is not very efficient because it utilizes only 2 values
in a given frequency distribution. However the smaller the value of the range, the less
dispersed the observations are from the arithmetic mean and vice versa
The range is not commonly used in business management because 2 sets of data may
yield the same range but end up having different interpretations regarding the degree of
dispersion
92
b) The absolute mean deviation

-
This is a useful measure of dispersion because it makes use of all the values given see
the following examples
Example 1
In a given exam the scores for 10 students were as follows
Student
Mark (x)
xx
A
B
C
D
E
F
G
H
I
J
Total
60
45
75
70
65
40
69
64
50
80
618
1.8
16.8
13.2
8.2
3.2
21.8
7.2
2.2
11.8
18.2
104.4
Required
Determine the absolute mean deviation
Mean, x =
618
10
= 61.8
X-X
Therefore
AMD =
104.4
10
= 10.44
Example 2
The following data was obtained from a given financial institution. The data refers to the loans
given out in 1996 to several firms
Firms (f)
3
4
1
5
6
f = 19
Amount of loan
per firm (x)
20000
60000
15000
12000
14000
fx
xx
x x .f
60000
240000
15000
60000
84000
fx = 459000
4157.9
35842.1
9157.9
12157.9
10157.9
12473.70
143368.40
9157.9
60789.50
60947.40
286736.90
Required
Calculate the mean deviation for the amount of items given
92
Lesson Three
X =
93
fx 459, 000
=
= 24157.9
19
f
AMD =
X-X
f
286736.90
19
= Shs 15, 091.40

NB if the absolute mean deviation is relatively small it implies that the data is more compact and
therefore the arithmetic mean is a fair sample representative.
c) The standard deviation

-
This is one of the most accurate measures of dispersion. It has the following
advantages;
i. It utilizes all the values given
ii. It makes use of both negative and positive values if they occur
The standard deviation reflects an accurate impression of how much the sample
iii.
data varies from the mean. This is because its suitability can also be tested using
other statistical methods
Example
A sample comprises of the following observations; 14, 18, 17, 16, 25, 31
Determine the standard deviation of this sample
Observation.
Total
X=
121
6
( x x)
( x x)
14
18
17
16
25
31
121
-6.1
-2.1
-3.1
-4.1
4.9
10.9
37.21
4.41
9.61
16.81
24.01
118.81
210.56
= 20.1
standard deviation, =
xx
n
210.56
6
= 5.93
94

Alternative method
x
14
18
17
16
25
31
121
Total
2
x x
=
n
n
X2
196
324
289
256
625
961
2651
2651 121
6
6
= 5.93
Example 2
The following table shows the part-time rate per hour of a given no. of laborers in the month of
June 1997.
No. of labourers
(f)
7
6
2
1
8
11
35
Rate per hr (x) Shs

230
400
350
450
200
150
Total
fx
fx2
1610
2400
700
450
1600
1650
8410
370300
960000
245000
202500
320000
247500
2345300
Calculate the standard deviation from the above table showing how the hourly payment were
varying from the respective mean
standard deviation, =
fx
f
fx
-
f
2345300 8410
-
35
35
67008.6 577372
9271.4
= 96.29
Example 3 Grouped data

94
Lesson Three
95
In business statistical work we usually encounter a set of grouped data. In order to determine the
standard deviation from such data, we use any of the three following methods
i. The long method
ii. The shorter method
iii. The coded method
The above methods are used in the following examples
Example 3.1
The quality controller in a given firm had an accurate record of all the iron bars produced in may
1997. The following data shows those records
i.
Using long method
Bar lengths
(cm)
201 250
251 300
301 350
351 400
401 450
451 500
501 - 550
No. of bars(f)
Class mid point

(x)
225.5
275.5
325.5
375.5
425.5
475.5
525.5
25
36
49
80
51
42
30
313
fx
fx2
5637.5
9918
15949.5
30040
21700.5
19971
15765
118981.50
1271256.25
2732409
5191562.25
11280020
9233562.75
9496210.50
8284507.50
47489526
Calculate the standard deviation of the lengths of the bars
standard deviation,
fx
f
=
fx
-
f
47489526 118981.50
-
313
313
= 84.99 cm
ii.
Using the shorter method
Bar lengths (cm)

201 250
251 300
301 350
351 400
401 450
451 500
501 - 550
Total
No. of
bars(f)
25
36
49
80
51
42
30
313
mid point (x)
x-A = d
fd
Fd2
225.5
275.5
325.5
375.5 (A)
425.5
475.5
525.5
-150
-100
-50
0
50
100
150
-3750
-3600
-2450
0
2550
4200
4500
1450
562500
360000
122500
0
127500
420000
675000
2267500
Calculate the standard deviation using the shorter method quagmire
96
Standard deviation,
2267500 1450
313
313
7244.40 21.50
7222.90
fd
fd
f
= 84.99 cm
iii.
Using coded method
Bar lengths
(cm)
201 250
251 300
301 350
351 400
401 450
451 500
501 - 550
(f)
mid point (x)
x-A = d
d/c = u
fu
fu2
25
36
49
80
51
42
30
313
225.5
275.5
325.5
375.5 (A)
425.5
475.5
525.5
-150
-100
-50
0
50
100
150
-3
-2
-1
0
1
2
3
-75
-72
-49
0
51
84
90
29
225
144
49
0
51
168
270
907
C = 50 where c is an arbitrary number, try picking a different figure say 45 the answer
should be the same.
Standard deviation using the coded method. This is the most preferable method among the three
methods
= c
fu
f
fu
-
f
907 29
= 50
313 313
= 50 1.6997
= 84.99
Variance
Square of the standard deviation is called variance.
96
Lesson Three
97
d) The semi interquartile range

-
This is a measure of dispersion which involves the use of quartile. A quartile is a

mark or a value which lies at the boundary of a division when any given set of data
is divided into four equal divisions
- Each of such divisions normally carries 25% of all the observations
- The semi interquartile range is a good measure of dispersion because it shows how
the rest of the data are generally spread around the mean
- The quartiles normally used are three namely;
i. The lower quartile (first quartile Q1) this usually binds the lower 25% of the
data
ii. The median (second quartile Q2)
iii. The upper quartile (third quartile Q3)
The semi-interquartile range,
SIR =
Q3 - Q1
2
Example 1
The weights of 15 parcels recorded at the GPO were as follows:
16.2, 17, 20, 25(Q1) 29, 32.2, 35.8, 36.8(Q2) 40, 41, 42, 44(Q3) 49, 52, 55 (in kgs)
Required
Determine the semi interquartile range for the above data
SIR =
Q3 Q1
2
44 - 25
2
19
2
= 8.5
Example 2 (Grouped Data)

The following table shows the levels of retirement benefits given to a group of workers in a
given establishment.
Retirement benefits
000
20 29
30 39
40 49
50 59
60 69
70 79
80 89
No of retirees (f)
UCB
cf
50
69
70
90
52
40
11
29.5
39.5
49.5
59.5
69.5
79.5
89.5
50
119
189
279
331
371
382
Required
i. Determine the semi interquartile range for the above data
ii. Determine the minimum value for the top ten per cent.(10%)
iii. Determine the maximum value for the lower 40% of the retirees
98
Solution
The lower quartile (Q1) lies on position
N +1
382 + 1
4
= 95.75
the value of Q1
= 29.5 +
(95.75 - 50)
69
x 10
= 29.5 + 6.63
= 36.13
The upper quartile (Q3) lies on position
N + 1
= 3
4
382 + 1
=3
= 287.25
the value of Q3
= 59.5 +
( 287.25 - 279 )
52
10
= 61.08
The semi interquartile range =
Q3 - Q1
2
61.08 - 36.13
=
2
= 12.475
= 12,475
ii. The top 10% is equivalent to the lower 90% of the retirees
The position corresponding to the lower 90%
90
(n + 1) = 0.9 (382 + 1)
100
= 0.9 x 383
= 344.7
98
Lesson Three
99
the benefits (value) corresponding to the minimum value for top 10%
= 69.5 +
( 344.7 - 331)
40
x 10
= 72.925
= 72925
iii. The lower 40% corresponds to position
=
40
100
(382 + 1)
= 153.20
retirement benefits corresponding to its position

= 39.5 +
(153.2 - 119 )
70
x 10
= 39.5 + 4.88
= 44.38
= 44380
e. The 10th 90th percentile range

This is a measure of dispersion which uses percentile. A percentile is a value which separates one
division from the other when a given data is divided into 100 equal divisions.
This measure of dispersion is very important when calculating the co-efficient of skewness (see
later)
Example
Using the above data for retirees calculate the 10th - 90th percentile. The tenth percentile 10th
percentile lies on position
10
(382 + 1) = 0.1 x 383
100
= 38.3
the value corresponding to the tenth percentile
= 19.5 +
(38.3 x 10)
50
= 19.5 + 7.66
= 27.16
The 90th percentile lies on position
100
90
(382 + 1) = 0.9 x 383
100
= 344.7
the value corresponding to the 90th percentile

= 69.5 +
( 344.7 - 331) x 10
40
= 69.5 + 3.425
= 72.925
the required value of the 10th 90th percentile = 72.925 27.16 = 45.765
Relative measures of dispersion

Definition:
A relative measure of dispersion is a statistical value which may be used to compare variations in
2 or more samples.
The measures of dispersion are usually expressed as decimals or percentages and usually they do
not have any other units
Example
The average distance covered by vehicles in a motor rally may be given as 2000 km with a
standard deviation of 5 km.
In another competition set of vehicles covered 3000 km with a standard deviation of 10 kms
NB: The 2 standard deviations given above are referred to as absolute measures of dispersion.
These are actual deviations of the measurements from their respective mean
However, these are not very useful when comparing dispersions among samples.
Therefore the following measures of dispersion are usually employed in order to assess the
degree of dispersion.
i.
Coefficient of mean deviation
=
Mean deviation
mean
Coefficient of quartile deviation
ii.
1 (Q - Q )
3
1
= 2
Q2
Where Q1 = first quartile

Q3 = third quartile
Coefficient of standard deviation
iii.
=
Standard deviation
mean
100
Lesson Three
iv.
101
Coefficient of variation
=
standard deviation
mean
100
Example (see information above)

First group of cars: mean = 2000 kms
Standard deviation = 5 kms
C.O.V
=5
x 100
2000
= 0.25%
Second group of cars: mean = 3000 kms
Standard deviation = 10kms
C.O.V
= 10 x 100
3000
= 0.33%
Conclusion
Since the coefficient of variation is greater in the 2nd group, than in the first group we may
conclude that the distances covered in the 1st group are much closer to the mean that in the 2nd
group.
Example 2
In a given farm located in the UK the average salary of the employees is 3500 with a standard
deviation of 150
The same firm has a local branch in Kenya in which the average salaries are Kshs 8500 with a
standard deviation of Kshs.800
Determine the coefficient of variation in the 2 firms and briefly comment on the degree of
dispersion of the salaries in the 2 firms.
First firm in the UK
C.O.V = 150 x 100
3500
= 4.29%
Second firm in Kenya
C.O.V = 800 x 100
8500
= 9.4%
102
Conclusively, since 4.29% < 9.4% then the salaries offered by the firm in UK are much closer to
the mean given them in the case to the local branch in Kenya
COMBINED MEAN AND STANDARD DEVIATION

Sometimes we may need to combine 2 or more samples say A and B. It is therefore essential to
know the new mean and the new standard deviation of the combination of the samples.
Combined mean
Let m be the combined mean
Let x1 be the mean of first sample
Let x2 be the mean of the second sample
Let n1 be the size of the 1st sample
Let n2 be the size of the 2nd sample
Let s1 be the standard deviation of the 1st sample
Let s2 be the standard deviation of the 2nd sample
combined mean =
n1 x1 + n2 x2
n1 + n2
n1s12 + n1 ( m x1 ) + n2 s22 + n2 ( m x2 )
2
combined standard deviation =
n1 + n2
Example
A sample of 40 electric batteries gives a mean life span of 600 hrs with a standard deviation of
20 hours.
Another sample of 50 electric batteries gives a mean lifespan of 520 hours with a standard
deviation of 30 hours.
If these two samples were combined and used in a given project simultaneously, determine the
combined new mean for the larger sample and hence determine the combined or pulled standard
deviation.
Size
40(n1)
50 (n1)
x
600 hrs(x1)
520 hrs (x2)
=
Combined mean
s
20hrs (s1)
30 hrs (s2)
40 ( 600 ) + 50 ( 520 )
40 + 50
50, 000
90
= 555.56
Combined standard deviation
40(202 ) + 40(555.56 - 660) 2 + 50(30) 2 + 50( 555.56 - 520) 2

40 + 50
1600 + 78996.54 + 45000 + 63225.68

90
= 47.52 hrs
102
Lesson Three
103
SKEWNESS
- This is a concept which is commonly used in statistical decision making. It refers to the
degree in which a given frequency curve is deviating away from the normal distribution
- There are 2 types of skew ness namely
Positive skew ness
i.
ii.
Negative skew ness
1. Positive Skewness
- This is the tendency of a given frequency curve leaning towards the left. In a
positively skewed distribution, the long tail extended to the right.
In this distribution one should note the following
The mean is usually bigger than the mode and median
i.
The median always occurs between the mode and mean
ii.
There are more observations below the mean than above the mean
iii.
This frequency distribution as represented in the skewed distribution curve is characteristic of
the age distributions in the developing countries
frequency
Positively skewed
frequency curve
frequency
Negatively skewed
frequency curve
Mode
Mean
Median
Long tail
Mean
Median
Mode
Normal distribution
2. Negative Skewness
This is an asymmetrical curve in which the long tail extends to the left
NB: This frequency curve for the age distribution is characteristic of the age distribution in
developed countries
- The mode is usually bigger than the mean and median
- The median usually occurs in between the mean and mode
- The no. of observations above the mean are usually more than those below the
mean (see the shaded region)
MEASURES OF SKEWNESS
- These are numerical values which assist in evaluating the degree of deviation of a
frequency distribution from the normal distribution.
- Following are the commonly used measures of skew ness.
1. Coefficient Skewness
( mean - median )
= 3
Standard deviation
2. Coefficient of skewness
104
mean - mode
Standard deviation
NB: These 2 coefficients above are also known as Pearsonian measures of skewness.
3. Quartile Coefficient of skewness
Q3 + Q1 - 2Q2
=
Q3 + Q1
Where Q1 = 1st quartile
Q2 = 2nd quartile
Q3 = 3rd quartile
NB: The Pearsonian coefficients of skewness usually range between ve 3 and +ve 3. These are
extreme value i.e. +ve 3 and ve 3 which therefore indicate that a given frequency is negatively
skewed and the amount of skewness is quite high.
Similarly if the coefficient of skewness is +ve it can be concluded that the amount of skew ness
of deviation from the normal distribution is quite high and also the degree of frequency
distribution is positively skewed.
Example
The following information was obtained from an NGO which was giving small loans to some
small scale business enterprises in 1996. the loans are in the form of thousands of Kshs.
Loans
46 50
51 55
56 60
61 65
66 70
71 75
76 80
81 85
86 90
91 95
Total
Units (f) Midpoints(x)

32
48
62
53
97
58
120
63
(A)
92
68
83
73
52
78
40
83
21
88
11
93
610
x-a=d
-15
-10
-5
0
5
10
15
20
25
30
d/c= u
-3
-2
-1
0
+1
+2
+3
+4
+5
+6
fu
-96
-124
-97
0
92
166
156
160
105
66
428
Fu2
288
248
97
0
92
332
468
640
525
396
3086
UCB
50.5
55.5
60.5
0
70.5
75.5
80.5
85.5
90.5
95.5
cf
32
94
191
0
403
486
538
57.8
599
610
Required
Using the Pearsonian measure of skew ness, calculate the coefficients of skew ness and hence
comment briefly on the nature of the distribution of the loans.
c ( fu )
Arithmetic mean = Assumed mean +
f
= 63 +
( 428 5 )
610
= 66.51
It is very important to note that the method of obtaining arithmetic mean (or any other statistic) by minusing
assumed mean (A) from X and then deviding by c can be abit confusing, if this is the case then just use the
straight forward method of:
Arithmetic mean =
f .x
f
where x is the midpoint, the answers are the same.
104
Lesson Three
105
The standard deviation = c
fu
f
fu
-
f
3086 428
-
610 610
=5
= 10.68
n +1
The Position of the median lies m =
2
=
= 60.5 +
= 60.5 +
( 305.5 - 191)
120
(114.4 )
120
610 + 1
2
= 305.5
Median = 65.27
Therefore the Pearsonian coefficient
= 3
( 66.51- 65.27 )
10.68
= 0.348
Comment
The coefficient of skewness obtained suggests that the frequency distribution of the loans given
was positively skewed
This is because the coefficient itself is positive. But the skewness is not very high implying the
degree of deviation of the frequency distribution from the normal distribution is small
Example 2
Using the above data calculate the quartile coefficient of skewness
Q3 + Q1 - 2Q2
Quartile coefficient of skewness =
Q3 + Q1
The position of Q1 lies on
actual value Q1 =55. 5 +
The position of Q3 lies on
610 + 1
4
= 152.75
(152.75 - 94 ) 5 = 58.53
97
= 3
( 610 +1) = 458.25

4
106
actual value Q3 =70.55 +
Q2 position:
i.e. 2
( 458.25 - 403) 5 = 73.83 5

83
( 610 +1)
4
Actual Q2 value = 60.5 +
= 305.5
( 305.5 -191) 5 = 65.27

120
The required coefficient of skew ness

73.83 + 58.53 2 ( 65.27 )
= 0.013
=
73.83 + 58.53
Conclusion
Same as above when the Pearsonian coefficient was used
KURTOSIS
- This is a concept, which refers to the degree of peaked ness of a given frequency
distribution. The degree is normally measured with reference to normal distribution.
- The concept of kurtosis is very useful in decision making processes i.e. if is a
frequency distribution happens to have either a higher peak or a lower peak, then it
should not be used to make statistical inferences.
- Generally there are 3 types of kurtosis namely;Leptokurtic
i.
Mesokurtic
ii.
Platykurtic
iii.
Leptokurtic
A frequency distribution which is lepkurtic has generally a higher peak than that
a)
of the normal distribution. The coefficient of kurtosis when determined will be
found to be more than 3. thus frequency distributions with a value of more
than 3 are definitely leptokurtic
b)
Some frequency distributions when plotted may produce a curve similar to that
of the normal distribution. Such frequency distributions are referred to as
mesokurtic. The degree of kurtosis is usually equal to 3
When the frequency curve contacted produces a peak which is lower that that
c)
of a normal distribution when such a curve is said to be platykurtic. The
coefficient of such is usually less than 3
- It is necessary to calculate the numerical measure of kurtosis. The commonly used
measure of kurtosis is the percentile coefficient of kurtosis. This coefficient is
normally determined using the following equation
( Q3 - Q1)
Percentile measure of kurtosis, K (Kappa) = 12
P90 - P10
Example
Refer to the table above for loans to small business firms/units
Required
Calculate the percentile coefficient of Kurtosis
90
P90 =
( n +1) = 0.9 ( 610 +1)
100
= 0.9 (611)
106
Lesson Three
107
= 549.9
The actual loan for a firm in this position
( 549.9 - 538 ) x 5 = 81.99
(549.9) = 80.5 +
40
P10 =
10
(n + 1) = 0.1 (611) = 61.1
100
The actual loan value given to the firm on this position is

50.5 +
( 61.1 32 )
62
x 5 = 52.85
= 0.9 (611)
= 549.9
percentile measure of kurtosis

K(Kappa)
( Q3 - Q1)
P90 - P10
( 73.83 - 58.53)
81.99 - 52.85
= 0.26
Since 0.26 < 3, it can be concluded that the frequency distribution exhibited by the distribution
of loans is platykurtic
Kurtosis is also measured by moment statistics, which utilize the exact value of each
observation.
i. M1 the first moment = M1 =
M2 =
= Mean M1 or M1
X
M3 =
X
M4 =
3. M2 second moment about the mean M2 or f2

M2 = M2 M12
4. M3 third moment about the mean M3 (a measure of the absolute skew ness)
M3 = M3 3M2M1 + 2M13
5. M4 fourth moment about the mean M4 (a measure of the absolute Kurtosis)
M4 = M4 4M3M1 + 6M2M12 + 3M14
An alternative formula
108
( x m)
M4 =
f
Moment coefficient of Kurtosis
Where m is mean
M4
S4
Example
Find the moment coefficient of the following distribution
x
f
12
1
14
4
16
6
18
10
20
7
22
2
X
12
14
16
18
20
22
f
1
4
6
10
7
2
30
xf
12
56
96
180
140
44
528
(x-m)
-5.6
-3.6
-1.6
.4
2.4
4.4
(x-m)2
31.36
12.96
2.56
0.16
5.76
19.36
(x-m)2f
31.36
51.84
15.36
1.60
40.32
38.72
179.20
(x-m)4f
983.45
671.85
39.32
0.256
232.24
749.62
2,676.74
528
= 17.6
30
179.20
30
= 35.677
( x m)
M4 =
f
= 5.973
2, 676.74
30
Moment coefficient of Kurtosis =
89.22
35.677
= 89.22
= 2.5
Note Coefficient of kurtosis can also be found using the method of assumed mean.
3.4
Indices
An index number is an attempt to summarize a whole mass of data into one figure. The single
figure shows how one year differs from another year.
It is a statistical devise used to measure the change in the level of prices, wages output and other
variables at given times, relative to their level at an earlier time which is taken as the base for
comparison purposes
108
Lesson Three
109
Pn
100 (an unweighted price index)
Po
A simple price index =
A simple quantity index =
Qn
100 (an unweighted quantity index)
Qo
Where pn is the price of a commodity in the current year (the year for which the price index to
be calculated)
Where po is the price of the same commodity in the base year (the year for comparison
purposes)
Similarly Qn and Qo are defined in the same way
AGGREGATE PRICE INDEX NUMBERS AND QUANTITY INDEX NUMBERS

PRICE INDEX
QUANTITY INDEX
p q
P q
p q
Pq
LASPEYRES INDEX
PAASCHES INDEX
n n
q p
q p
q q
q p
100
p q
P q
n
n n
100
o n
Value index =
100
100
100
MODIFIED FORM OF THE LASPEYRES PRICE INDEX NUMBER

Laspeyres Price index
( ) w
=
w
pn
po
100
Where w0 are the proportions of the total expected in the basic period. This formula is
frequently used to calculate retail price index.
CHANGING THE BASE OF THE INDEX

For comparison purposes if two series have different base years, it is difficult to compare them
directly. In such cases, it is necessary to change the base year of one of the series (or both) so
that both have the same base.
It is also necessary to keep the index relevant to current conditions hence the need to change the
base from time to time.
Example;
Year
Price index
1985
100
1986
104
1987
108
1988
109
1989
112
1990
120
Suppose we wish to change the base year to 1989

We recalculate each index by expressing it as a percentage of 1989
1991
125
1992
140
110
1985
Previous index
100
1986
104
1987
108
1988
109
1989 (new base year)
112
1990
120
1991
125
1992
140
Recalculated index
100
100 = 89.3
112
104
100 = 92.9
112
108
100 = 96.4
112
109
100 = 97.3
112
112
100 = 100
112
120
100 = 107.1
112
125
100 = 111.6
112
140
100 = 125.0
112
When changing the base year, it is advisable to update the weights used in the base year.
CHAIN BASED INDEX NUMBERS

A chain based index is one where the index is calculated every year using the previous year as the
base year. This type of index measures rate of change from year to year.
This method is suitable where weights are changing rapidly and items are constantly being
brought into the index and unwanted items taken out. It can be a price or quantity index
Previous index
1985
1986
100
104
1987
108
1988
109
1989
112
1990
120
1991
125
1992
140
Recalculated chainbased
index
100
104
100 = 104
100
108
100 = 103.8
104
109
100 = 100.9
108
112
100 = 102.8
109
120
100 = 107.1
112
125
100 = 104.2
120
140
100 = 112
120
fixedbased index
100(1985 base year
104
100 = 104
100
108
100 = 108
100
109
100 = 109
100
112
100 = 112
100
120
100 = 120
100
125
100 = 125
100
140
100 = 140
100
110
Lesson Three
111
The Fishers index

The Fishers index acts as a compromise between Laspeyres index and Paasche index. It is
calculated as a geometric mean of the two indexes.
Retail price index
It is weighted average of price relatives based upon an average household in the base year. The
items consumed are divided into groups such as food, housing, transport, alcoholic drinks,
footwear, fuel, light, water, household goods, services e.t.c. each item included in the index is
given a weighting and a price relative to the base is calculated. Modified form of laspeyres price
index formula is used as a weighted arithmetic mean of price relatives.
I.e. Retail Price index
( )W
=
W
pn
po
100
The index is used by the Government as a guide in determining the minimum wages, pension
rates unemployed benefits (in UK e.t.c). Trade unions use it as a basis for their wages claims.
Deflation
Indexes may be used to deflate time series so that comparisons between periods may be made in
real terms
It is a process of reducing a value measured in current period prices to its equivalent in the base
period prices. The deflated value is what would have been necessary to purchase the same
amount of goods as the present value can purchase in the current period
Deflation Factor =
p q
p q
n
100
Deflation of a time series

Year
Average monthly earnings (shs)
1
5,000
2
5,500
Retail index
100
120
6,000
140
6,500
170
7,200
200
Real earnings
5000 = 5000
100
= 4,583.3
5,500
120
100
6,000
= 4,285.7
140
100
= 3,823.5
6,500
170
100
= 3,600.0
7,200
200
The technique of index number construction

When preparing index numbers it is important to define
a) The exact purpose of the index
b) How the items are to be selected
c) The choice of the weights
d) The choice of the base
e) The type of average to be used
The base year should be as close to the normal trend as possible. The best methods should be
used for collection of data. The items should be selected in such a way that they are a fair
representation of all the relevant items.
Due consideration should be given to the weighting of all items selected
112
The index of industrial production

It is a quantity index compiled by the government. It measures changes in the volume of
production in major industries. The index is a good indication of the state of national economy.
It covers the following major industries in the UK
Mining and quarrying
i.
Manufacturing such as food, drinks and tobacco, chemicals, metal manufacture,
ii.
engineering e.t.c
iii.
Textile
Construction
iv.
Gas electricity, water e.t.c
v.
It excludes agriculture, fishing, trade, transport, finance and other such industries.
Each industries order is given a weighting. The weighting is based on average monthly
production in each industry in a fixed base year. It gives each item its relative importance
amongst all other items and thus gives a better estimate of the index for comparison purposes.
The Geometric Index (Industrial Share index)
This index is an index of 30 selected top industrial companies. It is calculated by taking an
unweighted geometric mean of the price relatives of the selected shares.
Example
The share prices of ordinary shares of four companies on 1st January 1990 and 1st January 1991
were as follows.
Share
Company A
Company B
Company C
Company D
Price on 1.1.1990
Shs 10
Shs 12
Shs 20
Shs 5
Price on 1.1.1991
Shs 12
Shs 15
Shs 25
Shs 6
Using an unweighted geometric index, calculate the index of share prices at 1.1.1991 if 1.1.1990
is the base date, index 100
Solution
1
1
12 15 25 6 4 27000 4
4
2.25
=
=
(
)
10 12 20 5 12000
= 1.225
percentage increase = 22.5%
index = 122.5
Inflation
The inflation rate for a given period can be calculated using the following formula;
Current retail price index
100
Inflation =
Retail price index in the base year
Marshal Hedge Worth Index
Marshal Hedge worth index =
p (p
p (q
n
+ qn )
+ qn )
100
112
Lesson Three
113
Tests For An Ideal Index Number

1. Factor Reversal Test
This test indicates that when the price index is multiplied with a quantity index i.e. factors are
reversed), it should result in the value index.
2. The time reversal test
If we reverse the time subscripts of a price or quantity index, the result should be reciprocal of
the original index.
114

LESSON 3 REINFORCING QUESTIONS
QUESTION ONE
a)
b)
c)
d)
Distinguish between discrete and continuous data.

What is dispersion and what is the formula for the standard deviation?
What is the measure of relative dispersion?
Draw diagrams showing positive and negative skewness
QUESTION TWO
The managers of an import agency are investigating the length of time that customers take to
pay their invoices, the normal terms for which are 30 days net. They have checked the payment
record of 100 customers chosen at random and have compiled the following table:
Payment in
5 to 9 days
10 to 14 days
15 to 19 days
20 to 24 days
25 to 29 days
30 to 34 days
35 to 39 days
40 to 44 days
Number of customers
4
10
17
20
22
16
8
3
Required:
a) Calculate the arithmetic mean.
b) Calculate the standard deviation
c) Construct a histogram and insert the modal value.
d) Estimate the probability that an unpaid invoice chosen at random will be between 30 and
39 days old.
QUESTION THREE
The price of the ordinary 25p shares of Manco PLC quoted on the stock exchange, at the close
of the business on successive Fridays is tabulated below
126
125
128
124
127
120
127
126
127
122
122
113
117
114
106
105
112
114
111
121
129
130
120
116
116
119
122
123
131
135
131
134
127
128
142
138
136
140
137
130
Required
a) Group the above date into eight classes.
b) Calculate cumulative frequency, the median value, quartile values and the
semi-quartile range.
c) Calculate the mean and standard deviation of your frequency distribution.
d) Compare and contrast the values that you have obtained for:
The median and mean
i)
The semi-interquartile range and the standard deviation
ii)
(Total:
(4 marks)
(4 marks)
(7 marks)
(5 marks)
20 marks)
114
Lesson Three
115
QUESTION FOUR
Define the coefficient of variation.
The following table gives profits (in ten thousands of shillings) of two supermarkets over a
duration of one year.
Month
January
February
March
April
May
June
July
August
September
October
November
December
i)
ii)
Supermarket A
65
48
15
28
41
59
41
10
24
56
92
120
Supermarket B
28
33
20
23
69
45
53
15
35
57
99
136
Required:
Compute the coefficient of variation for each supermarket.
Indicate for which supermarket the variability of profits is relatively greater.
QUESTION FIVE
Prodco PLC manufactures an item of domestic equipment which requires a number of
components which have varied as various modifications of the model have been used. The
following table shows the number of components required together with the price over the last
three years of production.
COMPONENT
A
B
C
D
Prices
3.63
2.11
10.03
4.01
1981
Quantity
3
4
1
7
Prices
4.00
3.10
10.36
5.23
1982
Quantity
2
5
1
6
Prices
4.49
3.26
12.05
5.21
1983
Quantity
2
6
1
5
Required:
a) Establish the base weighted price indices for 1982 and 1983 based on
1981 for the item of equipment.
b) Establish the current weighted price indices for 1982 and 1983 based on
1981 for the item of equipment.
c) Using the results of (a) and (b) as illustrations, compare and contrast
Laspeyres and Paasche price index numbers.
(Total:
(8 marks)
(8 marks)
(4 marks)
20 marks)
116
QUESTION SIX
a) A company manufacturing a product known as 257 uses five components in its assembly.
The quantities and prices of the components used to produce a unit of K257 in 1982, 1983
and 1984 are tabulated as follows:
COMPONENT
A
B
C
D
E
i)
ii)
iii)
1982
Quantity Prices
10
3.12
6
11.49
5
1.40
9
2.15
50
0.32
1983
Quantity Prices
12
3.17
7
11.58
8
1.35
9
2.14
53
0.32
1984
Quantity Prices
14
3.20
5
11.67
9
1.31
10
2.63
57
0.32
Required:
Calculate Laspyeres type price index number for the cost of one unit of K257
for 1983 and 1984 based on 1982.
(6 marks)
Calculate Paasche type price index numbers for the cost of one unit of K257
for 1983 and 1984 based on 1982.
(6 marks)
Compare and contrast the Laspeyre and Paasche price-index numbers you
have obtained in (i) and (ii)
(3 marks)
A number of employers manufacturing plastic components used in plumbing have formed

themselves into an association for the purpose of negotiating with the trade union for this
industrial sector.
The negotiations cover pay and contributions in this sector.
Required:
Explain the usefulness of an index of Industrial Production and an index of retail prices to both
sides in a series of pay negotiations.
(5 marks)
(Total: 20 marks)
QUESTION SEVEN
The data given below indicates the prices and production of some horticulatural products in
Central Territory:
Produce
Cabbages
Tomatoes
Onions
Spinach
Production
(1000 boxes)
1980
48,600
22,000
47,040
43,110
1990
62,000
37,440
61,430
55,720
Price per box (Shs)

1980
100
220
180
130
1990
150
310
200
170
116
Lesson Three
117
Required:
Calculate the increase or decrease in prices from 1980 on the basis of the following indices:
a)
b)
c)
d)
e)
Mean relatives
Laspeyres index
Paasche index
Marshall Hedgeworth index
Fishers index.
118
Measures of Relationships and Forecasting
LESSON FOUR
Correlation
Regression analysis
Multiple Linear Regression
Time series analysis and forecasting
118
Lesson Four
4.1
119
Correlation and Regression
Correlation
This is an important statistical concept which refers to interrelationship or association between
variables.
The purpose of studying correlation is for one to be able to establish a relationship, plan and
control the inputs (independent variables) and the output (dependent variables)
In business one may be interested to establish whether there exists a relationship between the
i.
Amount of fertilizer applied on a given farm and the resulting harvest
Amount of experience one has and the corresponding performance
ii.
Amount of money spent on advertisement and the expected incomes after sale of
iii.
the goods/service
There are two methods that measure the degree of correlation between two variables these are
denoted by R and r.
(a) Coefficient of correlation denoted by r, this provides a measure of the strength of
association between two variables one the dependent variable the other the independent
variable r can range between +1 and 1 for perfect positive correlation and perfect
negative correlation respectively with zero indicating no relation i.e. for perfect positive
correlation y increase linearly with x increament.
(b) Rank correlation coefficient denoted by R is used to measure association between two
sets of ranked or ordered data. R can also vary from +1, perfect positive rank
correlation and -1 perfect negative rank correlation where O or any number near zero
representing no correlation.
SCATTER GRAPHS
- A scatter graph is a graph which comprises of points which have been plotted but are
not joined by line segments
- The pattern of the points will definitely reveal the types of relationship existing between
variables
- The following sketch graphs will greatly assist in the interpretation of scatter graphs.
120
Perfect positive correlation
y
Dependant variable
x
x
x
x
x
x
x
x
Independent variable
NB: For the above pattern, it is referred to as perfect because the points may easily be
represented by a single line graph e.g. when measuring relationship between volumes of sales
and profits in a company, the more the company sales the higher the profits.
Perfect negative correlation
y
Quantity sold
x
x
X
x
x
x
x
x
x
10
20
Price
This example considers volume of sale in relation to the price, the cheaper the goods the bigger
the sale.
120
Lesson Four
121
High positive correlation
y
Dependant variable
xx
xx
x
x
xx
xx
xx
xx
x
xxx
x
x
independent variable
High negative correlation
y
quantity sold
x
x
xx
x
xx
x
x
x
x
xx
x
price
122
No correlation
600
x
x
400
x
x
x
200
x
x
20
30
40
50
0
10
h) Spurious Correlations
- in some rare situations when plotting the data for x and y we may have a group showing
either positive correlation or ve correlation but when you analyze the data for x and y
in normal life there may be no convincing evidence that there is such a relationship.
This implies therefore that the relationship only exists in theory and hence it is referred
to as spurious or non sense e.g. when high passrates of student show high relation with
increased accidents.
Correlation coefficient
- These are numerical measures of the correlations existing between the dependent and
the independent variables
- These are better measures of correlation than scatter graphs (diagrams)
- The range for correlation coefficients lies between +ve 1 and ve 1. A correlation
coefficient of +1 implies that there is perfect positive correlation. A value of ve shows
that there is perfect negative correlation. A value of 0 implies no correlation at all
- The following chart will be found useful in interpreting correlation coefficients
122
Lesson Four
123
__ 1.0 }
}
Perfect +ve correlation

High positive correlation
__ 0.5 }
}
__0
Low positive correlation
}
}
Low negative correlation
__-0.5}
}
__-1.0}
High negative correlation

Perfect ve correlation
There are usually two types of correlation coefficients normally used namely;-
Product Moment Coefficient (r)

It gives an indication of the strength of the linear relationship between two variables.
r=
n xy x y
n x 2 ( x ) n y 2 ( y )
2
note that this formula can be rearranged to have different outlooks but the resultant is always
the same.
Example
The following data was observed and it is required to establish if there exists a relationship
between the two.
X
15
24
25
30
35
40
45
65
70
75
Y
60
45
50
35
42
46
28
20
22
15
124
Solution
Compute the product moment coefficient of correlation (r)
X
Y
X2
Y2
15
60
225
3,600
24
45
576
2,025
25
50
625
2,500
30
35
900
1,225
35
42
1,225
1,764
40
46
1,600
2,116
45
28
2,025
784
65
20
4,225
400
70
22
4,900
484
75
15
5,625
225
X = 424
r=
r=
Y = 363
= 21,926
XY
900
1,080
1,250
1,050
1,470
1,840
1,260
1,300
1,540
1,125
2
= 15,123
XY = 12,815
n xy x y
n x 2 ( x ) n y 2 ( y )
2
10 12,815 424 363
(10 21, 926 424 ) (10 15,123 363 )

2
25, 762
( 39, 484 ) (19, 461)
= 0.93
The correlation coefficient thus indicates a strong negative linear association between the two
variables.
Interpretation of r Problems in interpreting r values

NOTE:
A high value of r (+0.9 or 0.9) only shows a strong association between the two variables
but doesnt imply that there is a causal relationship i.e. change in one variable causes change
in the other it is possible to find two variables which produce a high calculated r yet they
dont have a causal relationship. This is known as spurious or nonsense correlation e.g. high
pass rates in QT in Kenya and increased inflation in Asian countries.
Also note that a low correlation coefficient doesnt imply lack of relation between variables
but lack of linear relationship between the variables i.e. there could exist a curvilinear
relation.
A further problem in interpretation arises from the fact that the r value here measures the
relationship between a single independent variable and dependent variable, where as a
particular variable may be dependent on several independent variables (e.g. crop yield may
be dependent on fertilizer used, soil exhaustion, soil acidity level, season of the year, type of
seed etc.) in which case multiple correlation should be used instead.
The Rank Correlation Coefficient (R)
Also known as the spearman rank correlation coefficient, its purpose is to establish whether
there is any form of association between two variables where the variables arranged in a ranked
form.
124
Lesson Four
R=1-
125
6 d 2
n ( n 2 1)
Where d = difference between the pairs of ranked values.
n = numbers of pairs of rankings
Example
A group of 8 accountancy students are tested in Quantitative Techniques and Law II. Their
rankings in the two tests were.
Student
Q. T. ranking
Law II ranking
d
d2
A
2
3
-1
1
B
7
6
1
1
C
6
4
2
4
D
1
2
-1
1
E
4
5
-1
1
F
3
1
2
4
G
5
8
-3
9
H
8
7
1
1
= 22
d = Q. T. ranking Law II ranking

R=1-
6 d 2
n ( n 1)
2
= 1
6 22
8 ( 82 1)
= 0.74
Thus we conclude that there is a reasonable agreement between students performances in the
two types of tests.
NOTE: in this example, if we are given the actual marks then we find r. R varies
between +1 and -1.
Tied Rankings
A slight adjustment to the formula is made if some students tie and have the same ranking the
adjustment is
t3 t
where t = number of tied rankings the adjusted formula becomes
12
6
R=1-
( d + )
n ( n 1)
2
t 3 t
12
Example
Assume that in our previous example student E & F achieved equal marks in Q. T. and were
given joint 3rd place.
126
Solution
Student
A
B
C
D
E
F
G
H
Q. T. ranking
2
7
6
1
3
3
5
8
6
R = 1-
( d + )
n ( n 1)
t 3 t
12
Law II ranking
3
6
4
2
5
1
8
7
1-
d
-1
1
2
-1
-1
2
-3
1
6 26 1 2 + 212 2
8 ( 8 1)
d2
1
1
4
1
2
6
9
1
= 26 1 2
since t = 2
= 0.68
NOTE:It is conventional to show the shared rankings as above, i.e. E, & F take up the 3rd and
4th rank which are shared between the two as 3 each.
ii. Coefficient of Determination
This refers to the ratio of the explained variation to the total variation and is used to measure the
strength of the linear relationship. The stronger the linear relationship the closer the ratio will be
to one.
Coefficient determination =
Explained variation
Total variation
Example (Rank Correlation Coefficient)

In a beauty competition 2 assessors were asked to rank the 10 contestants using the professional
assessment skills. The results obtained were given as shown in the table below
Contestants
A
B
C
D
E
F
G
H
J
K
1st assessor
6
1
3
7
8
2
4
5
10
9
2nd assessor
5
3
4
6
7
1
8
2
9
10
126
Lesson Four
127
REQUIRED
Calculate the rank correlation coefficient and hence comment briefly on the value obtained
d
d2
A
B
C
D
E
F
G
H
J
K
6
1
3
7
8
2
4
5
10
9
5
3
4
6
7
1
8
2
9
10
1
-2
-1
1
1
1
-4
3
+1
-1
1
4
1
1
1
1
16
9
1
1
d2 = 36
The rank correlation coefficient R
R=1-
6 d 2
n ( n 2 1)
=1-
6 36
10 (102 1)
=1-
216
990
= 1 0.22
= 0.78
Comment: since the correlation is 0.78 it implies that there is high positive correlation between
the ranks awarded to the contestants. 0.78 > 0 and 0.78 > 0.5
Example
Contestant
A
B
C
D
E
F
G
H
1st
assessor
1
5 (5.5)
3
2
4
5 (5.5)
7
8
2nd assessor
2
3
4
1
5
6.5
6.5
8
-1
2.5
-1
1
-1
-1
-0.5
0
d2
1
6.25
1
1
1
1
0.25
0
d2 = 11.25
Required: Complete the rank correlation coefficient
R=
1-
6 d 2
n ( n 1)
=1
=1-
6 11.25
8 ( 63)
67.5
504
128

= 1 0.13
= 0.87
This implies high positive correlation
Example (Rank Correlation Coefficient)

Sometimes numerical data which refers to the quantifiable variables may be given after which a
rank correlation coefficient may be worked out.
Is such a situation, the rank correlation coefficient will be determined after the given variables
have been converted into ranks. See the following example;
Candidates
P
Q
R
S
T
U
V
W
X
Y
Math
92
82
60
87
72
60
52
50
47
59
Rank correlation r =
r
1
3
5(5.5)
2
4
5(5.5)
8
9
10
7
1=1-
Accounts
67
88
58
80
69
77
58
60
32
54
r
5
1
7(7.5)
2
4
3
7(7.5)
6
10
9
d
-4
2
-2
0
0
-2.50
0.5
3
0
-2
d2
16
4
4
0
0
6.25
0.25
9
0
4
d2 = 43.5
6 d 2
n ( n 2 1)
6 43.5
261
=1
2
990
10 (10 1)
= 0.74 (High positive correlation between mathematics

marks and accounts)
Example
(Product moment correlation)
The following data was obtained during a social survey conducted in a given urban area
regarding the annual income of given families and the corresponding expenditures.
Family
A
B
C
D
E
F
G
H
J
K
Total
(x)Annual
income 000
420
380
520
610
400
320
280
410
380
300
4020
(y)Annual
expenditure 000
360
390
510
500
360
290
250
380
240
270
3550
xy
x2
Y2
151200 176400 129600

148200 144400 152100
265200 270400 260100
305000 372100 250000
144000 160000 129600
92800 102400
84100
70000
78400
62500
155800 168100 144400
91200 144400
57600
81000
90000
72900
1504400 1706600 1342900
128
Lesson Four
129
Required
Calculate the product moment correlation coefficient briefly comment on the value obtained
The produce moment correlation
r=
n xy x y
n x 2 ( x ) n y 2 ( y )
2
Workings:
4020
= 402
10
X =
r=
Y=
3550
= 355
10
10 (1, 504, 400 ) ( 4020 )( 3550 )

10 (1, 706, 600 ) 40202 10 (1,342, 900 ) ( 3550 )
= 0.89
Comment: The value obtained 0.89 suggests that the correlation between annual income and
annual expenditure is high and positive. This implies that the more one earns the more one
spends.
4.2
REGRESSION
- This is a concept, which refers to the changes which occur in the dependent variable as
a result of changes occurring on the independent variable.
- Knowledge of regression is particularly very useful in business statistics where it is
necessary to consider the corresponding changes on dependant variables whenever
independent variables change
- It should be noted that most business activities involve a dependent variable and either
one or more independent variable. Therefore knowledge of regression will enable a
business statistician to predict or estimate the expenditure value of a dependant variable
when given an independent variable e.g. consider the above example for annual incomes
and annual expenditures. Using the regression techniques one can be able to determine
the estimated expenditure of a given family if the annual income is known and vice
versa
- The general equation used in simple regression analysis is as follows
y = a + bx
Where y = Dependant variable
a= Interception y axis (constant)
b = Slope on the y axis
x = Independent variable
The determination of the regression equation such as given above is normally
i.
done by using a technique known as the method of least squares.
Regression equation of y on x i.e. y = a + bx
130
x x
x
Line of best fit
x x
x
x
x
x
x
The following sets of equations normally known as normal equation are used to determine the
equation of the above regression line when given a set of data.
y = an + bx
xy = ax + bx2
Where y = Sum of y values
xy = sum of the product of x and y
x = sum of x values
x2= sum of the squares of the x values
a = The intercept on the y axis
b = Slope gradient line of y on x
NB: The above regression line is normally used in one way only i.e. it is used to estimate the y
values when the x values are given.
Regression line of x on y i.e. x = a + by
- The fact that regression lines can only be used in one way leads to what is known as a
regression paradox
- This means that the regression lines are not ordinary mathematical line graphs which
may be used to estimate the x and y simultaneously
- Therefore one has to be careful when using regression lines as it becomes necessary to
develop an equation for x and y before doing the estimation.
The following example will illustrate how regression lines are used
Example
An investment company advertised the sale of pieces of land at different prices. The following
table shows the pieces of land their acreage and costs
130
Lesson Four
131
Piece of land
A
B
C
D
E
F
G
H
J
(x)Acreage Hectares
2.3
1.7
4.2
3.3
5.2
6.0
7.3
8.4
5.6
x =44.0
(y) Cost 000

230
150
450
310
550
590
740
850
530
y = 4400
xy
529
255
1890
1023
2860
3540
5402
7140
2969
xy= 25607
Required
Determine the regression equations of
i. y on x and hence estimate the cost of a piece of land with 4.5 hectares
ii. Estimate the expected average if the piece of land costs 900,000
y = an + bxy
xy = ax + bx2
By substituting of the appropriate values in the above equations we have
4400 = 9a + 44b .. (i)
25607 = 44a + 254.96b ..(ii)
By multiplying equation . (i) by 44 and equation (ii) by 9 we have
193600 = 396a + 1936b .. (iii)
230463 = 396a + 2294.64b ..(iv)
By subtraction of equation . (iii) from equation (iv) we have
36863 = 358.64b
102.78 = b
by substituting for b in .. (i)
4400 = 9a + 44( 102.78)
4400 4522.32 = 9a
122.32 = 9a
-13.59 = a
Therefore the equation of the regression line of y on x is
Y = 13.59 + 102.78x
When the acreage (hectares) is 4.5 then the cost
(y) = -13.59 + (102.78 x 4.5)
= 448.92
= 448, 920
Note that
Where the regression equation is given by
y= a + bx
Where a is the intercept on the y axis and
b is the slope of the line or regression coefficient
n is the sample size
then,
intercept a =
y b x
n
x2
5.29
2.89
17.64
10.89
27.04
36
53.29
70.56
31.36
x2 = 254.96
132
Slope b =
n xy x y
n x 2 ( x )
Example
The calculations for our sample size n = 10 are given below. The linear regression model is
y = a + bx
Table
Distance x miles
3.5
2.4
4.9
4.2
3.0
1.3
1.0
3.0
1.5
4.1
x = 28.9
The Slope b =
Time y mins
16
13
19
18
12
11
8
14
9
16
y = 136
xy
56.0
31.0
93.1
75.6
36.0
14.3
8.0
42.0
13.5
65.6
xy = 435.3
x2
12.25
5.76
24.01
17.64
9.0
1.69
1.0
9.0
2.25
16.81
x2 = 99.41
y2
256
169
361
324
144
121
64
196
81
256
y2= 1972
10 435.3 28.9 136 422.6

=
10 99.41 28.92
158.9
= 2.66
and the intercept a =
136 ( 2.66 28.9 )

10
= 5.91
We now insert these values in the linear model giving
y = 5.91 + 2.66x
or
Delivery time (mins) = 5.91 + 2.66 (delivery distance in miles)
The slope of the regression line is the estimated number of minutes per mile needed for a
delivery. The intercept is the estimated time to prepare for the journey and to deliver the goods,
that is the time needed for each journey other than the actual traveling time.
PREDICTION WITHIN THE RANGE OF SAMPLE DATA

We can use the linear regression model to predict the mean of dependant variable for any given
value of independent variable
For example if the sample model is given by
Time (min) = 5.91 + 2.66 (distance in miles)
Then if the distance is 4.0 miles then our estimated mean time is
= 5.91 + 2.66 x 4.0 = 16.6 minutes
132
Lesson Four
133
4.3
Multiple Linear Regression Models
There are situations in which there is more than one factor which influence the dependent
variable
Example
Cost of production per week in a large department depends on several factors;
Total numbers of hours worked
i.
Raw material used during the week
ii.
iii.
Total number of items produced during the week
Number of hours spent on repair and maintenance
iv.
It is sensible to use all the identified factors to predict department costs
Scatter diagram will not give the relationship between the various factors and total costs
The linear model for multiple linear regression if of the type; (which is the line of best fit).
y = + b1x1 +b2x2 + + bnxn
We assume that errors or residuals are negligible.
In order to choose between the models we examine the values of the multiple correlation
coefficient r and the standard deviation of the residuals .
A model which describes well the relationship between y and xs has multiple correlation
coefficient r close to 1 and the value of which is small.
Example
Odino chemicals limited are aware that its power costs are semi variable cost and over the last
six months these costs have shown the following relationship with a standard measure of output.
Month
1
2
3
4
5
6
Output (standard units)

12
18
19
20
24
30
Total power costs 000

6.2
8.0
8.6
10.4
10.2
12.4
Required
i.
Using the method of least squares, determine an appropriate linear relationship
between total power costs and output
ii.
If total power costs are related to both output and time (as measured by the number
of the month) the following least squares regression equation is obtained
Power costs = 4.42 + (0.82) output + (0.10) month
Where the regression coefficients (i.e. 0.82 and 0.10) have t values 2.64 and 0.60
respectively and coefficient of multiple correlation amounts to 0.976
Compare the relative merits of this fitted relationship with one you determine in (a).
Explain (without doing any further analysis) how you might use the data to forecast
total power costs in seven months.
Solution
a)
Output (x)
12
18
19
20
24
30
x = 123
Power costs (y)

6.2
8.0
8.6
10.4
10.2
12.4
y = 55.8
x2
144
324
361
400
576
900
x2 = 2705
y2
38.44
64.00
73.96
108.16
104.04
153.76
y2 = 542.36
xy
74.40
144.00
163.40
208.00
244.80
372.00
xy= 1,206.60
134
b=
n xy x y
n x 2 ( x )
6 1206.6 123 55.8

=
6 2705 (123)
376.2
= 0.342
1101
1
(y bx)
n
1
(55.8 0.342) 123
6
= 2.29
(Power costs) = 2.29 + 0.342 (output)
b. For linear regression calculated above, the coefficient of correlation r is

r=
( 6 1206.6 ) (123 55.8 )
6 2705 123 123 6 542.36 55.8 55.8

376.2
1101 140.52
= 0.96
This show a strong correlation between power cost and output. The multiple correlation when
both output and time are considered at the same time is 0.976.
We observe that there has been very little increase in r which means that inclusion of time
variable does not improve the correlation significantly
The value for time variable is only 0.60 which is insignificant as compared with a t value of 2.64
for the output variable
In fact, if we work out correlation between output and time, there will be a high correlation.
Hence there is no necessity of taking both the variables. Inclusion of time does improve the
correlation coefficient but by a very small amount.
If we use the linear regression analysis and attempt to find the linear relationship between output
and time i.e.
Month
1
2
3
4
5
6
Output
12
18
19
20
24
30
134
Lesson Four
135
The value of b and a will turn out to be 3.11 and 9.6 i.e. relationship will be of the form
Output = 9.6 + 3.11 month
For this equation forecast for 7th month will be
Output = 9.6 + 3.11 7
= 9.6 + 21.77
= 31.37 units
Using the equation , Power costs = 2.29 + 0.34 output
= 2.29 + 0.34 31.37
= 2.29 + 10.67
= 12.96 i.e. 12,960
Non Linear Relationships

If the scatter diagram and the correlation coefficient do not indicate linear relationship, then the
relationship may be non linear
Two such relationships are of peculiar interest
y = ab x
y = ax b
and
Both of these can be reduced to linear model. Simple or multiple linear regression methods are
then used to determine the values of the coefficients
i.
Exponential model
y = ab x
Take log of both sides
log y = log a + log bx
log y = log a + xlog b
Let log y = Y and log a = A and log b = B
Thus we get Y = A + Bx. This is a linear regression model
ii.
Geometric model
y = ax b
using the same technique as above
log y = log a + blog x
Y = A + bX
Where Y = log y
A = log a
X = log x
Using linear regression technique (the method of least squares), it is possible to calculate the
value of a and b
TIME SERIES AND ANALYSIS

This is the mathematical or statistical analysis on past data arranged in a periodic sequence.
Decision making and planning in an organization involves forecasting which is one of the time
series analysis.
Impediments in time series analysis
Accuracy of data in reflecting
a) Drastic changes e.g. in the advent of a major competitor, period of war or sudden change of
taste.
b) For long term forecasting internal and external pressures makes historical data less effective.
136
1.
Moving Average
Periodical data e.g. monthly sales may have random fluctuation every month despite a general
trend being evident. Moving average helps in smoothing away these random changes.
A moving average is the forecast for a period that takes the average of the previous periods.
Example:
The table below represents company sales, calculate 3 and 6 monthly moving averages, for the
data
Months
Sales
January
1200
February
1280
March
1310
April
1270
May
1190
June
1290
July
1410
August
1360
September
1430
October
1280
November
1410
December
1390
Solution.
These are calculated as follows
Jan + Feb + Mar
1200 + 1280 + 1310
=
Aprils forecast =
3
3
Feb + Mar + Apr 1280 + 1310 + 1270
=
Mays forecast =
3
3
And so on
Similarly for 6 monthly moving average
July forecast =
Jan + Feb + Mar + Apr + May + Jun
1200 + 1280 + 1310 + 1270 + 1190 +1290
And so on
April
May
June
July
August
September
October
November
December
3 months moving average

1263
1287
1257
1250
1297
1353
1400
1357
1373
6 months moving average
1257
1292
1305
1325
1327
1363
Note:
When plotting moving average on graphs the points are plotted as the midpoint of the period of
the average, e.g. in our example the forecast for April (1263) is plotted on mid Feb.
136
Lesson Four
137
Characteristics of moving average

1) The more the number of periods in the moving average, the greater the smoothing
effect.
2) Different moving averages produce different forecasts.
3) The more the randomness of data with underlying trend being constant then the
more the periods should be involved in the moving averages.
Limitations of moving averages.
1) Equal weighing with disregard to how more recent data is more relevant.
2) Moving average ignores data outside the period of the average thus it doesnt fully
utilise available data.
3) Where there is an underlying seasonal variation, forecasting with unadjusted moving
average can be misleading.
2.
Exponential smoothing
This is a weighted moving average technique, it is given by:
New forecast = Old forecast + (Latest Observation Old forecast)
Where = Smoothing constant
This method involves automatic weighing of past data with weights that decrease exponentially
with time.
Example
Using the previous example and smoothing constant 0.3 generate monthly forecasts
Months
Sales
Forecasts: = 0.3
January
1200
February
1280
1200
March
1310
1224
April
1270
1250
May
1190
1256
June
1290
1233
July
1410
1250
August
1360
1283
September
1430
1327
October
1280
1358
November
1410
1335
December
1390
1357
Solution
Since there were no forecasts before January we take Jan to be the forecast for February.
Feb 1200
For March;
March forecast = Feb forecast + 0.3 ( Feb sales Feb forecast)
= 1200 + 0.3 (1280 1200)
=1224
Note:
The value lies between 0 and 1.
The higher the value, the more the forecast is sensitive to the current status.
Characteristics of exponential smoothing
More weight is given to the most recent data.
138

All past data are incorporated unlike in moving averages.
Less data is needed to be stored unlike in periodic moving averages.
Decomposition of time series

Time series has the following characteristics.
A long term trend (T) tendency of the whole series to rise and fall.
a)
Seasonal variation (S) short term periodic fluctuations in values. e.g. in Kenya maize
b)
yield is high in November and low in March or matatus have better business on Friday
and very low on Sundays.
Cyclical variation (C) These are medium term changes caused by factors which apply
c)
for a while then disappear, and come back again in a repetitive cycle. e.g. drought hits
Kenya every 7 years.
Note that cyclic variation has a longer term than seasonal variation e.g. seasonal
variation may occur once every year while cyclic variation occurs once every several
years.
d)
Random residual variation (R) These are non-recurring random variations e.g. war,
fire, coup e.t.c.
For accurate forecasts these aspects are qualified separately (i.e. T,C,S and R) from data.
This is known as time decomposition or time series analysis
The separate elements are then combined to produce a forecast.
Time series models:

Additive Model
Time series value = T +S +C +R
Where S, C and R are expressed in absolute value.
This model is best suited where the component factors are independent e.g. where the
seasonal variation is unaffected by trend.
Multiplicative Model:
Time series value = T S C R
Where S, C and are expressed as percentage or proportions.
This model is best applied where characteristics interact e.g. where high trends increase
seasonal variations. Multiplicative model is more commonly used in practice.
Of the four elements of time series the most important are trend and seasonal variation. The
following illustration shows how the trend (T) and seasonal variation (S) are separated out from
a time series and how the calculated T and S values are used to prepare forecast. The process of
separating out the trend and seasonal variation is known as deseasonalising the data.
There are two approaches to this process: one is based on regression through the actual data
points and the other calculates the regression line through moving average trend points. The
method using the actual data is demonstrated first followed by the moving average method.
138
Lesson Four
1.
139
Time series analysis: trend and seasonal variation using regression on the data
The following data will be used to illustrate how the trend and seasonal variation are calculated.
Example 1
1
2
3
4
Year
Sales of widgets in 000s

Quarter 1
Quarter 2
20
32
21
42
23
39
27
39
Quarter 3
62
75
77
92
Quarter 4
29
31
48
53
It will be apparent that there is a strong seasonal element in the above data (low in Quarter 1 and
high in Quarter 3) and there is a generally upward trend.
The steps in analyzing the data and preparing a forecast are:
Step 1:
Calculate the trend in the data using the least squares method.
Step 2:
Estimate the sales for each quarter using the regression formula
established in step 1.
Step 3:
Calculate the percentage variation of each quarters actual sales from the
estimates, obtained in step 2.
Step 4:
Average the percentage variations from step 3. This establishes the

average seasonal variations.
Step 5:
Prepare forecast based on trend percentage seasonal variations.
Solution
Step 1
Calculate the trend in the data by calculating the linear regression line y = a + bx.
Year 1
Year 2
Year 3
x (quarters)
1
2
3
4
x (sales)
20
32
62
29
xy
20
64
186
116
x2
1
4
9
16
5
6
7
8
21
42
75
31
105
252
525
248
25
36
49
64
9
10
11
12
23
39
77
38
207
390
847
576
81
100
121
144
140
Year 4
13
14
15
16
x=136
27
39
92
53
y= 710
351
546
1380
848
xy= 6661
169
196
225
256
x2 =1496
Least square equations

y = an + bx
xy = ax + bx2
710 = 16a + 136b
6661 = 136a + 1496b
626 = 340b
b = 1.84 and substituting we obtain
a = 28.74
Trend line = 28.74 + 1.84x
Steps 2 and 3
Use the trend line to calculate the estimated sales for each quarter.
For example, the estimate for the first quarter in year 1 is
estimate = 28.74 + 1.84 (1) = 30.58
The actual value of sales is then expressed as a percentage of this estimate. For example, actual
sales in the first quarter were 20 so the seasonal variation is
Actual sales
20
%=
= 65%
Estimate
30.58
Year 1
Year 2
Year 3
x (quarters)
y (sales)
Trend
Actual
%
Trend
1
2
3
4
20
32
62
29
30.58
32.42
34.26
36.10
65
99
181
80
5
6
7
8
21
42
75
31
37.94
39.78
41.62
43.46
55
106
180
71
9
10
11
12
23
39
77
48
45.30
47.14
48.98
50.82
51
83
157
94
140
Lesson Four
141
Year 4
13
14
15
16
27
39
92
53
52.66
54.50
56.34
58.18
51
72
163
91
Trend estimates and percentage variations table.

Step 4
Average the percentage variations to find the average seasonal variations.
Q1
Q2
Q3
%
%
%
65
99
181
55
106
180
51
83
157
72
163
51
360
681
222
56%
90%
170%
4=
Q4
%
80
71
94
91
336
84%
These then are the average variations expected from the trend for each of the quarters; for
example, on average the first quarter of each year will be 56% of the value of the trend. Because
the variations have been averaged, the amounts over 100% (Q3 in this example). This can be
checked by adding the average and verifying that they total 400% thus:
56% + 90% + 170% + 84% = 400%.
On occasions, roundings in the calculations will make slight adjustments necessary to the
average variations.
Step 5
Prepare final forecasts based on the trend line estimates from trend estimates and percentages
variation table (i.e. 30.58, 32.42, etc) and the averaged seasonal variations from the table above.
(i.e. 56%, 90%, 170% and 84%)
The seasonally adjusted forecast is calculated thus:
Seasonally adjusted forecast = Trend estimate Seasonal variation%

X (quarters)
Year 1
Year 2
Y (sales)
1
2
3
4
20
32
62
29
Seasonally adjusted
forecast
17.12
29.18
58.24
30.32
5
6
7
8
21
42
75
31
21.24
35.80
70.75
36.51
142
Year 3
Year 4
9
10
11
12
23
39
77
48
25.37
42.43
83.27
42.69
13
14
15
16
27
39
92
53
29.49
49.05
95.78
48.87
Seasonally adjusted forecasts

The forecasts are compared with the actual data to get some idea of how good extrapolated
forecasts might be. With further analysis they enable us to quantify the residual variations.
Extrapolation using the trend and seasonal factors
Once the formulae above have been calculated, they can be used to forecast (extrapolate) future
sales. If it is required to estimate the sales for the next year (i.e. Quarters 17, 18, 19 and 20 in our
series) this is done as follows:
Quarter 17
Basic trend = 28.74 + 1.84 (17)

= 60.02
Seasonal adjustment for a first quarter = 56%

Adjusted forecast = 60.02 56%
= 33.61
A similar process produces the following figures:
Adjusted forecasts
Notes:
a)
b)
c)
Quarter 18 = 55.67
19 = 108.29
20 = 55.05
Time series decomposition is not an adaptive forecasting system like moving averages
and exponential smoothing.
Forecasts produced by such an analysis should always be treated with caution.
Changing conditions and changing seasonal factors make long term forecasting a
difficult task.
The above illustration has been an example of a multiplicative model. This is the
seasonal variations were expressed in percentage or proportionate terms. Similar steps
would have been necessary if the additive model had been used except that the
variations from the trend would have been the absolute values. For example, the first
two variations would have been
Q1: 20 30.58 = absolute variation = -10.58
Q2: 32 32.42 = absolute variation = - 0.42
And so on.
The absolute variations would have been averaged in the normal way to find the
average absolute variation, whether + or -, and these values would have been used to
make the final seasonally adjusted forecasts.
142
Lesson Four
2.
143
Trend and seasonal variation using moving averages

When the correlation coefficient is low the method of calculating the regression line
through the actual data points should not be used. This is because the regression line is
too sensitive to changes in the data values.
In such circumstances, calculating a regression line through the moving average trend
points is more robust and stable.
Example 1 is reworked below using this method and, because there are many similarities
to the earlier method, only the key stages are shown.
x
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
20
32
62
29
21
42
75
31
23
39
77
48
27
39
92
53
3 point moving
average (1)
38
41
37.3
30.7
46
49.3
43
31
46.3
54.7
50.7
38
52.7
61.3
Trend line (2)

34.38
35.70
37.02
38.34
39.66
40.98
42.30
43.62
44.94
46.26
47.58
48.90
50.22
51.54
52.86
54.18
Actual
%
Trend
58
90
167
76
53
102
177
71
51
84
162
98
54
76
174
98
Trend estimates and percentage variations utilizing moving averages

The first three moving average is calculated as follows
20 + 32 + 62
= 38 which is entered opposite period 2
3
The next calculated:
32 + 62 + 29
= 41, and so on
3
The regression line y = a + bx of the moving average values is calculated in the normal manner
and results in the following:
y = 33.06 + 1.32x
This is used to calculate the trend line:
e.g.
For Period 1:y = 33.06 + 1.32(1) = 34.38

For Period 2:y = 33.06 + 1.32 (2) = 35.70
The percentage variations are averaged as previously shown, resulting in the following values:
144

Q1
54
Average seasonal variation %
Q2
89
Q3
170
Q4
86
The trend line and the average seasonal variations are then used in a similar manner to that
previously described.
For example, to extrapolate future sales for the next year (i.e. quarters 17, 18, 19 and 20) is as
follows:
Quarter 17
Forecast sales = (33.06 + 1.32(17)) 0.54 = 29.97
A similar process produces the following figures:
Quarter
18 = 50.57
19 = 98.84
20 = 51.13
Forecast errors
Differences between actual results and predictions may arise from many reasons. They may arise
from random influences, normal sampling errors, choice of the wrong forecasting system or
alpha value or simply that the future conditions turn out to be radically different from the past.
Whatever the cause(s) management wish to know the extent of the forecast errors and various
methods exist to calculate these errors.
A commonly used technique, appropriate to time series, is to calculate the mean squared error of the
deviations between forecast and actual values then choose the forecasting system and/or
parameters which gives the lowest value of mean squared errors, i.e. akin to the least squares
method of establishing a regression line.
Longer- term forecasting

Moving averages, exponential smoothing and decomposition methods tend to be used for short
to medium term forecasting. Longer term forecasting is usually less detailed and is normally
concerned with forecasting the main trends on a year to year basis. Any of the techniques of
regression analysis described in the preceding chapters could be used depending on the
assumptions about linearity or non- linearity, the number of independent variables and so on.
The least squares regression approach is often used for trend forecasting.
Forecasting using least squares
Example 2
Data have been kept of sales over the last seven years
Year
Sales (in 000 units
1
14
2
17
3
15
4
23
5
18
6
22
7
27
It is required to forecast the sales for the 8th year
144
Lesson Four
Solution
Years (x)
1
2
3
4
5
6
7
x=28
145
Sales (y)
14
17
15
23
18
22
27
y = 136
xy
14
34
45
92
90
132
189
x2
1
4
9
16
25
36
49
xy=596
x2= 140
136 = 7a + 28b
596 = 28a + 140b
b = 1.86
And substituting in one of the equations we obtain
a = 12
Regression line = y = 12 + 1.86x
Or,
Sales in (000s of units) = 12.00 + 1.86 (no of years)
We use this expression for forecasting, for 8th year sales
= 12 + 1.86 (8)
=26.88 i.e. 26,888 units
146

QUESTION ONE
a) What is meant by correlation?
b) Why is the co-efficient of determination calculated?
c) Define R. (coefficient of rank correlation)
QUESTION TWO
Explain the difference between Linear model, exponential model and geometric model, and
write down their formulas
QUESTION THREE
An analysis of representatives car expenses shows that the expenses are dependent on the miles
travelled (x) and the type of journey (x). the general form is:
y = a + b1x1 + b2x2
Calculations have produced the following values (where y is expenses per month)
y = 86 + 0.37x1 + 0.08x2
r2x1 = 0.78
r2x2 = 0.16
R = 0.88
Interpret these values
QUESTION FOUR
Month
January
February
March
April
May
June
July
August
September
October
November
December
January
Actual Sales (units)

450
440
460
410
380
400
370
360
410
450
470
490
460
Required.
Provide 3 month, 6 month and 12 month moving average.
146
Lesson Four
147
QUESTION FIVE
The manager of a company is preparing revenue plans for the last quarter of 1993/94 and for
the first three quarters of 1994/5. The data below refer to one of the main products:
Revenue
000
1990/91
1991/92
1992/93
1993/94
April-June
Quarter 1
000
49
50
51
50
July-Sept
Quarter 2
000
37
38
40
42
Oct-Dec
Quarter 3
000
58
59
60
61
Jan-March
Quarter 4
000
67
68
70
-
Required:
a) Calculate the four-quarterly moving average trend for this set of data.
b) Calculate the seasonal factors using either the additive model or the multiplicative model,
but not both.
c) Explain, but do not calculate how you would use the results in parts (a) and (b) of this
question to forecast the revenue for the last quarter of 1993/4 and for the first three
quarters of 1994/95.
148
QUESTION SIX
A company has a fleet of vehicles and is trying to predict the annual maintenance cost per
vehicle. The following data have been supplied for a sample of vehicles:
Vehicle number
Age in years
1
2
3
4
5
6
7
8
9
10
(x)
2
8
6
8
10
4
4
2
6
10
Maintenance cost
Per annum
X 10
(y)
60
132
100
120
150
84
90
68
104
140
Required:
a) Using the least squares technique calculate the values of a and b in the equation
y = a + bx, to allow managers to predict the likely maintenance cost, knowing the age of the
vehicle.
b) Prepare a table of maintenance costs covering vehicles from 1 to 10 years of age, based on
your calculations in (a).
c) Estimate the maintenance costs of a 12-year-old vehicle and comment on the validity of
making such an estimate.
QUESTION SEVEN
A company is building a model in order to forecast total costs based on the level of output. The
following data are available for last year:
Month
January
February
March
April
May
June
July
August
September
October
November
December
Output
000 units
(x)
16
20
23
25
25
19
16
12
19
25
28
12
Costs
000
(y)
170
240
260
300
280
230
200
160
240
290
350
200
Required:
a) State two possible reasons for the large variation in output per month.
b) Plot a graph of output and costs, and comment on the relationship observed.
148
Lesson Four
149
c) Using the least square technique, calculate the values of a an b in the equation y = a + bx in
order to predict costs given the output, and explain the meaning of the calculated values.
QUESTION EIGHT
Your company has been selling data base and spreadsheets for the last four years and has found
the business to vary with season. The quarterly sales figures for the last four years are shown in
table 6b1 and table 6b2 shows the deviation from the trend at the appropriate periods
Table 6b1
Quarterly sales in 000s
Year
1983
1984
1985
1986
1987
Q2
360
430
500
590
Q3
530
750
660
710
Q4
354
395
509
521
Q1
Q2
Q3
-128
-145
-165
-37
12
43
276
153
153
Q4
-42
-93
-15
Q1
304
340
374
440
Table 6b2
Seasonal deviation from trend in 000s
Year
1983
1984
1985
1986
Required
i.
establish the trend figures from the data in the two tables
establish the seasonal variations for the four year period
ii.
using your results from parts (i) and (ii) forecast sales for 1987 quarter 2
iii.
QUESTION NINE
1. The directors of your company wish to make a serious study of the heating costs of the ****
block. The data for the last sixteen quarterly periods are tabulated as follows.
Heating costs in
Quarter
Year
1980
1981
1982
1983
1984
Q1
Q2
Q3
1730
1950
1860
1910
1554
1595
1709
1721
1504
1540
1574
1640
Q4
1560
1630
1700
1790
150

Required
a) Assuming the additive model calculate the trend of heat costs using the method of
moving averages
b) Estimate the seasonal deviations from trend
c) Estimate the heating costs for quarter IV of 1984 and comment on any factors
affecting the reliability of your forecast
150
Lesson Four
151
COMPREHESIVE ASSIGNMENT TWO

Instructions:
SECTION I
QUESTION ONE
a) In the just concluded higher education seminar at Shoppers Paradise, Nairobi, the College of
Business Administration of Highland University states in some of its promotional material
that the average graduate of the college earns over Sh.3 million a year. Assume, for
simplicity, that only four people have graduated to date; Sam, Tom, Jackie and Mary who
earn Sh.1.6 million, Sh.1.8 million, Sh.1.8 million and Sh.2 million respectively in a year.
Required:
Compute the mean, median and the mode. Is the colleges claim correct?
(3 marks)
b) Let us change our assumption about the number of graduates in (a) above and instead
assume that five people have graduated. They consist of the four listed above and Suki who
earns Sh.5.3 million per year.
Required:
Compute the mean, median and the mode for the five graduates. Is the colleges claim
correct?
(3 marks)
c) Changing our assumption one more time about the number of graduates, let us assume that
six people have graduated. They consist of the four original ones, Suki, who earns Sh.5.3
million a year; and Bob who earns Sh.6.7 million a year.
Required:
i) Compute the mean, median and mode for the six graduates. Is the colleges claim correct?
(2 marks)
ii) Comment on what happened to the mean, median and mode as you moved from part (a)
to (b) to (c) of this problem.
(2 marks)
iii) What do the results in (ii) above suggest about the relative stability of the mean, median
and the mode?
(2 marks)
iv) How do you feel about the ethics of this college in claiming that their average graduates
earn over Sh.3 million a year?
(2 marks)
d) Genuine athletic Company Ltd., manufactures weight-lifting equipment. The companys
top-of-the-line equipment are used in events such as the Olympics and other prestigious
professional weight-lifting competitions. Consequently, it is very important that if a barbell
plate is stamped say, 50 kilogrammes, it weighs very close to 50 kilogrammes. In addition,
a barbell plate must have a hole just slightly larger than 1 centimeter in diameter so that it
will slip onto the 1 centimeter diameter bar easily but fit smoothly when it is in place.
A recent sampling of barbell plates of 10 and 50 kilogrammes revealed the following
information:
152
i)
ii)
iii)
iv)

Weights of 10 kilogram plates had an arithmetic mean of 10.013 kilogrammes and a
standard deviation of 0.124 kilogrammes.
Weights of 50 kilogram plates had an arithmetic mean of 50.032 kilogrammes and a
standard deviation of 0.465 kilogrammes.
Diameters of holes in the 10 kilogram plates had an arithmetic mean of 1.22 centimeters
and a standard deviation of 0.187 centimeters.
Diameter of holes in the 50 kilogram plates had an arithmetic mean of 1.20 centimeters
and a standard deviation of 0.183 centimeters.
Required:
Determine whether the production process associated with one size of barbell plant
produced more variable results than the production process associated with the other size.
(6 marks)
(Total: 20 marks)
QUESTION TWO
a) The index of industrial production in the Utopian country by July 2001 is given below:
Sector
Mining and quarrying
Manufacturing:
Food, drink and tobacco
Chemicals
Metal
Engineering
Textiles
Other manufacturing
Construction
- Gas, electricity and water
Weight
41
July 2001 Index

(1994 = 100)
361
77
66
47
298
67
142
182
80
106
109
72
86
70
91
84
115
Required:
i) Calculate the index of industrial production for all industries and manufacturing industries.
(6 marks)
(4 marks)
ii) Comment on your results.
b) Explain some of the uses of index numbers.
c) What are some of the limitations of index numbers?
(5 marks)
(5 marks)
(Total: 20 marks)
QUESTION THREE
Leisure Publishers Ltd. recently published 20 romantic novels by 20 different authors. Sales
ranged from just over 5,000 copies for one novel about 24,000 copies for another novel. Before
publishing, each novel had been assessed by a reader who had given it a rating between 1 and 10.
The managing director suspects that the main influence on sales is the cover of the book. The
illustrations on the front covers were drawn either by artist A or artist B. the short description
on the back cover of the novel was written by either editor C or editor D.
A multiple regression analysis was done using the following variables:
Y
Sales (million of shillings)
X1
1 if front cover is by artist A
2 if front cover is by artist B
152
Lesson Four
X2
X1
153
readers rating
1 if the short description of the novel is by editor C
2 if the short description of the novel is by editor D
The computer analysis produced the following results:

Correlation coefficient r = 0.921265
Standard error of estimate
= 2.04485
Analysis of variance
Degrees of freedom
Regression
3
Residue
16
Individual analysis of variables
Variable
Coefficient
Constant
15.7588
1
-6.25485
2
0.0851136
3
5.86599
Correlation coefficients
1
-0.307729
1
Sum of squares
375.37
66.903
Mean square
125.12
1.1814
Standard error
2.54389
0.961897
0.298272
0.922233
F Value
38.375
42.284
0.081428
40.457741
0
0.123094
1
-0.674104
0.310838
0.627329
1
F ratio
29.923
Required:
a) The regression equation.
(3 marks)
(3 marks)
b) Does the regression analysis provide useful information? Explain.
c) Explain whether the covers were more important for sales than known quality of the novels.
(4 marks)
d) State with 95% confidence the difference in sales of a novel if its cover illustrations were
done by artist B instead of artist A.
(5 marks)
e) State with 95% confidence the difference in sales of a novel if its short description was by
editor D and not editor C.
(5 marks)
(Total: 20 marks)
QUESTION FOUR
a) Explain the difference between regression and correlation analysis .
(4 marks)
b) Explain why the existence of a significant correlation does not imply causation. (2 marks)
c) A bakery bakes cakes under the brand name super cakes. Irene Juma, the manageress does
not know the cost of each cake. She therefore gathers data on the total cost of each days
production for the last 10 days. The results are shown in the table below;
154
i)
ii)
Day
Number of cakes
(00 units)
Total cost (Sh.

000)
1
2
3
4
5
6
7
8
9
10
22.5
21.0
27.5
21.5
30.0
20.0
24.0
26.5
18.3
17.0
23.0
21.6
23.3
24.0
28.2
22.4
23.1
25.3
20.1
16.5
Required:
Estimate the total cost function using the ordinary least squares method. State the fixed
cost and unit cost.
(11 marks)
If each cake is sold at Sh.10, determine the break even number of cakes.
(3 marks)
(Total: 20 marks)
QUESTION FIVE
Differentiate between additive model and the multiplicative model as used in time series analysis.
(4 marks)
The sales data of XYZ Ltd. (in millions of shillings) for the years 2001 and 2004 inclusive are as
given below:
Quarter
Year
2001
2002
2003
2004
1
40
42
46
54
2
64
84
78
78
3
124
150
154
184
Required:
i) The trend in the data using the least squares method.
ii) The estimated sales for each quarter of year 2004.
iii) The percentage variation of each quarters actual sales for year 2004.
4
58
62
96
106
(8 marks)
(4 marks)
(4 marks)
(Total: 20 marks)
SECTION II
QUESTION SIX
a) Explain the following terms as used in index numbers:
i) Price index
ii) Quantity index
iii) Composite index
iv) Value index
(2 marks)
(2 marks)
(2 marks)
(2 marks)
b) The following prices and quantities reflect the average weekly consumption pattern of a
certain family for the years 2001 and 2002.
154
Lesson Four
155
Year 2001
Price (p0)
Quantity (q0)
Sh.
15
2
30
2
30
3
50
1
Item
Oranges (kg)
Milk (Litres)
Bread (Loafs)
Eggs (Dozens)
Year 2002
Price (p1)
Quantity (q1)
Sh.
25
1
35
2
40
3
65
1
Required:
i) Price relatives for each item
ii) Laspeyres price index
iii) Paasche price index
(4 marks)
(4 marks)
(4 marks)
(Total: 20 marks)
QUESTION SEVEN
Explain three methods of fitting a trend in time series analysis.
(6 marks)
The quarterly sales data for Chuce hardware are given below:
Year
2000
2001
2002
2003
2004
1
(Sh. Million)
8.5
9.5
10.4
9.5
10.9
Quarter
2
3
(Sh. Million)
(Sh. Million)
10.4
7.5
12.2
8.8
13.5
9.7
11.7
8.4
13.7
10.1
Required:
(a)
The centred four-quarter moving averages.
(b)
The specific seasonal variation for each quarter
The typical seasonal indices
(c)
Explain the third quarter typical seasonal index
(d)
4
(Sh. Million)
11.8
13.6
13.1
12.9
15.0
(6 marks)
(3 marks)
(3 marks)
(2 marks)
(Total: 20 marks)
QUESTION EIGHT
a) A machine produces circular bolts and as a quality control test, 250 bolts were selected
randomly and the diameter of their heads measured as follows:
Diameter of head (cm)
0.9747
0.9749
0.9750
0.9752
0.9753
0.9755
0.9756
0.9758
0.9759
0.9761
0.9762
0.9764
0.9765
0.9767
0.9768
0.9770
0.9771
0.9773
0.9774
0.9776
0.9777
0.9779
0.9780
0.9782
Number of components
2
6
8
15
42
68
49
25
18
12
4
1
156
Required:
b)
Determine whether the customer is getting reasonable value if the label on the circular
i)
bolt advertises that the average diameter of the head is 0.97642 cm.
(8 marks)
In what situation would weighted mean be used?
(3 marks)
ii)
Describe briefly how to estimate the median on a grouped frequency distribution
iii)
graphically?
(3 marks)
iv)
Why is the mode not used extensively in statistical analysis?
(3 marks)
The standard deviation is the natural partner to the mean. Explain
(3 marks)
(Total: 20 marks)
156
Probability
157
LESSON FIVE
Probability
Contents
- Probability theory
- Bayes Theorem and conditional probability
- Permutations and combinations
- Discrete probability distributions
- Continuous probability distribution
158
Lesson Five
5.1
-
PROBABILITY
Probability is a very popular concept in business management. This is because it covers
the risks which may be involved in certain business situations. It is a fact that when a
business investment is being arranged, the outcome is usually uncertain. Therefore the
concept of probability may be used to describe the degree of uncertainty of a
particular business outcome
Probability may therefore be defied as the chances of a given event occurring.
Numerically, probability values range between 0 and 1. a probability of 0 implies that
the event cannot occur at all. A probability of 1 implies that the event will certainly
occur.
Therefore other events have their probabilities with values lying between 0 and 1
The formular used to determine probability is as follow
r Favourable outcomes
Probability (x) = =
n
Total outcomes
Application of Probability in Business

1. Business games of chance e.g. Raffles Lotteries e.t.c.
2. Insurance firms: this is usually done when a new client or property is being insured. The
company has to be certain about the chances of the insured risks occurring.
3. Business decision making regarding viability of projects thus the projects with a greater
probability has greater chances.
Example
A bag contains 80 balls of which 20 are red, 25 are blue and 35 are white. A ball is picked at
random what is the probability that the ball picked is:
(i)
Red ball
(ii)
Black ball
(iii)
Red or Blue ball.
Solution
(i)
Probability of a red ball =

=
(ii)
Probability of black ball =

=
(iii)
P(R or B) =
20
80
or
=
25
80
9
Number of red balls in the bag

Total number of balls in the bag
20
80 4
Number of black balls in the bag
Total number of balls
0
80
= P(R )
= P(B )
=0
20 25
+
80 80
16
Note: in probability or is replaced by a plus (+) sign. See addition rule.
Common terms
Events: an event is a possible outcome of an experiment or a result of a trial or an observation.
158
Probability
159
Mutually exclusive events

A set of events is said to be mutually exclusive if the occurance of any one of the events
precludes the occurrence of any of the other events e.g. when tossing a coin, the events are a
head or a tail these are said to be mutually exclusive since the occurrence of heads for instance
implies that tails cannot and has not occurred.
It can be represented in venn diagram as.
E1
E2
E1 E2 =
E1
Non-mutually exclusive events

(independent events)
E2
E1 E2
Consider a survey in which a random sample of registered voters is selected. For each voter
selected their sex and political party affiliation are noted. The events KANU and woman
are not mutually exclusive because the selection of KANU does not preclude the possibly that
the voter is also a woman.
Independent Events
Events are said to be independent when the occurance of any of the events does not affect the
occurrence of the other(s).
e.g. the outcome of tossing a coin is independent of the outcome of the preceeding or
succeeding toss.
Example
From a pack of playing cards what is the probability of;
(i)
Picking either a Diamond or a Heart mutually exclusive
(ii)
Picking eigher a Flower or an Ace indepent events
Solutions.
(i)
P(Diamond or Heart)
= P(Diamond) + P(Heart)
160
Lesson Five
13
52
= 0.5
(ii)
13
52
26
52
P(Flower or Ace)
= P(Flower) + P(Ace) P(Flower and Ace)
13 4
1
+
=
52 52 52
4
=
= 0.31
52
Note: that the formula used incase of independent events is different to the one of
mutually exclusive.
Rules of Probability
(a)
Additional Rule This rule is used to calculate the probability of two or more mutually
exclusive events. In such circumstances the probability of the separate events must be
added.
Example
What is the probability of throwing a 3 or a 6 with a throw of a die?
Solution
P(throwing a 3 or a 6) = 1 + 1 = 1
6
6
3
(b)
Multiplicative rule
This is used when there is a string of independent events for which individual
probability is known and it is required to know the overall probability.
Example
What is the probability of a 3 and a 6 with two throws of a die?
Solution
P(throwing a 3) and P(6)
= P(3) and P(6) = 1 1 = 1
6
6
36
Note:
1)
2)
In probability and is replaced by x multiplication.

P(x) and P(y) P(x and y) note that these two are different. The first
implies P(x) happening and P(y), but if the order of which happened
first is unimportant then we have p(x and y).
In the example above:
P(3) and P(6) = 1
36
but
P(3 and 6) = P(3 followed by 6) or P(6 followed by 3)
= [P(3) P(6)] or [P(6) P(3)]
= 1 + 1 = 1
36
36
18
160
Probability
(c)
161
Conditional probability
This is the probability associated with combinations of events but given that some prior
result has already been achieved with one of them.
Its expressed in the form of
P(x|y) = Probability of x given that y has already occurred.
P ( xy )
conditional probability formula.
P(x|y) =
P( y )
Example:
In a competitive examination. 30 candidates are to be selected. In all 600 candidates appear in a
written test, and 100 will be called for the interview.
(i)
What is the probability that a person will be called for the interview?
(ii)
Determine the probability of a person getting selected if he has been called for the
interview?
(iii)
Probability that person is called for the interview and is selected?
Solution:
Let event A be that the person is called for the interview and event B that he is selected.
(i)
P(A) =
(ii)
P(B|A) =
(iii)
100
600
30
= 1
3
100 10
P(AB) = P(A) P(B|A)
= 1 3 = 3 = 1
6
10
60
20
=
Example:
From past experience a machine is known to be set up correctly on 90% of occasions. If the
machine is set up correctly then 95% of good parts are expected but if the machine is not set up
correctly then the probability of a good part is only 30%.
On a particular day the machine is set up and the first component produced and found to be
good. What is the probability that the machine is set up correctly.
Solution:
This is displayed in the form of a probability tree or diagram as follows:
GP = 0.95
CS = 0.9
IS = 0.1
CS GP
CS Correct Setting
IS Incorrect Setting
BP = 0.05
GP = 0.3
CS BP
GP Good Product
IS GP
BP Bad Product
BP = 0.7
IS BP
162
Lesson Five
P(CSGP) = 0.9 0.95 = 0.855

P(CSBP) = 0.9 0.05 = 0.045
P(ISGP) = 0.1 0.3 = 0.03
P(ISBP) = 0.1 0.7 = 0.07
1.00
- Probability of getting a good part (GP) = CSGP or ISGP
= CSGP + ISGP
= 0.855 + 0.03 = 0.885
Note: Good parts may be produced when the machine is correctly set up and also when its
incorrectly setup. In 1000 trials, 855 occasions when its correctly setup and good parts produced
(CSGP) and 30 occasions when its incorrectly setup and good parts produced (ISGP).
- Probability that the machine is correctly set up after getting a good part.
Number of favourable outcomes P(CSGP) 0.855
=
=
=
= 0.966
Total possible outcomes
P(GP)
0.885
Or
P(CSGP) 0.855
= P(CS|GP) =
=
= 0.966
P(GP)
0.885
Example
In a class of 100 students, 36 are male and studying accounting, 9 are male but not studying
accounting, 42 are female and studying accounting, 13 are female and are not studying
accounting.
Use these data to deduce probabilities concerning a student drawn at random.
Solution:
Accounting A
Male M
Female F
Total
45
P(M) =
P(F) =
P(A) =
()
36
42
78
100
55
100
78
P A =
Not accounting A
9
13
22
Total
45
55
100
= 0.45
= 0.55
100
22
100
= 0.78
= 0.22
P(M and A) = P(A and M) =
36
100
= 0.36
P(M and A ) = 0.09

P(F and A ) = 0.13
162
Probability
163
These probabilities can be express differently as;

P(M) = P(M and A) or P(M and A )
= 0.36 + 0.09 = 0.45
P(F) = P(F and A) or P(F and A )
= 0.42 + 0.13 = 0.55
P(A) = P(A and M) + P(A and F) = 0.36 + 0.42 = 0.78
()
P A = P( A and M) + P( A and F) = 0.09 + 0.13 = 0.22

Now calculate the probability that a student is studying accounting given that he is male.
This is a conditional probability given as P(A|M)
P(A and M) 0.36
P(A|M) =
=
= 0.80
P(M)
0.45
From the formula above we get that,
P(A and M) = P(M) P(A|M) .. (i)
Note that P(A|M) P(M|A)
Since P(M|A) =
P(A and M )
this is known as the Bayes rule.
P(A)
Bayes rule/Theorem
This rule or theorem is given by
P(A|B) =
( )
P (A ) P B A
P(B)
Its used frequently in decision making where information is given the in form of conditional
probabilities and the reverse of these probabilities must be found.
Example
Analysis of questionnaire completed by holiday makers showed that 0.75 classified their holiday
as good at Malindi. The probability of hot weather in the resort is 0.6. If the probability of
regarding holiday as good given hot weather is 0.9, what is the probability that there was hot
weather if a holiday maker considers his holiday good?
Solution
P(A|B) =
( )
P (A ) P B A
P(B)
Let H = hot weather

G = Good
P(G) = 0.75
P(H) = 0.6 and P(G|H) = 0.9 (Probability of regard holiday as
good given hot weather)
Now the question requires us to get
164
Lesson Five
P(H|G) = Probability of (there was) hot weather given that the holiday has been rated
as good).
P(H )P G H (0.6 )(0.9 )
=
=
P(G)
0.75
= 0.72.
Worked examples on probability

1. A machine comprises of 3 transformers A, B and C. The machine may operate if at least 2
transformers are working. The probability of each transformer working are given as shown
below;
P(A) = 0.6,
P(B) = 0.5,
P(C) = 0.7
A mechanical engineer went to inspect the working conditions of those transformers. Find the
probabilities of having the following outcomes
Only one transformer operating
i.
Two transformers are operating
ii.
All three transformers are operating
iii.
None is operating
iv.
v.
At least 2 are operating
At most 2 are operating
vi.
Solution
P(A) =0.6
P( A ) = 0.4
P(C) = 0.7
P( C ) = 0.3
i.
P(B) = 0.5
P(~B) = 0.5
P(only one transformer is operating) is given by the following possibilities

2nd
3rd
1st
P
(A
C)
= 0.6 x 0.5 x 0.3 = 0.09
(A
C)
= 0.4 x 0.5 x 0.3 = 0.06
(A
C)
= 0.4 x 0.5 x 0.7 = 0.14
P(Only one transformer working)

= 0.09 + 0.06 + 0.14 = 0.29
P(only two transformers are operating) is given by the following possibilities.
1st
2nd
3rd
ii.
P
(A
C)
= 0.6 x 0.5 x 0.3 = 0.09
(A
C)
= 0.6 x 0.5 x 0.7 = 0.21
(A
C)
= 0.4 x 0.5 x 0.7 = 0.14
P(Only two transformers are operating)

= 0.09 + 0.21 + 0.14 = 0.44
P(all the three transformers are operating).
iii.
= P(A) x P(B) x P(C)
= 0.6 x 0.5 x 0.7
= 0.21
iv.
P(none of the transformers is operating).
164
Probability
165
= P( A ) x P( B ) x P( C )
= 0.4 x 0.5 x 0.3
= 0.06
P(at least 2 working).
= P(exactly 2 working) + P(all three working)
v.
= 0.44 + 0.21
= 0.65
P(at most 2 working).
= P(Zero working) + P(one working) + P(two working)
= 0.06 + 0.29 + 0.44
= 0.79
vi.
5.3
Permutations and Combinations
Definition
Permutation
- This is an order arrangement of items in which the order must be strictly observed
Example
Let x, y and z be any three items. Arrange these in all possible permutations
2nd
Y
Z
X
Z
Y
X
1st
X
X
Y
Y
Z
Z
3rd
Z
Y
Z
X
X
Y
Six different permutations
NB: The above 6 permutations are the maximum one can ever obtain in a situation where there
are only 3 items but if the number of items exceeds 3 then determining the no. of permutations
by outlining as done above may be cumbersome. Therefore we use a special formula to
determine such permutations. The formula is given below
The number of permutations of r items taken from a sample of n items may be provided as
n
Pr =
n!
(n - r )!
where; ! = factorial
e.g.
i.
P3 =
3!
(3 3)!
3 21
0!
note; 0! = 1
166
Lesson Five
ii.
P3
=
=
iii.
P5
6
=6
1
5!
(5 - 3)!
5 4 3 21
1 2
60
(7 - 5)!
7!
7 65 4 3 21
21
5040
2
2520
Example
There are 6 contestants for the post of chairman secretary and treasurer. These positions can be
filled by any of the 6. Find the possible no. of ways in which the 3 positions may be filled.
Solution
Chairman
Secretary
Treasurer
6
5
4
Therefore the no of ways of filing the three positions is 6 x 5 x 4 = 120
6
P3
=
=
6!
(6 - 3)!
65 4 3 21
3 2 1
720
6
120
Combinations
Definition
A combination is a group of times in which order is not important.
For a combination to hold at any given time it must comprise of the same items but if a new
item is added to the group or removed from the group then we have a new combination
Example
3 items x, y and z will have 6 different permutations but only one combination.
The following formular is usually used to determine the no. of combinations in a given situation.
n
Cr =
n!
r !( n r ) !
Example
i.
C7 =
8!
7!( 8 7 ) !
166
Probability
167
8! 8 7!
=
7!1! 1 7!
=8
ii.
C4 =
=
6!
4!( 6 4 ) !
6! 6 5 4!
=
4!2! 4! 2 1
= 15
iii.
8!
3!5!
8 7 6 5!
=
3 2 1 5!
C3 =
= 56
Example
There is a committee to be selected comprising of 5 people from a group of 5 men and 6
women. If the selection is randomly done. Find the possibility of having the following
possibilities (combinations)
i.
Three men and two women
At least one man and at least one woman must be in the committee
ii.
One particular man and one particular woman must not be in the committee (one
iii.
man four women)
Solution
i.
The committee size = 5 people

The group size = 5m + 6w
assuming no restrictions the committee can be selected in 11C5

the committee has to consist of 3m & 2w
these may be selected as follows.
5C 6C
3
2
P(committee 3m and 2w)
5
C3 6C2
11
C5
note that this formula can be fed directly to your scientific calculator and attain a solution.
5!
6!
3!2! 4!2!
11!
5!6!
5 4 3 2 1 6 5 4! 5 4 3 2 1 6!
3 2 1 2 1 2 1 4! 1110 9 8 7 6!
27
77
168
Lesson Five
ii.
P(at least one man and at least one woman must be in the committee)
The no. of possible combinations of selecting the committee without any woman = 5C5
The probability of having a committee of five men only
5
C5
1
=
C5 462
11
the probability of having a committee of five women only
6!
5!1!
11!
5!6!
C
= 11 5 =
C5
6 5!
5!6!
5!1! 11 10 9 8 7 6!
1
77
P(at least one man and at least one woman)

= 1 {P(no man) + P(no woman)}
1
1
=1 +
77 462
= 1 (6 + 1)
=1
=
iii.
462
7
462
455
465
P(one particular man and one particular woman must not be in the committee
would be determined as follows
The group size
= 5m + 6w
Committee size
= 5 people
Actual groups size from which to
Select the committee
= 4m + 5w
Committee
= 1m + 4w
The committee may be selected in 9C5
The one man may be selected in 4C1 ways
The four women may be selected in 5C4 ways
P(committee of 4w1man).
5
C4 4C1
9
C5
5!
4!
4!1! 1!3!
9!
4!5!
168
Probability
169
5 4! 4 3!
4!5!
1!4! 1 3! 9 8 7 6 5!
10
63
5.4 DISCRETE PROBABILITY DISTRIBUTIONS

BINOMIAL PROBABILITY DISTRIBUTION
Binomial probability distribution is a set of probabilities for discrete events. Discrete events are
those whose results or outcomes can be counted. Binomial probabilities are commonly
encountered in business situations e.g. in quality control activities the binomial probabilities are
frequently used especially when determining the probability of having a certain no. of defective
items in a given consignment.
- The binomial probability distribution is usually characterized by the fact that the
binomial events have to fulfill the following properties
Each event has 2 possible outcomes only known as success or failure
i.
The probability of each outcome is independent of the previous outcomes
ii.
iii.
The sample size is generally fixed
The probabilities of success and failure tend to approach 0.5 if the sample size
iv.
increases (in the event when an unbiased coin is thrown a number of times)
The probabilities are given by the following equation
v.
P ( r ) = 9C5 p r (1 p )
nr
n!
nr
p r (1 p )
r !( n r )
Where
p = Probability of success
r = no. of successes
n = sample size
q = 1 P = Probability of failure
Example 1
A medical survey was conducted in order to establish the proportion of the population which
was infected with cancer. The results indicated that 40% of the population were suffering from
the disease.
A sample of 6 people was later taken and examined for the disease. Find the probability that the
following outcomes were observed
a) Only one person had the disease
b) Exactly two people had the disease
c) At most two people had the disease
d) At least two people had the disease
e) Three or four people had the disease
Solution
P(a persona having cancer)
= 40% = 0.4
P(a person not having cancer)
= 60% = 0.6
a) P(only one person having cancer)
= 6C1 (0.4)(0.6)5
=
=P
=1p=q
6!
(0.4)1(0.6)5
5 !1!
170
Lesson Five
= 0.1866
Note that from the formula
nC prqn-r:
n = sample size = 6
r
p = 0.4
r = 1 = only one person having cancer
b) P(2 people had the disease)
= 6C2 (0.4)2 (0.6)4
=
=
6!
(0.4) 2 (0.6)5
4!2!
654 !
(0.4) 2 (0.6)5
4 ! 2 1
= 15 (0.4) 2 (0.6)5
= 0.311
c) P(at most 2) = P(0) + P(1) + P(2) = P(0) or P(1) or P(2)
So we calculate the probability of each and add them up.
P(0) = P(nobody having cancer)
= 6C0 (0.4) 0(0.6)6
=
6!
(0.4) 0(0.6)6
0 !6 !
= (0.6)6
= 0.0467
The probabilities of P(1) and P(2) have been worked out in part (a) and (b)
Therefore P(at most 2) = 0.0467 + 0.1866 + 0.311 = 0.5443
d) P(at least 2)
= P(2) + P(3) + P(4) + P(5) + P(6)
= 1 [P(0) + P(1)] This is a shorter way of working out the solution since
[P(0) + P(1) + P(2) + P(3) + P(4) + P(5) + P(6) = 1]

= 1 (0.0467 + 0.1866)
= 0.7667
e) P(3 or 4 people had the disease)
= P(3) +P(4)
= 6C3(0.4)3(0.6)3 + 6C4(0.4)4(0.6)2
=
6!
6!
(0.4) 3(0.6)3 +
(0.4) 4(0.6)2
3!3!
2! 4!
= 6 5 4 3! (0.4) 3(0.6)3 + 6 5 4! (0.4) 4(0.6)2

3 2 1 3!
2 1 4!
= 20(0.4)3(0.6)3 + 15(0.4)4(0.6)2
= (20 0.013824) + (15 0.009216)
= 0.27648 + 0.13824
= 0.41472
170
Probability
171
Example 2
An insurance company takes a keen interest in the age at which a person is insured.
Consequently a survey conducted on prospective clients indicated that for clients having the
same age the probability that they will be alive in 30 years time is 2 3 . This probability was
established using the actuarial tables. If a sample of 5 people was insured now, find the
probability of having the following possible outcomes in 30 years
a) All are alive
b) At least 3 are alive
c) At most one is alive
d) None is alive
e) At least 1 is alive
Sample size = 5
P ( alive ) = p =
a)
where as
2
3
P ( not alive ) = q = 13
P ( all alive ) = P ( r = 5 )
= 5C5 ( 23 )
=
( 13 )
5! 2 5 1 0
( )( )
5!0! 3 3
= ( 23 )
32
243
P ( atleast 3 alive ) = P ( r 3)
=
b)
= P ( 3) orP ( 4 ) orP ( 5 ) = P ( 3) + P ( 4 ) + P ( 5 )
P ( 3) = C3 (
5
P ( 4 ) = 5C4 ( 23 )
)( )
2 3
3
1 2
3
5! 2 5 1 0
3
2
=
( 3 ) ( 3 ) = 10 ( 23 ) ( 13 )
3!2!
80
=
243
P (5) =
32
243
80
80
32
+
+
243 243 243
192
=
243
P ( 3) =
( 13 )
5! 2 4 1 1 5 4! 2 4 1
( 3) (3) =
( 3) (3)
4!1!
4! 1
80
=
243
=
172
c)
Lesson Five
P ( atmost 1 is alive ) = P ( r 1)
d)
= P ( 0 ) + P (1)
= 5C0 ( 23 )
( 13 )
P ( none is alive ) = P ( r = 0 )
= 5C0 ( 23 )
+ 5C1 ( 23 )
( 13 )
5! 1 5 5! 2 1 4
( ) + ( 3 )( 3 )
0!5! 3
1!4!
1
10
=
+
243 243
11
=
243
=
e)
( 13 )
1
243
P ( atleast 1 alive ) = P ( r 1)
= 1 P ( none alive )
1
243
242
=
243
= 1
POISSON PROBABILITY DISTRIBUTION

- This is a set of probabilities which is obtained for discrete events which are described as
being rare. Occasions similar to binominal distribution but have very low probabilities
and large sample size.
Examples of such events in business are as follows:
Telephone congestion at midnight
i.
ii.
Traffic jams at certain roads at 9 oclock at night
Sales boom
iii.
Attaining an age of 100 years (Centureon)
iv.
- Poisson probabilities are frequently applied in business situations in order to determine
the numerical probabilities of such events occurring.
The formula used to determine such probabilities is as follows
P ( x) =
e x
x!
Where
x = No. of successes
= mean no. of the successes in the sample ( = np)

e = 2.718
Example 1
A manufacturer assures his customers that the probability of having defective item is 0.005. A
sample of 1000 items was inspected. Find the probabilities of having the following possible
outcomes
Only one is defective
i.
At most 2 defective
ii.
More than 3 defective
iii.
P(x) =
i.
e x
x!
( = np = 1000 0.005) = 5
P(only one is defective) = P(1) = P(x = 1)
=
2.718 5 51
1!
Note that 2.718-5=
1
2.718 5
172
Probability
173
5
2.718 5
5
148.33
ii.
0.0337
P(at most 2 defective) = P(x 2)
=
P(0) + P(1) + P(2)
e 5 50
0!
P(x = 0) =
= 2.718-5
=
=
P(1) = 0.0337
P(2) =
2.718+ 5
1
148.336
= 0.00674
2.7185 52
2!
25
2 148.336
= 0.08427
P(x2) = 0.00674 + 0.0337 + 0.08427

= 0.012471
iii.
P(more than 3 defective) = P(x > 3)

= 1 [P(0 ) + P(1) + P(2 ) + P(3)]
BINOMIAL MATHEMATICAL PROPERTIES

1. The mean or expected value = n p = np
Where; n = Sample Size
2. The variance = npq
Where; q = probability of failure = 1 - p
3. The standard deviation =
npq
Example
A firm is manufacturing 45,000 units of nuts. The probability of having a defective nut is 0.15
Calculate the following
The expected no. of defective nuts
i.
The variance and standard deviation of the defective nuts in a daily consignment of
ii.
45,000
Solution
Sample size n = 45,000
P(defective) = 0.15 = p
P(non defective) = 0.85 = q
the expected no of defective nuts
= 45,000 0.15 = 6,750
The variance = npq
ii.
= 45000 0.85 0.15
i.
174
Lesson Five
= 5737.50
The standard deviation =
npq
= 5737.50
= 75.74
POISSON MATHEMATICAL PROPERTIES

1. The mean or expected value = np =
Where; n = Sample Size
2. The variance = np =
3. Standard deviation =
np =
Example
The probability of a rare disease striking a given population is 0.003. A sample of 10000 was
examined. Find the expected no. suffering from the disease and hence determine the variance
and the standard deviation for the above problem
Solution
Sample size
n = 10000
P(a person suffering from the disease) = 0.003 = p
expected number of people suffering from the disease
Mean = = 10000 0.003
= 30
= np =
variance = np = 30
Standard deviation =
np
= 30
= 5.477
5.5
PROBABILITY DISTRIBUTION FOR CONTINUOUS RANDOM
VARIABLES.
In a continuous distribution, the variable can take any value within a specified range, e.g. 2.21 or
1.64 compared to the specific values taken by a discrete variable e.g 1 or 3. The probability is
represented by the area under the probability density curve between the given values.
The uniform distribution, the normal probability distribution and the exponential distribution
are examples of a continuous distribution
- The normal distribution is a probability distribution which is used to determine
probabilities of continuous variables
Examples of continuous variables are
o Distances
o Times
o Weights
o Heights
o Capacity e.t.c
- Usually continuous variables are those, which can be measured by using the appropriate
units of measurement.
- Following are the properties of the normal distribution
174
Probability
175
1. The total area under the curve is = 1 which is equivalent to the maximum value
of probability
Normal probability
Distribution curve
Line of symmetry
Tail end
Tail end
Age (Yrs)
2. The line of symmetry divides the curve into two equal halves
3. The two ends of the normal distribution curve continuously approach the
horizontal axis but they never cross it
4. The values of the mean, mode and median are all equal
NB: The above distribution curve is referred to as normal probability distribution curve because
if a frequency distribution curve is plotted from measurements of a given sample drawn from a
normal population then a graph similar to the normal curve must be obtained.
- It should be noted that 68% of any population lies within one standard deviation, 1
- 95% lies within two standard deviations 2
- 99% lies within three standard deviations 3
Where = standard deviation
176
Lesson Five
STANDARDIZATION OF VARIABLES
- Before we use the normal distribution curve to determine probabilities of the
continuous variables, we need to standardize the original units of measurement, by
using the following formular.

Z=
Where = Value to be standardized

Z = Standardization of x
= population mean
= Standard deviation
Example
A sample of students had a mean age of 35 years with a standard deviation of 5 years. A student
was randomly picked from a group of 200 students. Find the probability that the age of the
student turned out to be as follows
i.
Lying between 35 and 40
ii.
iii.
Lying beyond 45 yrs
iv.
Lying beyond 30 yrs
v.
Lying below 25 years
vi.
Solution
(i). The standardized value for 35 years

Z=
The standardized value for 40 years

40 - 35
=
=
Z=
5
35 - 35
= 0
5
the area between Z = 0 and Z = 1 is 0.3413 (These values are checked from the normal
tables see appendix)
The value from standard normal curve tables.
When z = 0,
p=0
And when z = 1, p = 0.3413
Now the area under this curve is the area between z = 1 and z = 0
= 0.3413 0 = 0.3413
the probability age lying between 35 and 40 yrs is 0.3413
(ii). 30 and 40 years

30 35
5
=
=
Z=
5
5

40 35
=
=
Z=
5
= -1
1
the area between Z = -1 and Z = 1 is

= 0.3413 (lying on the positive side of zero) + 0.3413 (lying on the negative side of
zero)
176
Probability
177
P = 0.6826
the probability age lying between 30 and 40 yrs is 0.6826

(iii). 25 and 30 years

25 35
10
Z=
=
=
= -2
5
5
Z=
30 35
5
= -1
the area between Z = -2 and Z = -1

Probability area corresponding to Z = -2
= 0.4772 (the z value to check from the tables is 2)
Probability area corresponding to Z = -1
= 0.3413 (the z value for this case is 1)
the probability that the age lies between 25 and 30 yrs
= 0.4772 0.3413 (The area under this curve)
P = 0.1359
iv). P(beyond 45 years) is determined as follow = P(x > 45)

45 35 + 10
=
=
=+2
Z=
5
5
Probability corresponding to Z = 2 = 0.4772 = probability of between 35 and 45
P(Age > 45yrs) = 0.5000 0.4772

= 0.0228
The exponential distribution
The exponential distribution is of particular importance because of the wide ranging nature of
the practical situations in which it is used.
Examples
1. The length of time until an electronic device fails
2. The time required to wait for the first emission of a particle from a radio active
source
3. The length of time between successive accidents in a large factory
Assume that a probability density function f(x) is valid between the values a and b, then
b
(i)..
f ( x)dx = 1 i.e. The area under the curve is equal to 1

a
(ii).The mean of the distribution E ( x ) = xf ( x )dx
(iii)
The variance of the distribution = E(x2) [E(x)]2

b
( )
Where E x 2 = x 2 f ( x )dx
a
178
Lesson Five
Example of continuous probability distribution function

The distribution of a random variable x has a probability density function f(x) given by
f(x) = kx for 0 x 1
f(x) = 0 elsewhere
Where k is constant
Required.
i.
Show the value of k is 2
ii.
Find the mean of f(x)
Find the variance of f(x)
iii.
Solution
1
i)
ii)
f ( x )dx = 1
Mean = E ( x ) = xf ( x )dx
0
kx.dx =
0
k
2
k
2
x 2 = 1
0
2 x dx =
2
(1 0 ) = 1
2
3
x 3
( 23 0 ) = 23
k = 2
iii)
Variance = E ( x 2 ) E ( x )
b
= x 2 f ( x ) dx [ Mean ]
= ( x 2 2 x )dx ( 23 )
= 12 x 4 94
0
=
1
2
4
9
Variance =
1
18
Exponential distribution
Example
The mean life of an electrical component is 100 hours and its life has an exponential
distribution.
Find
a. The probability that it will last less than 60 hours
b. The probability that it will last more than 90 hours
Solution
A continuous random variable X has an exponential distribution, if for some constant k >0 it
has the probability density function
= k .e kx for
f ( x)
= 0
x0
elsewhere
178
Probability
179
The function f(x) is positive for all values of x and the area under the curve
f ( x ) dx = ke kx dx = 1
0
The mean of an exponential distribution with parameter k is
1
k
and its variance is
1
k2
Example
The mean of an exponential distribution is 100, find;
a) P(x<60)
b) P(x>90)
solution.
a) P ( x < 60 ) =
60
1
1 100 x
100
mean = 100 thus k =
dx
1
100
60
= e 100 = 1e0.6
0
= 0.45
b) P x > 90 = 1 P ( x 90 )
90
1
= 1 100
e
x
100
dx
90
= 1 e0.9 = e0.9
0
= 0.41
The students t distribution
The students t distribution was presented by W. S. Gosset in 1908 under the pen name of
student. The t distribution is of great importance in the so called small sample tests and is
profoundly used in statistical inference
The t distribution has a single parameter, known as the number of degrees of freedom. It is
denoted by the Greek symbol (read as nu). It can be interpreted as the number of useful items
of information generated by a sample of given size. The degrees of freedom are sample size less
one (v = n-1)
Properties of t distribution
1.
The t distribution ranges from to first as does the normal distribution
The t distribution like the standard normal distribution is bell shaped and
2.
symmetrical around mean zero
The shapes of the t distribution changes as the number of degrees of freedom
3.
changes
4.
The t distribution is more platykurtic that the normal distribution
The t distribution has a greater dispersion than the standard normal
5.
distribution. As n gets larger the t distribution approaches the normal
distribution when n = 30 the difference is very small
180
Lesson Five
Relation between the t distribution and standard normal distribution is shown in the following
diagram
Standard normal distribution

T distribution n = 15
T distribution n = 5
-4
-3
-2
-1
Note that the t distribution has different shapes depending on the size of the sample. When the
sample is quite small the height of the t distribution is shorter than the normal distribution and
the tails are wider.
Assumptions of t distribution
1. The sample observations are random
2. Samples are drawn from normal distribution
3. The size of sample is thirty or less n 30
Application of t distribution
- Estimation of population mean from small samples
- Test of hypothesis about the population mean
- Test of hypothesis about the difference between two means
Chi Square distribution
Chi square was first used by Karl Pearson in 1900. It is denoted by the Greek letter 2. it
contains only one parameter, called the number of degrees of freedom (d-f), where the term
degree of freedom represents the number of independent random variables that express the chi
square
Properties
1.
Its critical values vary with the degree of freedom. For every increase in the number
of degrees of freedom there is a new 2 distribution.
This possesses additional property so that when 21 and 22 are independent and
2.
have a chi square distribution with n1 and n2 degrees of from 21 + 22 will also be
distributed as a chi square distribution with n1 + n2 degrees of freedom
Where the degrees of freedom is 3.0 and less the distribution of 2 is skewed. But,
3.
for degrees of freedom greater than 30 in a distribution, the values of 2 are
normally distributed
180
Probability
181
4.
The 2 function has only one parameter, the number of degrees of freedom.
P(x)
=1
=2
=3
=4
=5
10
5. 2 distribution is a continuous probability distribution which has the value zero

at its lower limit and extends to infinity in the positive direction. Negative value
of 2 is not possible because the differences between the observed and expected
frequencies are always squared
F distribution or Variance ratio distribution

It was developed by R. A Fisher in 1924 and is usually defined in terms of the ratio of the
variances of two normally distributed populations
It is used to test the hypothesis that the two normally distributed populations have two equal
variances
F distribution ratio of the variances between two normally distributed population may be
s12
expressed as
s 22
d12
d 22
With 1 = n11 and 2 = n21 degrees of freedom

Where normal population means are unknown
n1 sample size of independent random 1
n2 sample size of independent random 2
182
Lesson Five
s12 - Sample variance of 1

s22 sample variance of 2
d12 - Population variance of 1
d 22 Population variance of 2
s12 and s22 are given by
s12
(x x )
=
s22
(x
=
as the unbiased estimator of d12
n1 1
2
x2
n1 1
as the unbiased estimator of d22
if d12 = d 22 then the statistic F =
S12
S 22
Larger estimate of variance

Smaller estimate of variance
F Distribution with n11 and n21 degrees of freedom. F distribution depends on the degrees
of freedom 1 for the numerator and 2 for the denominator. It has parameters 1 and 2
such that for different values of 1 and 2 will have different distributions.
Properties
1.
2.
The shape of the f distribution depends upon the number of degrees of

freedom
The mean and variance of the f distribution are
Mean =
Variance =
1
-v2 - 2
2v12 ( v1 + v2 2 )
v1 ( v2 2 ) ( v2 4 )
2
for 2 >2
for 2 > 4
3.
The f distribution is positively skewed and its skewness decreases with increases
4.
in 1 and 2
The value of f must be positive or zero since variances are squares and can
never assume negative values
Assumptions
a) All sample observations are randomly selected and independent
b) The total variance of the various sources of variance should be additive.
c) The ratio of S12 to S22 should be equal to or greater than 1
d) The population for each sample must be normally distributed with identical mean of
variance
e) F value can never be negative
182
Probability
183

QUESTION ONE
The quality controller, Mr. Brooks, at Queensville Engineers has become aware of the need for
an acceptance sampling programme to check the quality of bought-in components. This is of
particular importance for a problem the company is currently having with batches of pump
shafts bought in from a local supplier. Mr. Brooks proposes the following criteria to assess
whether or not to accept a large batch of pump shafts from this supplier.
From each batch received take a random sample of 50 shafts, and accept that batch if no more
than two defectives are found in the sample.
Mr. Brooks needed to calculate the probability of accepting a batch Pa , when the proportion of
defectives in the batch, p, is small (under 10%, say)
Required:
a)
Explain why the Poisson distribution is appropriate to invstigate this situation.
Using the Poisson distribution, determine the probability of accepting a batch Pa,
b)
containing p=2% defectives if the method is used.
Determine Pa, for p = 0%, 2%, 5%, 10%, 15%
QUESTION TWO
A woven cloth is liable to contain faults and is subjected to an inspection procedure. Any fault
has a probability of 0.7 that it will be detected by the procedure, independent of whether any
other fault is detected or not.
Required:
a) If a piece of cloth contains three faults, A, B and C,
Calculate the probability that A and C are detected, but that B is undetected;
i)
ii)
Calculate the probability that any two of A, B and C be detected, the other fault
being undetected;
State the relationship between your answers to parts (i) and (ii) and give reasons for
iii)
this.
b) Suppose now that, in addition to the inspection procedure given above, there is a secondary
check which has a probability of 0.6 of detecting each fault missed by the first inspection
procedure. This probability of 0.6 applies independently to each and every fault undetected
by the first procedure.
i)
Calculate the probability that a piece of cloth with one fault has this fault
undetected by both the inspection procedure and the secondary check;
ii)
Calculate the probability that a piece of cloth with two faults has one of these faults
detected by either the inspection procedure or the secondary check, and one fault
undetected by both;
Of the faults detected, what proportion are detected by the inspection procedure
iii)
and what proportion by the secondary check?
184
Lesson Five
QUESTION THREE
A company has three production sections S1,S2 and S3 which contribute 40%,35% and 25%,
respectively, to total output. The following percentages of faulty units have been observed:
S1
2%
(0.02)
S2
3%
(0.03)
S3
4%
(0.04)
There is a final check before output is dispatched. Calculate the probability that a unit found
faulty at this check has come from section 1, S1
QUESTION FOUR
Assuming a Binomial Distribution what is the probability of a salesman making 0,1,2,3,4,5 or 6
sales in 6 visits if the probability of making a sale on a visit is 0.3?
Do not use tables for this question.
QUESTION FIVE
Records show that 60% of students pass their examinations at first attempt. Using the normal
approximation to the binomial, calculate the probability that at least 65% of a group of 200
students will pass at the first attempt.
QUESTION SIX
A batch of 5000 electric lamps has a mean life of 1000 hours and a standard deviation of 75
hours. Assume a normal distribution.
a) How many lamps will fail before 900 hours?
How many lamps will fail between 950 and 1000 hours?
b)
c) What proportion of lamps will fail before 925 hours?
d)Given the same mean life, what would the standard deviation have to be to ensure that
no more than 20% of lamps fail before 916 hours?
184
Sampling and Estimation
185
LESSON SIX
Sampling techniques
Central limit theorem
Sampling distribution of statistical parameters
Test of hypothesis
6.1
Methods of Sampling
a . Random or probability sampling methods
they include
Simple random sampling
i.
ii.
Stratified sampling
Systematic sampling
iii.
Multi stage sampling
iv.
b. Non random probability sampling methods
these consist of
i.
ii.
iii.
Judgment sampling
Quota sampling
Cluster sampling
Simple Random Sampling

This refers to the sampling technique in which each and every item of the population is given an
equal chance of being included in the sample. Since selection of items in the sample depends
entirely on chance, this method is also called chance selection or representative sampling.
It is assumed that if the sample is chosen at random and if the size of the sample is sufficiently
large, it will represent all groups in the population
Random sampling is of 2 types; sampling with replacement and sampling without replacement
Sampling is said to be with replacement when from a finite population a sampling unit is drawn
observed and then returned to the population before another unit is drawn. The population in
this case remains the same and a sampling unit might be selected more than once
If on the other hand a sampling unit is chosen and not retuned to the population after it has
been observed the sampling is said to be without replacement.
Random samples may be selected by the help of lottery method or table of random numbers
(such as tippets table of random numbers, fischer and Yates numbers or Kendall and Babington
Smith numbers.)
Stratified sampling
In this case the population is divided into groups in such a way that units within each group are
as similar as possible in a process called stratification. The groups are called strata. Simple
random samples from each of the strata are collected and combined into a simple. This
technique of collecting a sample from a population is called stratified sampling. Stratification
may be by age, occupation income group e.t.c.
186
Lesson Six
Systematic Sampling
This sampling is a part of simple random sampling in ascending or descending orders. In
systematic sampling a sample is drawn according to some predetermined object. Suppose a
population consists of 1000 units, then every tenth, 20th or 50th item is selected. This method is
very easy and economical. It also saves a lot of time
Multistage sampling
This is similar to stratified sampling except division is done on geographical/location basis, e.g. a
country can be divided into provinces and then survey is done in 4 towns in each province. This
helps to cut traveling costs for a surveyor.
Cluster Sampling
This is where a few geographical regions e.g. a location, town or village are selected at random
and say every single household or shop in that area is interviewed. This again cuts on costs.
Judgment Sampling
Here the interviewer selects whom to interview believing that their view is more fundamental
since they might be directly affected e.g. to find out effects of public transport one may chose to
interview only people who dont own cars and travel frequently to work.
6.2
THE CENTRAL LIMIT THEOREM
The theory was introduced by De Moivre and according to it; if we select a large number of
simple random samples, say from any population and determine the mean of each
sample, the distribution of these sample means will tend to be described by the normal
probability distribution with a mean and variance 2/n. This is true even if the population
itself is not normal distribution. Or the sampling distribution of sample means approaches to a
normal distribution irrespective of the distribution of population from where the sample is taken
and approximation to the normal distribution becomes increasingly close with increase in sample
sizes
Types of distribution
Population distribution
It refers to the distribution of the individual values of population. Its mean is denoted by
Sample distribution
It is the distribution of the individual values of a single sample. Its mean is generally written as
x . it is not usually the same as
Distribution of Sample Means or sampling distribution
A sample of size n is taken from the parent population and mean of the sample is calculated.
This is repeated for a number of samples so that we have a distribution of sample means, which
approaches a normal distribution.
Standard errors of the mean
The series of sample means X 1 , X 2 , X 3 .. is normally distributed or nearly so (according
to the central limit theorem). It can be described by its mean and its standard deviation. This
standard deviation is known as the standard error.
186
187
Standard error of the mean = S x =
s
n
Note: this formula is satisfactory for larger samples and a large population i.e. n > 30 and n >
5% of N.
- The word error is in place of deviation to emphasize that variation among sample means
is due to sampling errors.
- The smaller the standard error the greator the precision of the sample value.
6.3
Statistical inference
It is the process of drawing conclusions about attributes of a population based upon information
contained in a sample (taken from the population).
It is divided into estimation of parameters and testing of hypothesis. Symbols for statistic of
population parameters are as follows.
Arithmetic mean
Standard deviation
Number of items
Sample Statistic
x
s
n
Population Parameter
Statistical estimation
It is the procedure of using statistic to estimate a population parameter
It is divided into point estimation (where an estimate of a population parameter is given by a
single number) and interval estimation (where an estimate of a population is given by a range in
which the parameter may be considered to lie) e.g. a bus meant to take a class of 100 students
(population N) for trip has a limit to the maximum weight of 600kg of which it can carry, the
teacher realizes he has to find out the weight of the class but without enough time to weigh
everyone he picks 25 students selected at random (sample n = 25). These students are weighed
and their average weight recorded as 64kg ( X - mean of a sample) with a standard deviation (s),
now using this the teacher intends to estimate the average weight of the whole class (
population mean) by using the statistical parameters standard deviation (s), and mean of the
sample ( x ).
Characteristic of a good estimator

(i)
Unbiased: where the expected value of the statistic is equal to the population
parameter e.g. if the expected mean of a sample is equal to the population mean
(ii)
Consistency: where an estimator yields values more closely approaching the
population parameter as the sample increases
Efficiency: where the estimator has smaller variance on repeated sampling.
(iii)
Sufficiency: where an estimator uses all the information available in the data
(iv)
concerning a parameter
Confidence Interval
The interval estimate or a confidence interval consists of a range (an upper confidence limit and
lower confidence limit) within which we are confident that a population parameter lies and we
assign a probability that this interval contains the true population value
The confidence limits are the outer limits to a confidence interval. Confidence interval is the
interval between the confidence limits. The higher the confidence level the greater the
confidence interval. For example
A normal distribution has the following characteristic
Sample mean 1.960 includes 95% of the population
i.
188
Lesson Six
ii.
Sample mean 2.575 includes 99% of the population
1.
LARGE SAMPLES
These are samples that contain a sample size greater than 30(i.e. n>30)
(a)
Estimation of population mean
Here we assume that if we take a large sample from a population then the mean of the
population is very close to the mean of the sample
Steps to follow to estimate the population mean includes
i.
Take a random sample of n items where (n>30)
ii.
Compute sample mean ( X ) and standard deviation (S)
Compute the standard error of the mean by using the following formular
iii.
s
Sx =
n
where S x = Standard error of mean
S = standard deviation of the sample
n = sample size
iv.
Choose a confidence level e.g. 95% or 99%
Estimate the population mean as under
v.
Population mean = (appropriate number) S x
Appropriate number means confidence level e.g. at 95% confidence level is 1.96
this number is usually denoted by Z and is obtained from the normal tables.
Example
The quality department of a wire manufacturing company periodically selects a sample of wire
specimens in order to test for breaking strength. Past experience has shown that the breaking
strengths of a certain type of wire are normally distributed with standard deviation of 200 kg. A
random sample of 64 specimens gave a mean of 6200 kgs. Find out the population mean at 95%
level of confidence
Solution
Population mean = 1.96 S x
Note that sample size is alredy n > 30 whereas s and x are given thus step i), ii) and iv) are
provided.
Here: X = 6200 kgs
s
200
Sx =
=
=
25
n
64
Population mean
= 6200 1.96(25)
= 6200 49
= 6151 to 6249
At 95% level of confidence, population mean will be in between 6151 and 6249
FINITE POPULATION CORRECTION FACTOR (FPCF)

If a given population is relatively of small size and sample size is more than 5% of the
population then the standard error should be adjusted by multiplying it by the finite population
correction factor
188
189
FPCF is given by
N n
n 1
where N = population size

n = sample size
Example
A manager wants an estimate of sales of salesmen in his company. A random sample 100 out of
500 salesmen is selected and average sales are found to be Shs. 75,000. if a sample standard
deviation is Shs. 15000 then find out the population mean at 99% level of confidence
Solution
Here N = 500, n = 100, X = 75000 and S = 15000
Now
Standard error of mean
s
N n
= Sx =
x
n 1
n
(500 100 )
(500 1)
15000
x
100
15000
400
x
499
10
15000
=
(0.895)
10
Sx
= 1342.50 at 99% level of confidence

= X 2.58 S x
Population mean
=shs 75000 2.58(1342.50)

=shs 75000 3464
= Shs 71536 to 78464
b)
Estimation of difference between two means
We know that the standard error of a sample is given by the value of the standard deviation
()divided by the square root of the number of items in the sample (
But, when given two samples, the standard errors is given by
S (X
AX B
)=
n ).
S A2 S B2
+
n A nB
Also note that we do estimate the interval not from the mean but from the difference between
the two sample means i.e. X A X B .
The appropriate number of confidence level does not change
Thus the confidence interval is given by;
X A X B Confidence level S (X X )
A B
= X A X B Z S (X X )
A B
190
Lesson Six
Example
Given two samples A and B of 100 and 400 items respectively, they have the means X 1 = 7 ad
X 2 = 10 and standard deviations of 2 and 3 respectively. Construct confidence interval at 70%

confidence level?
Solution
Sample A
X1 = 7
X 2 = 10
n2 = 400
n1 = 100
S2 = 3
S1 = 2
The standard error of the samples A and B is given by
S (X
AX B )
25
=
400
4
9
+
100 400
5
20
0.25
At 70% confidence level, then appropriate number is equal to 1.04 (as read from the normal
tables)
X 1 X 2 = 7 10 = - 3 = 3
We take the absolute value of the difference between the means e.g. the value of X = absolute
value of X i.e. a positive value of X.
Confidence interval is therefore given by
= 3 1.04 (0.25 )
From the normal tables a z value of 1.04 gives a value of 0.7.
= 3 0.26
= 3.26 and 2.974
Thus 2.974 X 3.26
Example 2
A comparison of the wearing out quality of two types of tyres was obtained by road testing.
Samples of 100 tyres were collected. The miles traveled until wear out were recorded and the
results given were as follows
Tyres
T1
T2
X 1 = 26400 miles
X 2 = 25000 miles
Mean
Variance
S21= 1440000 miles
S22= 1960000 miles
Find a confidence interval at the confidence level of 70%
Solution
X 1 = 26400
X 2 = 25000
Difference between the two means
(X
X 2 = (26400 25000)
190
191
= 1,400
Again we take the absolute value of the difference between the two means
We calculate the standard error as follows
S (X
AX B )
S12 S 22
+
n1 n2
1, 440, 000 1,960, 000

+
100
100
=
184.4
Confidence level at 70% is read from the normal tables as 1.04 (Z = 1.04).
Thus the confidence interval is calculated as follows
= 1400 (1.04) (184.4)
= 1400 191.77
or (1400 191.77) to (1400 + 191.77)
1,208.23 X 1591.77
c) Estimation of population proportions

This type of estimation applies at the times when information cannot be given as a mean or as a
measure but only as a fraction or percentage
The sampling theory stipulates that if repeated large random samples are taken from a
population, the sample proportion p will be normally distributed with mean equal to the
population proportion and standard error equal to
Sp =
Pq
= Standard error for sampling of population proportions
n
Where n is the sample size and q = 1 p.

The procedure for estimating a proportion is similar to that for estimating a mean, we only have
a different formula for calculating standard.
Example 1
In a sample of 800 candidates, 560 were male. Estimate the population proportion at 95%
confidence level.
Solution
Here
Sample proportion (P) =
560
= 0.70
800
q = 1 p = 1 0.70 = 0.30
n = 800
pq
=
n
( 0.70 )( 0.30 )
800
Sp = 0.016
population proportion
192
Lesson Six
= P 1.96 Sp where 1.96 = Z.
= 0.70 1.96 (0.016)
= 0.70 0.03
= 0.67 to 0.73
= between 67% to 73%
Example 2
A sample of 600 accounts was taken to test the accuracy of posting and balancing of accounts
where in 45 mistakes were found. Find out the population proportion. Use 99% level of
confidence
Solution
Here
n = 600; p =
45
= 0.075
600
q = 1 0.075 = 0.925
Sp =
pq
=
n
( 0.075 )( 0.925 )
600
= 0.011
Population proportion
= P 2.58 (Sp)
= 0.075 2.58 (0.011)
= 0.075 0.028
= 0.047 to 0.10
= between 4.7% to 10%
d) Estimation of difference between population proportions

Let the two proportions be given by P1 and P2, respectively
Then the difference (absolute) between the two proportions is given by (P1 P2)
The standard error is given by
S (P
1 P2 )
pq pq
p n + p2 n2
+
where p = 1 1
and q = 1 - p
n1 n2
n1 + n2
Then given the confidence level, the confidence interval between the two population
proportions is given by
(P1 P2) Confidence level S (P P )
1 2
= (P1 P2) Z
pq pq
+
n1 n2
192
Where P =
193
p1n1 + p2 n2
always remember to convert P1 & P2 to P.
n1 + n2
2.
SMALL SAMPLES
(a)
Estimation of population mean
If the sample size is small (n<30) the arithmetic mean of small samples are not normally
distributed. In such circumstances, students t distribution must be used to estimate the
population mean.
In this case
Population mean = X tsx
X = Sample mean
s
Sx =
n
S = standard deviation of samples =
( x x)
n 1
for small samples.
n = sample size
v = n 1 degrees of freedom.
The value of t is obtained from students t distribution tables for the required confidence level
Example
A random sample of 12 items is taken and is found to have a mean weight of 50 grams and a
standard deviation of 9 grams
What is the mean weight of population
a) with 95% confidence
b) with 99% confidence
Solution
X = 50; S = 9; v = n 1 = 12 1 = 11;
Sx =
s
9
=
n
12
= x tsx
At 95% confidence level
12
= 50 2.262
= 50 5.72 grams
Therefore we can state with 95% confidence that the population mean is between 44.28 and
55.72 grams
At 99% confidence level
12
= 50 3.25
= 50 8.07 grams
194
Lesson Six
Therefore we can state with 99% confidence that the population mean is between 41.93 and
58.07 grams
Note: To use the t distribution tables it is important to find the degrees of freedom (v = n 1).
In the example above v = 12 1 = 11
From the tables we find that at 95% confidence level against 11 and under 0.05, the value of t =
2.201
6.4
Hypothesis Testing
Definition
- A hypothesis is a claim or an opinion about an item or issue. Therefore it has to be tested
statistically in order to establish whether it is correct or not correct
- Whenever testing an hypothesis, one must fully understand the 2 basic hypothesis to be tested
namely
i.
The null hypothesis (H0)
The alternative hypothesis(H1)
ii.
The null hypothesis
This is the hypothesis being tested, the belief of a certain characteristic e.g. Kenya Bureau of
Standards (KBS) may walk to a sugar making company with an intention of confirming that the
2kgs bags of sugar produced are actually 2kgs and not less, they conduct hypothesis testing with
the null hypothesis being: H0 = each bag weighs 2kgs. The testing will set out to confirm this or
to refute it.
The alternative hypothesis
While formulating a null hypothesis we also consider the fact that the belief might be found to
be untrue hence we will reject it. We therefore formulate an alternative hypothesis which is a
contradiction to the null hypothesis, thus when we reject the null hypothesis we accept the
alternative hypothesis.
In our example the alternative hypothesis would be
H1 = each bag does not weigh 2kg
Acceptance and rejection regions
All possible values which a test statistic may either assume consistency with the null hypothesis
(acceptance region) or lead to the rejection of the null hypothesis (rejection region or critical
region)
The values which separate the rejection region from the acceptance region are called critical
values
Type I and type II errors
While testing hypothesis (H0) and deciding to either accept or reject a null hypothesis, there are
four possible occurrences.
a) Acceptance of a true hypothesis (correct decision) accepting the null hypothesis and it
happens to be the correct decision. Note that statistics does not give absolute information,
thus its conclusion could be wrong only that the probability of it being right are high.
b) Rejection of a false hypothesis (correct decision).
c) Rejection of a true hypothesis (incorrect decision) this is called type I error, with
probability = .
d) Acceptance of a false hypothesis (incorrect decision) this is called type II error, with
probability = .
194
195
Levels of significance
A level of significance is a probability value which is used when conducting tests of hypothesis.
A level of significance is basically the probability of one making an incorrect decision after the
statistical testing has been done. Usually such probability used are very small e.g. 1% or 5%
0.5000
0.4900
1% provision for errors
0
Critical value
0.45
5% = 0.05
Critical region
0
Crititical value = -1.65
NB: If the standardized value of the mean is less than 1.65 we reject the null hypothesis (H0)
and accept the alternative Hypothesis (H1) but if the standardized value of the mean is more
than 1.65 we accept the null hypothesis and reject the alternative hypothesis
The above sketch graph and level of significance are applicable when the sample mean is < (i.e.
less than the population mean)
196
Lesson Six
The following is used when sample mean > population mean
Acceptance region
Critical region (rejection region)
5% = 0.05
Z = 1.65 (critical value)
NB: If the sample mean standardized value < 1.65, we accept the null hypothesis but reject the
alternative. If the sample mean value > 1.65 we reject the null hypothesis and accept the
alternative hypothesis
The above sketch is normally used when the sample mean given is greater than the population
mean
Accept null hyp( reject Alternative hyp)
Reject null hyp (accept alt hyp)
0.05% = 0.05
Reject null hyp (accept alt hyp)
0.495
0.495
-2.58
0.5% = 0.05
+2.58
NB: if the standardized value of the sample mean is between 2.58 and +2.58 accept the null
hypothesis but otherwise reject it and therefore accept the alternative hypothesis
196
197
TWO TAILED TESTS

A two tailed test is normally used in statistical work(tests of significance) e.g. if a complaint
lodged by the client is about a product not meeting certain specifications i.e. the item will
generate a complaint if its measurements are below the lower tolerance limit or above the
upper tolerance limit
Region of acceptance for
H0
Critical region
Critical region
15cm
17 cm
NB: Alternative hypothesis is usually rejected if the standardized value of the sample mean lies
beyond the tolerance limits (15cm and 17 cm).
ONE TAILED TEST

This is a test where the alternative hypothesis (H1:) is only concerned with one of the tails of the
distribution e.g. to test a business complaint if the complaint is above the measurements of item
being shorter than is required.
E.g. a manufacturer of a given brand of bread may state that the average weight of the bread is
500 gms but if a consumer takes a sample and weighs each of the pieces of bread and happens
to have a mean of 450 gms he will definitely complain about the bread which is underweight.
The statistical analysis to be done will concentrate on the left tail of the normal distribution in
which one will have to establish whether 450 gms being less than 500g is statistically significant.
Such a test therefore is referred to as one tailed test.
198
Lesson Six
left
On the other hand the test may compuliate on the right hand tail of the normal distribution
when this happens the major complaint is likely to do with oversize items bought. Therefore the
test is known as one tailed as the focus is on one end of the normal distribution.
5% level of significance
1% level of significance
Number of standard errors

Two tailed test
One tailed test
1.96
1.65
2.58
2.33
HYPOTHESIS TESTING PROCEDURE

Whenever a business complaint comes up there is a recommended procedure for conducting a
statistical test. The purpose of such a test is to establish whether the null hypothesis or
alternative hypothesis is to be accepted.
The following are steps normally adopted
1. Statement of the null and alternative hypothesis
2. Statement of the level of significance to be used.
3. Statement about the test statistic i.e. what is to be tested e.g. the sample mean, sample
proportion, difference between sample means or sample proportions
4. Type of test whether two tailed or one tailed.
5. Statement on critical values using the appropriate level of significance
6. Standardizing the test statistic
7. Conclusion showing whether to accept or reject the null hypothesis
STANDARD HYPOTHESIS TESTS
In principal, we can test the significance of any statistic related to any probability distribution.
However we will be interested in a few standard cases. The sample statistics mean, proportion
and variance, are related to the normal, t, F, and chi squared distributions
Thus
198
199
1. Normal test
Test a sample mean ( X ) against a population mean () (where samples size n > 30 and
population variance 2 is known) and sample proportion, P(where sample size np >5 and nq
>5 since in this case the normal distribution can be used to approximate the binomial
distribution
2. t test
Tests a sample mean ( X ) against a population mean and especially where the population
variance is unknown and n < 30.
3. Variance ratio test or f test

It is used to compare population variances and it is used with samples of any size drawn
from normal populations.
4. Chi squared test
It can be used to test the association between attributes or the goodness of fit of an
observed frequency distribution to a standard distribution
Example 1
A certain NGO carried out a survey in a certain community in order to establish the average at
which the girls are married. The results of the survey indicated that the marriage age for the girls
is 19 years
In order to establish the validity of the mean marital age, a sample of 50 women was interviewed
and the average age indicated that they got married at the age of 16 years. However the different
ages at which they were married differed with the standard deviation of 2.1years
The sample data indicates that the marital age is less 19 years. Is this conclusion true or not ?
Required
Conduct a statistical test to either support the above conclusion drawn from the sample statistics
i.e. the marriage age is less than 19 years, use a level of significance of 5%
Solution
1. Null hypothesis
H0: (mean marital age) = 19 years
Alternative hypothesis H1: (mean marital age) < 19 years
2. The level of significance is 5%
3. The test statistics is the sample mean age, X = 16 years
4. The critical value of the one tailed test (one tailed because the alternative hypothesis is
an inequality) at 5% level of significance is 1.65
200
Lesson Six
Acceptance region
Rejection region
- 1.65
5. The standardizes value of the sample mean is

X -
S
where S x =
Z
=
Sx
n
Where,
X = Sample mean
= Population mean
S = sample standard deviation
n = sample size
z = standard value (as per computation)
The standard value Z must fall within the acceptance region for us to accept the null
hypothesis. Thus it must be > - 1.65 otherwise we accept the alternative hypothesis.
16 19
2.1
50
= - 10.1
6. Since 10.1 < -1.65, we reject the null hypothesis but accept the alternative hypothesis
at 5% level of significance i.e. the marriage age in this community is significantly lower
than 19 years
Example 2
A foreign company which manufactures electric bulbs has assured its customers that the lifespan
of the bulbs is 28 month with a standard deviation of 4months
Recently the company embarked on a quality improvement research for their product. After the
research using new technology, a sample of 70 bulbs was tested and they gave a mean lifespan of
30.2 months
Does this justify the research undertaken? Use 1% level of significance to conduct a statistical
test in order to establish the truth about the above question.
Testing procedure
1. Null hypothesis H0: = 28
Alternative hypothesis H1: > 28
2. The level of significance is 1% (one tailed test)
200
201
3. The test statistics is the sample mean age, x = 30.2

4. The critical value of the one tailed test at 5% level of significance is + 2.33
0.4900
1% = 0.01
2.33
5. The standardized value of the sample mean is
Z
X
Sx
30.2 28
4
70
= 4.6
6. Since 4.6 > 2.33, we reject the null hypothesis but accept the alternative hypothesis at
1% level of significance i.e. the new sample mean life span is statistically significant
higher than the population mean
Therefore the research undertaken was worth while or justified
Example 3
A construction firm has placed an order that they require a consignment of wires which have a
mean length of 10.5 meters with a standard deviation of 1.7 m
The company which produces the wires delivered 90 wires, which had a mean length of 9.2 m.,
The construction company rejected the consignment on the grounds that they were different
from the order placed.
Required
Conduct a statistical test to indicate whether you support or not support the action taken by the
construction company at 5% level of significance.
Solution
Null hypothesis = 10.5 m
Alternative hypothesis 10.5 m
Level of significance be 5%
The test statistics is the sample mean X = 9.2m
The critical value of the two tailed test at 5% level of significance is 1.96 (two tailed test).
202
Lesson Six
- 1.96
+1.96
The standardized value of the test Z =

Z
X -
SX
9.2 10.5
1.7
90
= - 7.25
Since 7.25 < 1.96, reject the null hypothesis but accept the alternative hypothesis at 5% level
of significance i.e. the sample mean is statistically different from the consignment ordered by
the construction company. Therefore support the action taken by the construction company
TESTING THE DIFFERENCE BETWEEN TWO SAMPLE MEANS (LARGE

SAMPLES)
A large sample is defined as one which contains 30 or more items (n30) Where n is the sample
size
In a business those involved are constantly observant about the standards or specifications of
the item which they sell e.g. a trader may receive a batch of items at one time and another batch
at a later time at the end he may have concluded that the two samples are different in certain
specifications e.g. mean weight mean lifespan, mean length e.t.c. further it may become
necessary to establish whether the observed differences are statistically significant or not. If the
differences are statistically significant then it means that such differences must be explained i.e.
there are known causes but if they are not statistically significant then it means that the
difference observed have no known causes and are mainly due to chance
If the differences are established to be statistically significant then it implies that the complaints,
which necessitated that kind of test, are justified
Let X1 and X2 be any two samples whose sizes are n1 and n2 and mean X 1 and X 2. Standard
deviation S1 and S2 respectively. In order to test the difference between the two sample means,
we apply the following formulas
Z
X1 X 2
S X1 X 2
where S X 1 X 2 =
S12 S 22
+
n1 n2
Example 1
An agronomist was interested in the particular fertilizer yield output. He planted maize on 50
equal pieces of land and the mean harvest obtained later was 60 bags per plot with a standard
202
203
deviation of 1.5 bags. The crops grew under natural circumstances and conditions without the
soil being treated with any fertilizer. The same agronomist carried out an alternative experiment
where he picked 60 plots in the same area and planted the same plant of maize but a fertilizer
was applied on these plots. After the harvest it was established that the mean harvest was 63
bags per plot with a standard deviation of 1.3 bags
Required
Conduct a statistical test in order to establish whether there was a significant difference between
the mean harvest under the two types of field conditions. Use 5% level of significance.
Solution
H 0 : 1 = 2
H 1 : 1 2
Critical values of the two tailed test at 5% level of significance are 1.96
The standardized value of the difference between sample means is given by Z where
=
=
- 1.96
X1 X 2
S X1 X 2
where S X 1 X 2 =
1.52 1.32
+
50
60
( 60 63)
0.045 + 0.028
11.11
+1.96
Since 11.11 < -1.96, we reject the null hypothesis but accept the alternative hypothesis at 5%
level of significance i.e. the difference between the sample mean harvest is statistically significant.
This implies that the fertilizer had a positive effect on the harvest of maize
Note: You dont have to illustrate your solution with a diagram.
204
Lesson Six
Example 2
An observation was made about reading abilities of males and females. The observation lead to a
conclusion that females are faster readers than males. The observation was based on the times
taken by both females and males when reading out a list of names during graduation ceremonies.
In order to investigate into the observation and the consequent conclusion a sample of 200 men
were given lists to read. On average each man took 63 seconds with a standard deviation of 4
seconds
A sample of 250 women were also taken and asked to read the same list of names. It was found
that they on average took 62 seconds with a standard deviation of 1 second.
Required
By conducting a statistical hypothesis testing at 1% level of significance establish whether the
sample data obtained does support earlier observation or not
Solution
H0: 1 = 2
H1: 1 2
Critical values of the two tailed test is at 1% level of significance is 2.58.
Z
X1 X 2
S X1 X 2
63 62
42
200
1
+ 250
)
=
3.45
Acceptance region
Rejection region
- 2.58
+2.58
+3.45
Since 3.45 > 2.33 reject the null hypothesis but accept the alternative hypothesis at 1% level
of significance i.e. there is a significant difference between the reading speed of Males and
females, thus females are actually faster readers.
TEST OF HYPOTHESIS ON PROPORTIONS

This follows a similar method to the one for means exept that the standard error used in this
case:
204
Sp =
205
Pq
n
Z score is calculated as, Z =
P
Sp
Where P = Proportion found in the sample.

the hypothetical proportion.
Example
A member of parliament (MP) claims that in his constituency only 50% of the total youth
population lacks university education. A local media company wanted to acertain that claim thus
they conducted a survey taking a sample of 400 youths, of these 54% lacked university
education.
Required:
At 5% level of significance confirm if the MPs claim is wrong.
Solution.
Note: This is a two tailed tests since we wish to test the hypothesis that the hypothesis is
different () and not against a specific alternative hypothesis e.g. < less than or > more
than.
H0 : = 50% of all youth in the constituency lack university education.
H1 : 50% of all youth in the constituency lack university education.
pq
0.5 x0.5
=
= 0.025
n
400
0.54 0.50
= 1.6
Z=
0.025
Sp =
at 5% level of significance for a two-tailored test the critical value is 1.96 since calculated Z value
< tabulated value (1.96).
i.e. 1.6 < 1.96 we accept the null hypothesis.
Thus the MPs claim is accurate.
HYPOTHESIS TESTING OF THE DIFFERENCE BETWEEN PROPORTIONS

Example
Ken industrial manufacturers have produced a perfume known as fianchetto. In order to test
its popularity in the market, the manufacturer carried a random survey in Back rank city where
10,000 consumers were interviewed after which 7,200 showed preference. The manufacturer
also moved to area Rook town where he interviewed 12,000 consumers out of which 1,0000
showed preference for the product.
Required
Design a statistical test and hence use it to advise the manufacturer regarding the differences in
the proportion, at 5% level of significance.
Solution
H0 : 1 = 2
H1 : 1 2
206
Lesson Six
The critical value for this two tailed test at 5% level of significance = 1.96.
Now Z =
( P1 P2 ) ( 1 2 )
S ( P1 P2 )
But since the null hypothesis is 1 = 2, the second part of the numerator disappear i.e.
1 - 2 = 0 which will always be the case at this level.
Then Z =
( P1 P2 )
S ( P1 P2 )
Where;
Sample size
Sample proportion of success
Population proportion of success.
Now
Sample 2
n2 = 12,000
P2 = 0.83
2
pq pq
+
n1 n2
S ( p1 p2 ) =
Where P =
Sample 1
n1 = 10,000
P1 =0.72
1
p1n1 + p2 n2
n1 + n2
And q = 1 p
in our case
P=
10, 000(0.72) + 12, 000(0.83)

10, 000 + 12, 000
84, 000
=
22, 000
= 0.78
q = 0.22
S ( P1 P2 ) =
0.78 ( 0.22 )
10, 000
0.78 ( 0.22 )
12, 000
= 0.00894
Z=
0.72 0.83
0.00894
12.3
Since 12.3 > 1.96, we reject the null hypothesis but accept the alternative. the differences
between the proportions are statistically significant. This implies that the perfume is much
more popular in Rook town than in Back rank city.
HYPOTHESIS TESTING ABOUT THE DIFFERENCE BETWEEN TWO

PROPORTIONS
Is used to test the difference between the proportions of a given attribute found in two random
samples.
206
207
The null hypothesis is that there is no difference between the population proportions. It means
two samples are from the same population.
Hence
H0 : 1 = 2
The best estimate of the standard error of the difference of P1 and P2 is given by pooling the
samples and finding the pooled sample proportions (P) thus
P=
p1n1 + p2 n2
n1 + n2
Standard error of difference between proportions
S ( p1 p2 ) =
pq pq
+
n1 n2
P1 P2
S ( p1 p2 )
And Z =
Example
In a random sample of 100 persons taken from village A, 60 are found to be consuming tea. In
another sample of 200 persons taken from a village B, 100 persons are found to be consuming
tea. Do the data reveal significant difference between the two villages so far as the habit of
taking tea is concerned?
Solution
Let us take the hypothesis that there is no significant difference between the two villages as far
as the habit of taking tea is concerned i.e. 1 = 2
We are given
n1 = 100
P1 = 0.6;
n2 = 200
P2 = 0.5;
Appropriate statistic to be used here is given by
P
=
=
p1n1 + p2 n2
n1 + n2
( 0.6 )(100 ) + ( 0.5 )( 200 ) = 60 + 100

100 + 200
300
= 0.53
= 1 0.53
= 0.47
S ( P1 P2 ) =
=
pq pq
+
n1 n2
( 0.53)( 0.47 ) + ( 0.53)( 0.47 )

100
200
= 0.0608
0.6 0.5
Z=
0.0608
= 1.64
208
Lesson Six
Since the computed value of Z is less than the critical value of Z = 1.96 at 5% level of
significance therefore we accept the hypothesis and conclude that there is no significant
difference in the habit of taking tea in the two villages A and B
t distribution (students t distribution) tests of hypothesis (test for small samples n < 30)
For small samples n < 30, the method used in hypothesis testing is exactly similar to the one for
large samples exept that t values are used from t distribution at a given degree of freedom v,
instead of z score, the standard error Se statistic used is also different.
Note that v = n 1 for a single sample and n1 + n2 2 where two sample are involved.
a) Test of hypothesis about the population mean
When the population standard deviation (S) is known then the t statistic is defined as
t
X
SX
where S X =
S
n
Follows the students t distribution with (n-1) d.f. where
X = Sample mean
= Hypothesis population mean
n = sample size
and S is the standard deviation of the sample calculated by the formula
S=
( X X )
n 1
for n < 30
If the calculated value of t exceeds the table value of t at a specified level of significance, the null
hypothesis is rejected.
Example
Ten oil tins are taken at random from an automatic filling machine. The mean weight of the tins
is 15.8 kg and the standard deviation is 0.5kg. Does the sample mean differ significantly from the
intended weight of 16kgs. Use 5% level of significance.
Solution
Given that n = 10; x = 15.8; S = 0.50; = 16; v = 9
H0 : = 16
H1 : 16
0.5
10
= SX =
t
=
=
15.8 16
0.5
10
0.2
0.16
= -1.25
The table value for t for 9 d.f. at 5% level of significance is 2.26. the computed value of t is
smaller than the table value of t. therefore, difference is insignificant and the null hypothesis is
accepted.
208
209
b) Test of hypothesis about the difference between two means

The t test can be used under two assumptions when testing hypothesis concerning the difference
between the two means; that the two are normally distributed (or near normally distributed)
populations and that the standard deviation of the two is the same or at any rate not significantly
different.
Appropriate test statistic to be used is
t
X1 X 2
S X X 2
( 1 )
at (n1 + n2 2) d.f.
The standard deviation is obtained by pooling the two sample standard deviation as shown
below.
( n1 1) S12 + ( n2 1) S22
Sp =
n1 + n2 2
Where S1 and S2 are standard deviation for sample 1 & 2 respectively.

Now S X 1 =
Sp
Sp
and S X 2 =
n1
n2
S X1X 2 =
(
)
S X2 + S X2 2
Alternatively S
( X1X 2 )
= Sp
n1 + n2
n1n2
Example
Two different types of drugs A and B were tried on certain patients for increasing weights, 5
persons were given drug A and 7 persons were given drug B. the increase in weight (in pounds)
is given below
Drug A
Drug B
8
10
12
8
16
12
9
15
3
6
11
Do the two drugs differ significantly with regard to their effect in increasing weight? (Given that
v= 10; t0.05 = 2.23)

Solution
H 0 : 1 = 2
H 1 : 1 2
t=
X1 X 2
S X1 X 2
(
)
Calculate for X 1 , X 2 and S
210
Lesson Six
X1
X2
8
12
13
9
3
X1 X 1
-1
+3
+4
0
-6
(X1 X 1 )2
1
9
16
0
36
X1 = 45
(X1 X 1 ) = 0
(X1 X 1 )2= 62
X1 =
S1 =
Sp =
X
n1
45
=9
5
X2 =
62
= 3.94
4
S2 =
X
n2
10
8
12
15
6
8
11
X2= 70
(X2 X 2 )
0
-2
+2
+5
-4
-2
+1
(X2 X )2
0
4
4
25
16
4
1
(X2 X 2 ) = 0
(X2 X 2 )2= 54
70
= 10
7
54
=3
6
( 4 )15.4 + ( 6 ) 9
10
= 3.406
S( X 1 X 2 ) =
11.6 11.6
7+5
+
or 3.406
5
7
5(7)
= 1.99
t
X1 X 2
S X1 X 2
(
)
9 10
1.99
= 0.50
Now t0.05 (at v = 10) = 2.23 > 0.5
Thus we accept the null hypothesis.
Hence there is no significant difference in the efficacy of the two drugs in the matter of
increasing weight
Example
Two salesmen A and B are working in a certain district. From a survey conducted by the head
office, the following results were obtained. State whether there is any significant difference in the
average sales between the two salesmen at 5% level of significance.
210
211
A
20
170
20
No. of sales
Average sales in shs
Standard deviation in shs
Solution
H 0 : 1 = 2
H 1 : 1 2
Where
Sp =
B
18
205
25
( n1 1) S12 + ( n2 1) S 22
n1 + n2 2
n1 + n2
n1n2
S X 1 X 2 = Sp
(
)
Where: X 1 =170, X 2 = 205, n1 = 20, n2 = 18, S1 = 20, S2 = 25, V = 36
Sp =
(19 ) ( 202 ) + (17 ) ( 252 )

20 + 18 2
= 22.5
S( X 1 X 2 ) = 22.5
38
360
= 7.31
t=
170 205
7.31
= 4.79
t0.05(36) = 1.9 (Since d.f > 30 we use the normal tables)
The table value of t at 5% level of significance for 36 d.f. when d.f. >30, that t distribution is the
same as normal distribution is 1.9. since the value computed value of t is more than the table
value, we reject the null hypothesis. Thus, we conclude that there is significant difference in the
average sales between the two salesmen
Testing the hypothesis equality of two variances

The test for equality of two population variances is based on the variances in two independently
selected random samples drawn from two normal populations
Under the null hypothesis 12 = 22
F=
s12
12
s 22
22
Now under the H0 : 12 = 22 it follows that
212
Lesson Six
F=
S12
which is the test statistic.
S 22
Which follows F distribution with V1 and V2 degrees of freedom. The larger sample variance is
placed in the numerator and the smaller one in the denominator
If the computed value of F exceeds the table value of F, we reject the null hypothesis i.e. the
alternate hypothesis is accepted
Example
In one sample of observations the sum of the squares of the deviations of the sample values
from sample mean was 120 and in the other sample of 12 observations it was 314. test whether
the difference is significant at 5% level of significance
Solution
Given that n1 = 10, n2 = 12, (x1 X 1 )2 = 120
(x2 X 2 )2 = 314
Let us take the null hypothesis that the two samples are drawn from the same normal population
of equal variance
H0 : 12 = 22
H1: 12 22
Applying F test i.e.
F=
S12
S 22
( X1 X 1 )
n1 1
X2 X 2
( n2 1)
120
9
314
11
13.33
28.55
since the numerator should be greater than denominator

F=
28.55
= 2.1
13.33
The table value of F at 5% level of significance for V1 = 9 and V2 = 11. Since the calculated
value of F is less than the table value, we accept the hypothesis. The samples may have been
drawn from the two population having the same variances.
Chi square hypothesis tests (Non-parametric test)(X2)

They include amongst others
Test for goodness of fit
i.
ii.
Test for independence of attributes
iii.
Test of homogeneity
iv.
Test for population variance
212
213
The Chi square test (2) is used when comparing an actual (observed) distribution with a
hypothesized, or explained distribution.
It is given by; 2 =
(O E )
Where O = Observed frequency
E = Expected frequency
The computed value of 2 is compared with that of tabulated 2 for a given significance level and
degrees of freedom.
i.
Test for goodness of fit
This tests are used when we want to determine whether an actual sample distribution matches a
known theoretical distribution
The null hypothesis usually states that the sample is drawn from the theoretical population
distribution and the alternate hypothesis usually states that it is not.
Example
Mr. Nguku carried out a survey of 320 families in Ateka district, each family had 5 children and
they revealed the following distribution
No. of boys
5
4
3
2
1
0
No. of girls
0
1
2
3
4
5
No. of families
14
56
110
88
40
12
Is the result consistent with the hypothesis that male and female births are equally probable at
5% level of significance?
Solution
If the distribution of gender is equally probable then the distribution conforms to a binomial
distribution with probability P(X) = .
Therefore
H0 = the observed number of boys conforms to a binomial distribution with P =
H1 = The observations do not conform to a binomial distribution.
On the assumption that male and female births are equally probable the probability of a male
birth is P = . The expected number of families can be calculated by the use of binomial
distribution. The probability of male births in a family of 5 is given by
(for x = 0, 1, 2, 3, 4, 5,)
P(x)
= 5cX Px q5-x
= 5cX ( )5
(Since P = q = )
To get the expected frequencies, multiply P(x) by the total number N = 320. The calculations are
shown below in the tables
x
0
1
2
3
4
P(x)
Expected frequency = NP(x)
5c
0
( )5
= 1
5c
1
( )5
= 5
5c
2
( )5
5c
3
( )5
5c
4
( )5
32
32
= 10
32
= 10
32
=5
32
320 1
320 5
32
= 10
= 50
32
320 10
= 100
32
320 10
= 100
32
320 5
= 50
32
214
Lesson Six
5c
5
( )5
=1
32
320 1
32
= 10
Arranging observed and expected frequencies in the following table and calculating x2
(O E) 2 /E
O
E
(O E) 2
14
10
16
1.60
56
50
16
0.72
110
100
100
1.00
88
100
144
1.44
40
50
100
2.00
12
10
4
0.40
(0 E) 2 /E = 7.16
(O E )
= 7.16
The table of 2 for V = 6 1 = 5 at 5% level of significance is 11.07. The computed value of 2
= 7.16 is less than the table value. Therefore the hypothesis is accepted. Thus it can be
concluded that male and female births are equally probable.
ii)
Test of independence of attributes
This test disclosed whether there is any association or relationship between two or more
attributes or not. The following steps are required to perform the test of hypothesis.
1. The null and alternative hypothesis are set as follows
H0: No association exists between the attributes
H1: an association exists between the attributes
2. Under H0 an expected frequency E corresponding to each cell in the
contingency table is found by using the formula
E=
RC
n
Where R = a row total, C = a column total and n = sample size

3. Based upon the observed values and corresponding expected frequencies the 2
statistic is obtained using the formular
2
(O E )
4. The characteristic of this distribution are defined by the number of degrees of

freedom (d.f.) which is given by
d.f. = (r-1) (c-1),
Where r is the number of rows and c is number of columns corresponding to a
chosen level of significance, the critical value is found from the chi squared
table
5. The calculated value of 2 is compared with the tabulated value 2 for (r-1) (c-1)
degrees of freedom at a certain level of significance. If the computed value of 2
is greater than the tabulated value, the null hypothesis of independence is
rejected. Otherwise we accept it.
214
215
Example
In a sample of 200 people where a particular devise was selected, 100 were given a drug and the
others were not given any drug. The results are as follows
Drug
No drug
Total
Cured
65
55
120
Not cured
35
45
80
Total
100
100
200
Test whether the drug will be effective or not, at 5% level of significance.
Solution
Let us take the null hypothesis that the drug is not effective in curing the disease.
Applying the 2 test
The expected cell frequencies are computed as follows
E11 =
R1C1
=
n
120 100
200
60
E12 =
R1C2
=
n
120 100
200
60
E21 =
R2C1
=
n
80 100
200
40
E22 =
R2C2
=
n
80 100
200
40
The table of expected frequencies is as follows

60
60
40
40
100
100
O
65
55
35
45
E
60
60
40
40
120
80
200
(O E) 2
25
25
25
25
(O E) 2 /E
0.417
0.625
0.417
0.625
(O E) 2 /E = 2.084
Arranging the observed frequencies with their corresponding frequencies in the following table
we get
(O E )
= 2.084
2
V= (r 1) (c-1) = (2 1) (2 1) = 1; tabulated
( 0.05 ) = 3.841
216
Lesson Six
The calculated value of 2 is less than the table value. The hypothesis is accepted. Hence the
drug is not effective in curing the disease.
Test of homogeneity
It is concerned with the proposition that several populations are homogenous with respect to
some characteristic of interest e.g. one may be interested in knowing if raw material available
from several retailers are homogenous. A random sample is drawn from each of the population
and the number in each of sample falling into each category is determined. The sample data is
displayed in a contingency table
The analytical procedure is the same as that discussed for the test of independence
Example
A random sample of 400 persons was selected from each of three age groups and each person
was asked to specify which types of TV programs be preferred. The results are shown in the
following table
Type of program
Age group
A
B
C
Total
Under 30
120
30
50
200
30 44
10
75
15
100
45 and above
10
30
60
100
Total
140
135
125
400
Test the hypothesis that the populations are homogenous with respect to the types of television
program they prefer, at 5% level of significance.
Solution
Let us take hypothesis that the populations are homogenous with respect to different types of
television programs they prefer
Applying 2 test
O
E
(O E) 2
(O E) 2 /E
120
70.00
2500.00
35.7143
10
35.00
625.00
17.8571
10
35.00
625.00
17.8571
30
67.50
1406.25
20.8333
75
33.75
1701.56
50.4166
30
33.75
14.06
0.4166
50
62.50
156.25
2.500
15
31.25
264.06
8.4499
60
31.25
826.56
26.449
(O E) 2 /E = 180.4948
2 =
(O E )
The table value of 2 for 4d.f. at 5% level of significance is 9.488

The calculated value of 2 is greater than the table value. We reject the hypothesis and conclude
that the populations are not homogenous with respect to the type of TV programs preferred,
thus the different age groups vary in choice of TV programs.
216
217
SUMMARY OF FORMULAE IN HYPOTHESIS

Testing
(a)
Hypothesis testing of mean
For n>30
Z=
S
X
Where S X =
at level of significance.
SX
n
For n < 30
t=
S
X
where S X =
SX
n
at n 1 d.f
level of significance
(b)
Difference between means (Independent samples)

For n > 30
Z=
X1 X 2
S X1 X 2
(
)
Where S
( X1X 2 )
S12 S 22
+
n1 n2
At = level of significance
For n < 30
t=
X1 X 2
at n1 + n2 2 d.f
S X1 X 2
(
)
where S
( X1X 2 )
= Sp
and S p =
(c)
n1 + n2
n1n2
( n1 1) S12 + ( n2 1) S22
n1 + n2 2
Hypothesis testing of proportions

Z=
p
Sp
Where: Sp =
pq
n
p = Proportion found in sample

q=1p
= hypothetical proportion
(d)
Difference between proportions

Z=
P1 P2
S( P1 P2 )
218
Lesson Six
Where:
pq pq
+
n1 n2
p n + p2 n2
p= 1 1
n1 + n2
S( P1 P2 ) =
(e)
q=1P
Chi-square test
X2 =
(f)
(O E )
Where O = observed frequency

Column total Row total
= expected frequency
E=
Sample Size
F test (variance test)
F=
S12
S 22
here the bigger value between the standard deviations make the numerator.
218
219

QUESTION ONE
A firm purchases a very large quantity of metal off-cuts and wishes to know the average weight
of an off-cut. A random sample of 625 off-cuts is weighed and it is found that the mean sample
weight is 150 grams with a sample standard deviation of 30 grams. What is the estimate of the
population mean and what is the standard error of the mean? What would be the standard error
if the sample size was 1225?
QUESTION TWO
A sample of 80 is drawn at random from a population of 800. The sample standard deviation
was found to be 6 grams.
- What is the finite population correction factor?
- What is the approximation of the correction factor?
- What is the standard error of the mean?
QUESTION THREE
State the Central Limit Theorem
QUESTION FOUR
a)
b)
c)
d)
e)
f)
g)
h)
What is statistical inference?

What is the purpose of estimation?
What are the properties of good estimators?
What is the standard error of the mean?
What are confident limits?
When is the Finite Population Correction Factor used? What is the formula?
How are population proportions estimated?
What are the characteristics of the t distribution?
QUESTION FIVE
A market research agency takes a sample of 1000 people and finds that 200 of them know of
Brand X. After an advertising campaign a further sample of 1091 people is taken and its found
that 240 know of Brand X.
It is required to know if there has been an increase in the number of people having an awareness
of Brand X at the 5% level.
QUESTION SIX
The monthly bonuses of two groups of salesmen are being investigated to see if there is a
difference in the average bonus received. Random samples of 12 and 9 are taken from the two
groups and it can be assumed that the bonuses in both groups are approximately normally
distributed and that the standard deviations are about the same. The same level of significance is
to be used.
The sample results were
n1=12
x1=1060
s1=63
n2=9
x2=970
s2=76
220
Lesson Six
QUESTION SEVEN
Torch bulbs are packed in boxes of 5 and 100 boxes are selected randomly to test for the
number of defectives
Number of
Number
Total
of boxes
defectives
Defectives
0
40
0
1
37
37
2
17
34
3
5
15
4
1
4
5
0
0
100
90
The number of any individual bulb being a reject is
90
5 = 0.18
100
and it is required to test at the 5% level whether the frequency of rejects conforms to a binomial
distribution.
QUESTION EIGHT
a) Define type I and type II errors.
b) What is a two-tail test?
c) What is the best estimate of the population standard deviation when the two samples are
taken
QUESTION NINE
Express Packets guarantee 95%of their deliveries are on time. In a recent week 80 deliveries
were made and 6 were late and the management says that, at the 95%level there has been a
significant improvement in deliveries.
Can the MDs statement be supported?
If not, at what level of confidence can it be supported?
A batch of weighing machines has been purchased and one machine is selected at random for
testing. Ten weighing tests have been conducted and the errors found are noted as follows:
Test
1
2
3
4
5
6
7
8
9
10
Errors (gms)
4.6
8.2
2.1
6.3
5.0
3.6
1.4
4.1
7.0
4.5
The purchasing manager has previously accepted machines with a mean error of 3.8 gms and
asserts that these tests are below standard.
Test the assertions at 5% level.
Compare your answers with those given in lesson 9 of the study pack
220
221
COMPREHESIVE ASSIGNMENT THREE

Instructions:
Time allowed: Three hours

SECTION I
QUESTION ONE
a) Explain what is meant by the following terms as used in statistical inference:
i)
ii)
iii)
iv)
v)
b)
Statistical hypothesis;
(2 marks)
Test of a hypothesis;
(2 marks)
Type I error;
(2 marks)
Type II error;
(2 marks)
Level of significance
(2 marks)
Cross Lines Group (CLG) has two factories in different parts of the country. Their
Resources, including the labour force skills are regarded as identical and both factories
were built at the same time.
A random sample of output data during a given period has been taken from each factory
and converted to standard hours of output per employee. The data are given below:
Factory 1
Factory 2
42
39
50
45
43
36
39
42
41
52
49
37
52
43
41
41
46
40
48
39
You are given that for factory 1 mean = 45.1 and variance = 20.10 and that for factory 2
mean = 41.4 and variance = 21.16.
Required:
i) Test the hypothesis that the mean of standard hours for employees in the two factories
is the same.
(7 marks)
(3 marks)
ii) Comment briefly on the conditions of the test and interpret the outcome.
(Total: 20 marks)
QUESTION TWO
a) State clearly what is meant by two events being statistically independent.
(2 marks)
b) In a certain factory which employs 500 men, 2% of all employees have a minor accident in a
given year. Of these, 30% had safety instructions whereas 80% of all employees had no
safety instructions.
Required
Find the probability of an employee being accident-free given that he had:
i) No safety instructions
ii) Safety instructions
(5 marks)
(5 marks)
c) An electric utility company has found that the weekly number of occurrences of lightning
striking the transformers is a Poisson distribution with mean 0.4.
222
Lesson Six
Required:
i) The probability that no transformer will be struck in a week.
ii) The probability that at most two transformers will be struck in a week
(3 marks)
(5 marks)
(Total: 20 marks)
QUESTION THREE
Explain the difference between the paired t-test and the two-sample t-test
(4 marks)
Trendy Tyres Ltd. Has introduced a new brand of tyres which in their advertisements claim to
be superior to their only competitor brand. The Roadmaster Tyres. The brand manager of
Roadmaster Tyres disputes this claim which he says is an advertisement gimmick. The brand
managers of the two companies agree to run a road test for the brands. Ten (10) saloon cars of
uniform weight and identical specifications are to be used for the test. Each car is fitted with
both brands of tyres: One brand at the front the other brand at the rear. The cars cover a
distance of 5,000 kilometers and the trend wear is recorded as follows:
1
2
3
4
5
6
7
8
9
10
Trend tyres
centimeters
1.08
1.06
1.24
1.20
1.17
1.21
1.18
1.10
1.22
1.60
Roadmaster tyres
Centimeters
1.12
1.09
1.16
1.24
1.23
1.25
1.20
1.15
1.19
1.13
Required:
i) Determine whether Trendy Tyres Ltd.s claim is true using = 0.01
ii) What are the assumptions you have made in (i) above?
(15 marks)
(1 mark)
(Total: 20 marks)
QUESTION FOUR
Kenear Commercial bank Ltd. commissioned a research whose results indicated that automatic
teller machines (ATM) reduces the cost of routine banking transactions.
Following this information, the bank installed an ATM facility at the premises of Joy Processing
Company Ltd., which for the last several months has exclusively been, used by JoyS 605
employees. Survey on the usage of the ATM facility by 100 of the employees in a month
indicated the following:
222
223
Number of times Frequency

ATM used
0
20
1
32
2
20
3
13
4
10
5
5
Required:
a) An estimate of the proportion of Joys employees who do not use the ATM facility in a
month
(2 marks)
b)
Determine the 95% confidence interval for the estimate in (a) above
(5 marks)
i)
Can the bank be certain that at least 40% of Joys employees will use the ATM facility?
ii)
(1 mark)
c) The number of ATM transactions on average an employee of Joy makes per month
(3 marks)
d) Determine the 95% confidence interval of the mean number of transactions made by an
employee in a month.
(6 marks)
e) Is it possible that the population mean number of transactions is four? Explain. (3 marks)
(Total: 20 marks)
QUESTION FIVE
State any five problems encountered in the construction of the consumer price index. (5 marks)
An investment analyst gathered the following data on the 91-day Treasury bill rates for the years
2003 and 2004.
Month
January
February
March
April
May
June
July
August
September
October
November
December
Treasury bill rates (%)

2003
2004
3.2
5.5
3.0
5.2
2.8
4.3
2.5
3.6
2.9
3.3
3.4
2.7
3.7
2.4
4.0
2.0
3.8
2.3
4.2
2.8
4.5
3.1
5.1
3.7
The analyst would like to determine if on average there was a significant change in the Treasury
bill rates over the two years.
Required:
i) The mean and variance of the Treasury bill rates for each year.
(10 marks)
ii) Determine if there is a significant difference in the average Treasury bill rates (use a
significance level of 1%)
(5 marks)
224
Lesson Six
Note:
n1 1) S12 + ( n2 1) S 22
(
=
n1 + n2 2
(Total: 20 marks)
SECTION II
QUESTION SIX
a) Describe the characteristics of the following distributions:
i) Binomial distribution
ii) Poisson distribution.
(3 marks)
(3 marks)
b) High Grade Meat Ltd. Produces beef sausages and sells them to various supermarket. In
order to satisfy the industrys requirement, the firm may only produces 0.2 per cent of
sausages below a weight of 80 grammes. The sausage producing machine operates with a
standard deviation of 0.5 grammes. The weights of the sausages are normally distributed.
The firms weekly output is 300,000 sausages and the sausage ingredients cost Sh.5.00 per
100 grammes. Sausages with weights in excess of 82 grammes require additional ingredients
costing Sh.2.50 per sausage.
Required:
i) The mean weight at which the machine should be set.
ii) The firms weekly cost of production
(4 marks)
(10 marks)
(Total: 20 marks)
QUESTION SEVEN
a) The past records of Salama Industries indicate that about 4 out of 10 of the companys
orders are for export. Further, their records indicate that 48 per cent of all orders are for
export in one particular financial quarter. They expect to satisfy about 80 orders in the next
financial quarter.
Required:
Determine the probability that they will break their previous export record.
Explain why you have used the approach you have chosen to solve part (i) above.
(7 marks)
(2 marks)
b) Gear Tyre Company has just developed a new steel-belted radial tyre that will be sold
through a national chain of discount stores. Because the tyre is a new product, the
companys management believes that the mileage guarantee offered with the tyre will be an
important factor in the consumer acceptance of the product. Before finalizing the tyre
mileage guarantee policy, the actual road test with the tyres shows that the mean tyre mileage
is = 36,500 kilometers and the standard deviation is = 5,000 kilometers. In addition, the
data collected indicate that a normal distribution is a reasonable assumption.
Required:
Gear Tyre Company will distribute the tyres if 20 per cent of the tyres manufactured can be
expected to last more than 40,000 kilometres. Should the company distributed the tyres?
(4 marks)
ii)
The company will provide a discount on a new set of tyres if the mileage on the original
tyres does not exceed the mileage stated on the guarantee.
i)
224
225
What should the guarantee mileage be if the company wants no more

than 10% of the tyres to be eligible for the discount?
(4 marks)
c) Explain briefly some of the advantages of the standard normal distribution.

(3 marks)
(Total: 20 marks)
QUESTION EIGHT
a) Explain the following terms used in statistical inference:
i)
Null hypothesis
Parametric test
ii)
Coefficient of correlation
iii)
Rank correlation coefficient
iv)
(2 marks)
(2 marks)
(2 marks
(2 marks)
b) State four areas that the chi-square distribution is used
(4 marks)
c) In an analysis of the results of telecommunication students, the examining board classified

the results as either credit, pass or discontinued. Further, the board analyzed the students
method of study which was either full-time, part-time or private. An employee of the board
cross-classified the examination results and the method of study of 300 students. He then
computed a test statistic of 42.28
Required:
i)
State the null and alternative hypotheses that should be tested.
(4 marks)
What conclusion can be drawn from the results of the data? (use = 0.05) (4 marks)
ii)
(Total: 20 marks)
226
Decesion Theory
LESSON SEVEN
Decision Theory
7.1
Decision theory
Decision trees and sequential decisions
Game theory
Decision Theory
Types of decisions
There are many types of decision making
1. Decision making under uncertainty
These refer to situations where more than one outcome can result from any single decision
2. Decision making under certainty
Whenever there exists only one outcome for a decision we are dealing with this category e.g.
linear programming, transportation assignment and sequencing e.t.c.
3. Decision making using prior data
It occurs whenever it is possible to use past experience (prior data) to develop probabilities for
the occurrence of each data
4. Decision making without prior data
No past experience exists that can be used to derive outcome probabilities in this case the
decision maker uses his/her subjective estimates of probabilities for various outcomes
Decision making under uncertainty
Several methods are used to make decision in circumstances where only the pay offs are known
and the likelihood of each state of nature are known
a) Maximin Method
This criteria is based on the conservative approach to assume that the worst possible is going
to happen. The decision maker considers each strategy and locates the minimum pay off for
each and then selects that alternative which maximizes the minimum payoff
Illustration
Rank the products A B and C applying the Maximin rule using the following payoff table
showing potential profits and losses which are expected to arise from launching these three
products in three market conditions
(see table 1 below)
226
Lesson Eight
227
Boom condition
Product A
Product B
Product C
+8
-2
+16
Pay off table in 000s

Steady state
Recession
1
+6
0
-10
+12
-26
Mini profits row

minima
-10
-2
-26
Table 1
Ranking the MAXIMIN rule = BAC
b) MAXIMAX method
This method is based on extreme optimism the decision maker selects that particular strategy
which corresponds to the maximum of the maximum pay off for each strategy
Illustration
Using the above example
Max. profits row maxima
Product A
+8
Product B
+12
Product C
+16
Ranking using the MAXIMAX method = CBA
c) MINIMAX regret method

This method assumes that the decision maker will experience regret after he has made the
decision and the events have occurred. The decision maker selects the alternative which
minimizes the maximum possible regret.
Illustration
Boom condition
Product A
Product B
Product C
8
18
0
Regret table in 000s

Steady state
Recession
5
0
6
22
0
38
Mini regret row

maxima
22
18
38
A regret table (table 2) is constructed based on the pay off table. The regret is the opportunity
loss from taking one decision given that a certain contingency occurs in our example whether
there is boom steady state or recession
The ranking using MINIMAX regret method = BAC
d) The expected monetary value method

The expected pay off (profit) associated with a given combination of act and event is obtained
by multiplying the pay off for that act and event combination by the probability of occurrence
of the given event. The expected monetary value (EMV) of an act is the sum of all expected
conditional profits associated with that act
Example
A manager has a choice between
228
Decesion Theory
i.
A risky contract promising shs 7 million with probability 0.6 and shs 4 million with
probability 0.4 and
ii.
A diversified portfolio consisting of two contracts with independent outcomes each
promising Shs 3.5 million with probability 0.6 and shs 2 million with probability 0.4
Can you arrive at the decision using EMV method?
Solution
The conditional payoff table for the problem may be constructed as below.
(Shillings in millions)
Event Ei Probability (Ei)
Conditional pay offs decision Expected pay off decision
(i)
Contract (ii) Portfolio(iii)
Contract (i) x (ii)
Portfolio (i) x (iii)
Ei
0.6
7
3.5
4.2
2.1
E2
0.4
4
2
1.6
0.8
EMV
5.8
2.9
Using the EMV method the manager must go in for the risky contract which will yield him a
higher expected monetary value of shs 5.8 million
e) Expected opportunity loss (EOL) method

This method is aimed at minimizing the expected opportunity loss (OEL). The decision maker
chooses the strategy with the minimum expected opportunity loss
f) The Hurwiz method
This method was the concept of coefficient of optimism (or pessimism) introduced by L.
Hurwicz. The decision maker takes into account both the maximum and minimum pay off for
each alternative and assigns them weights according to his degree of optimism (or pessimism).
The alternative which maximizes the sum of these weighted payoffs is then selected
g) The Laplace method
This method uses all the information by assigning equal probabilities to the possible payoffs
for each action and then selecting that alternative which corresponds to the maximum
expected pay off
Example
A company is considering investing in one of three investment opportunities A, B and C
under certain economic conditions. The payoff matrix for this situation is economic condition
Investment
opportunities
A
B
C
5000
-2000
4000
7000
10000
4000
3000
6000
4000
Determine the best investment opportunity using the following criteria

i.
Maximin
Maximax
ii.
iii. Minimax
iv. Hurwicz (Alpha = 0.3
228
Lesson Eight
229
Solution
Economic condition
Investment
1
2
3
Minimum
opportunities
A
5000
7000
3000
3000
B
-2000
10000
6000
-2000
C
4000
4000
4000
4000
i.
Using the Maximin rule Highest minimum = 4000
ii.
iii.
A
B
C
Maximum
7000
10000
4000
Choose investment C
Using the Maximax rule Highest maximum = 10000
Choose investment B
Minimax Regret rule
1
0
7000
1000
3000
0
6000
3000
0
2000
Maximum
regret
3000
7000
6000
Choose the minimum of the maximum regret i.e. 3000

Choose investment A
Hurwicz rule: expected values
iv.
For A (7000 x 0.3) + (3000 x 0.7) = 2100 + 2100 = 4200
For B (10000 x 0.3) + (-2000 x 0.7) = 3000 + 1400 = 1600
For C (4000 x 0.3) + (4000 x 0.7) = 1200 + 2800 = 4000
Best outcome is 4200 choose investment A
Value of perfect information

It relates to the amount that we would pay for an item of information that would enable us to
forecast the exact conditions of the market and act accordingly.
The expected value of perfect information EVPI is the expected outcome with perfect
information minus the expected outcome without perfect information namely the maximum
EMV
Example
From table 1 above and given that the probabilities are Boom 0.6, steady state 0.3 and recession
0.1 then
When conditions of the market are; boom launch product C: profit = 16
When conditions of the market are; steady state launch product B: profit = 6
When conditions of the market are; recession launch product B: profit = 12
The expected profit with perfect information will be
(16 x 0.6) + (6 x 0.3) + (12 x 0.1) = 12.6
our expected profit choosing product C is 7
the maximum price that we would pay for perfect information is 12.6 7 = 5.6
7.2
DECISION TREES AND SUB SEQUENTIAL DECISIONS
A decision tree is a graphic display of various decision alternatives and the sequence of events as
if they were branches of a tree.
230
-
Decesion Theory
The symbol
and
indicates the decision point and the situation of uncertainty
or event respectively. The node depicted by a square is a decision node while outcome
nodes are depicted by a circle.

-
Decision nodes: points where choices exist between alternatives and managerial decisions is
made based on estimates and calculations of the returns expected.
Outome nodes are points where the events depend on probabilities
Illustration of a tree diagram

Event
111
event
ACT
E1
A1
E2
D1
B1
D2
B2
C1
122
D3
A2
112
C2
121
131
For example 111 represents the payoff of the act event combination A1 E1 B1
When probabilities of various events are known they are written along the corresponding
branches. Joint probabilities are obtained by multiplying the probabilities along the branches
Example
Kauzi Agro mills ltd (KAM) is considering whether to enter a very competitive market. In case
KAM decided to enter this market it must either install a new forging process or pay overtime
wages to the entire workers. In either case, the market entry could result in
i.
high sales
medium sales
ii.
low sales
iii.
no sales
iv.
a) Construct an appropriate tree diagram
b) Suppose the management of KAM has estimated that if they enter the market
there is a 60% chance of their stakeholders approving the installation of the
new forge. (this means that there is a 40% chance of using overtime) a random
sample of the current market structure reveals that KAM has a 40% chance of
achieving high sales, a 30% chance of achieving medium sales, a 20% chance of
achieving low sales and a 10% chance of achieving no sales. Construct the
appropriate probability tree diagram and determine the joint probabilities for
various branches
c) Market analysts of KAM have indicated that a high level of sales will yield shs
1,000,000 profit; a medium level of sales will result in a shs 600000 profit a low
level of sales will result in a shs 200000 profit and a no sales level will cause
KAM a loss of shs 500000 apart from the cost of any equipment. Entering the
230
Lesson Eight
231
market will require a cash outlay of either shs 300000 to purchase and install a
forge or shs 10000 for overtime expenses should the second option be selected.
Draw the appropriate decision tree diagram
Solution
a) The tree diagram for this problem is illustrated as follows
The 1st stage of drawing a tree diagram is to show all decision points and outcome points done
from left to right, concentrate first on the logic of the problem and on probabilities or values
involved. This is called forward pass.
The resultant is the figure below:
Act
Act/event
Install forge
1
0
Use overtime
stop
Outcome/event
High sales
Medium sales
Low sales
No sales
High sales
10
Medium sales
11
Low sales
12
No sales
Do not enter market
Tree diagram
The entire sample space of act event choices is available to KAM are summarized in the table
shown below
Path
0135
0136
0137
0138
0149
0 1 4 10
0 1 4 11
0 1 4 12
02
Summary of alternative Act event sequence

Enter market, install forge, high sales
Enter market, install forge, medium sales
Enter market, install forge, low sales
Enter market, install forge, no sales
Enter market, use overtime, high sales
Enter market, use overtime, medium sales
Enter market, use overtime, low sales
Enter market, use overtime, no sales
Do not enter the market
232
Decesion Theory
b) The appropriate probability tree is shown in the figure below. The alternatives available
to the management of KAM are identified. The joint probabilities are the result of the
path sequence that is followed. For example, the sequence enter market install forge,
low sales yields (0.6) (0.2) = 0.12 = probability to install forge and get low sales.
Pay offs
Install forge
(300,000)
0.4
0.3
HS = 0.24 = 1,000,000
MS = 0.18 = 600,000
3
Enter Market
0.2
0.6
0.1
LS = 0.12 = 200,000
1
NS = 0.06 = - 500,000
0.4
Use overtime
(10,000)
4
Dont enter market
0.4
0.3
0.2
0.1
HS = 0.16 = 1,000,000
MS = 0.12 = 600,000
LS = 0.08 = 200,000
NS = 0.04 = - 500,000
(c)
The overall decision is determined after analysis of the expected values at various points
so the correct decision (with the highest expected value is made. The stage is worked
from right to left and is known as the backward pass.
The expected value for a decision is the highest pay off value where as the E.V
for an outcome is the summation of probability x pay off value of each branch.
In both cases any expenditure incurred due to the selection of the said option is
deducted.
In our case
Node 3 = [(0.4 1,000,000 ) + (0.3 600,000) + (0.2 200,000 ) + (0.1 50,000)]
- 300,000
E.V. = 615,000 300,000 = 315,000
Node 4 = [(0.4 1,000,000 ) + (0.3 600,000) + (0.2 200,000 ) + (0.1 50,000)]
- 10,000
E.V. = 615,000 10,000 = 605,000
Node 1 = (0.6 315,000) + (0.4 605,000)
E.V. = 431,000
Node 0 = The highest of (0;431,000)
Since not entering the market has a 0 expected value
= 431,000 = thus the decision should be to enter the market.
232
Lesson Eight
233
This is represented as below in a tree diagram.
0.4
Install forge
0.3
1,000,000
600,000
3
Enter Market
0.6
0.2
EV = 315,000
0.1
1
0
200,000
- 500,000
EV = 431,000
0.4
Use overtime
Dont enter market
4
EV = 605,000
0.4
0.3
0.2
0.1
1,000,000
600,000
200,000
- 500,000
0
BAYES THEORY AND DECISION TREES

It makes an application of Bayes Theorem to solve typical decision problems. This is examined
a lot so it is important to clearly understand it.
Example:
Magana Creations is a company producing Ruy Lopez brand of cars. It is contemplating
launching a new model, the Guioco. There are several possibilities that could be opted for.
- Continue producing Ruy Lopez which has profits declining at 10% per annum on a
compounding basis. Last year its profit was Shs. 60,000.
- Launch Guioco without any prior market research. If sales are high annual profit is put at
Shs. 90,000 with a probability which from past data is put at 0.7. Low sales have 0.3
probability and estimated profit of Shs. 30,000.
- Launch Guioco with prior market research costing Shs. 30,000 the market research will
indicate whether future sales are likely to be good or bad. If the research indicates good
then the management will spend Shs. 35,000 more on capital equipment and this will
increase annual profits to Shs. 100,000 if sales are actually high. If however sales are actually
low, annual profits will drop to Shs. 25,000. Should market research indicate good and
management not spend more on promotion the profit levels will be as for 2nd scenario
above.
- If the research indicate bad then the management will scale down their expectations to give
annual profit of Shs. 50,000 when sales are actually low, but because of capacity constrints if
sales are high profit will be Shs. 70,000.
Past history of the market research company indicated the following results.
234
Decesion Theory
Actual sales
Predicted sales
level
High
0.8*
0.2
Good
Bad
Low
0.1
0.9
*When actual sales were high the market research company had predicted good sales level
80% of the time.
Required:
Use a time horizon of 6 years to indicate to the management of the company which option
theory should adopt (Ignore the time value of money).
Solution
(a)
First draw the decision tree diagram
Ruy Lopez
(option 1)
60,000 (declining)
High 0.7
90,000
GUIOCO
(option 2)
Low 0.3
30,000
P(H|G)
0.95
Market
Research
(option 3)
Extra 35,000
Good
0.05
No extra
P(L|G)
P(H|G)
0.95
P(L|G)
0.05
P(H|B)
Bad
0.34
P(L|B)
0.66
100,000
25,000
90,000
30,000
70,000
50,000
Computations; note how probability figures are arrived at.

The decision tree dictates that the following probabilities need to be calculated.
P(G)
P(B)
For market research
234
Lesson Eight
235
P(H|G)
P(L|G)
P(H|B)
P(L|B)
For sales outcome;
P(G|H) = 0.8
P(B|H) = 0.2
P(G|L) = 0.1
P(B|L) = 0.9
P(H) = 0.7
P(L) = 0.3
Given
Good
P(G&H) = P(H) P(G|H)

0.7 0.8 = 0.56
B&H = P(H) P(B|H)
0.7 0.2 = 0.14
High 0.7
Bad
P(G) = P(G and H) + P(G and L)

= 0.56 + 0.03 = 0.59
P(B) = P(B and H) + P(B and L)
= 0.14 + 0.27 = 0.41
Note that P(G) + P(B) = 0.59 + 0.41 = 1.00
From Bayes rule;
P ( H |G ) =
P ( G|H ) P ( H ) 0.56
=
= 0.95
P (G )
0.59
P ( L|G ) =
P ( G|L ) P ( L ) 0.03
=
= 0.05
0.59
P (G )
P ( H |B ) =
P ( B|H ) P ( H ) 0.14
=
= 0.34
0.41
P ( B)
P ( L|B ) =
P ( B|L ) P ( L ) 0.27
=
= 0.66
0.41
P (B)
Evaluating financial outcome:

Option 1:
Last year Shs. 60,000 profits
Year
1=
2=
3=
4=
5=
6=
60,000 0.9 =
60,000 0.92 =
60,000 0.93 =
60,000 0.94 =
60,000 0.95 =
60,000 0.96 =
Shs.
54,000.0
48,000.0
43,740.0
39,366.0
35,429.5
31,886.5
253,022.0
P(G&L) = P(L) P(G|L)

0.3 0.1 = 0.03
P(B&L) = P(L) P(B|L)
0.3 0.9 = 0.27
Low 0.3
236
Decesion Theory
Option 2
Expected value of Giuoco
Node (A): 0.7(90,000 6) + 0.3(30,000 6)
= 378,000 + 54,000 = Shs. 432,000
Note that the figures a multiplied by 6 to account for the 6 years.
Option 3
Expected value of market research
Node (B): 0.95(100,000 6) + 0.05(25,000 6)
= 570,000 + 7,500 = Shs. 577,500
Deduct Shs. 35,000 for extensions
= 542,500.
Node (C): 0.95(90,000 6) + 0.05(30,000 6)
= 513,000 + 9,000 = Shs. 522,000
Node 1:
Compare B and C
B is higher, thus = 542,000.
Node (D):
0.34(70,000 6) + 0.66(50,000 6)
142,800 + 198,000 = Shs. 340,800
Node 2:
Shs. 340,800 or 0 no launch
Node (E):
0.59 542,500 + 0.41 340,800

320,075 + 139,728 = Shs. 459,803
Less market research expenditure
459,803 30,000 = Shs. 429,803
Node 2:
Final decision summary

Option 1 EMV = 253,022
Therefore we chose option 2 since it has the highest EMV.
Advantages of decision trees

1. it clearly brings out implicit assumptions and calculations for all to see question and revise
2. it is easy to understand
Disadvantages
1. it assumes that the utility of money is linear with money
2. it is complicated by introduction of more variables and decision alternatives
3. it is complicated by presence of interdependent alternatives and dependent variables
7.3
Game Theory
Game theory is used to determine the optimum strategy in a competitive situation
When two or more competitors are engaged in making decisions, it may involve conflict of
interest. In such a case the outcome depends not only upon an individuals action but also upon
236
Lesson Eight
237
the action of others. Both competing sides face a similar problem. Hence game theory is a
science of conflict
Game theory does not concern itself with finding an optimum strategy but it helps to improve
the decision process.
Game theory has been used in business and industry to develop bidding tactics, pricing policies,
advertising strategies, timing of the introduction of new models in the market e.t.c.
RULES OF GAME THEORY

i.
The number of competitors is finite
There is conflict of interests between the participants
ii.
Each of these participants has available to him a finite set of available courses of
iii.
action i.e. choices
iv.
The rules governing these choices are specified and known to all players
While playing each player chooses a course of action from a list of choices available
to him
the outcome of the game is affected by choices made by all of the players. The choices
v.
are to be made simultaneously so that no competitor knows his opponents choice until
he is already committed to his own
vi.
the outcome for all specific choices by all the players is known in advance and
numerically defined
When a competitive situation meets all these criteria above we call it a game
NOTE: only in a few real life competitive situation can game theory be applied because all the
rules are difficult to apply at the same time to a given situation.
Example
Two players X and Y have two alternatives. They show their choices by pressing two types of
buttons in front of them but they cannot see the opponents move. It is assumed that both
players have equal intelligence and both intend to win the game.
This sort of simple game can be illustrated in tabular form as follows:
Player X
Button m
Button n
Player Y
Button R
X wins 2 points
Y wins 2 points
Button t
X wins 3 points
X wins 1 point
The game is biased against Y because if player X presses button m he will always win. Hence Y
will be forced to press button r to cut down his losses
Alternative example
Player X
Button m
Button n
Player Y
Button R
X wins 3 points
Y wins 2 points
Button t
Y wins 4 points
X wins 1 point
In this case X will not be able to press button m all the time in order to win(or button n).
similarly Y will not be able to press button r or button t all the time in order to win. In such a
situation each player will exercise his choice for part of the time based on the probability
238
Decesion Theory
Standard conventions in game theory

Consider the following table
Y
3
X
-2
-4
1
X plays row I, Y plays columns I, X wins 3 points

X plays row I, Y plays columns II, X looses 4 points
X plays row II, Y plays columns I, X looses 2 points
X plays row II, Y plays columns II, X wins 1 points
3, -4, -2, 1 are the known pay offs to X(X takes precedence over Y)
here the game has been represented in the form of a matrix. When the games are expressed in
this fashion the resulting matrix is commonly known as PAYOFF MATRIX
STRATEGY
It refers to a total pattern of choices employed by any player. Strategy could be pure or a mixed
one
In a pure strategy, player X will play one row all of the time or player Y will also play one of this
columns all the time.
In a mixed strategy, player X will play each of his rows a certain portion of the time and player Y
will play each of his columns a certain portion of the time.
VALUE OF THE GAME
The value of the game refers to the average pay off per play of the game over an extended
period of time
Example
Player Y
3 4
Player X
6 2
in this game player X will play his first row on each play of the game. Player y will have to play
first column on each play of the game in order to minimize his looses
so this game is in favour of X and he wins 3 points on each play of the game.
This game is a game of pure strategy and the value of the game is 3 points in favour of X
Example
Determine the optimum strategies for the two players X and Y and find the value of the game
from the following pay off matrix
Player Y
3 -1 4 2
Player X -1 -3 -7 0
4 -7 3 -9
Strategy assume the worst and act accordingly
if X plays first
238
Lesson Eight
239
if X plays first with his row one then Y will play with his 2nd column to win 1 point similarly if X
plays with his 2nd row then Y will play his 3rd column to win 7 points and if x plays with his 3rd
row then Y will play his fourth column to win 9 points
In this game X cannot win so he should adopt first row strategy in order to minimize losses
This decision rule is known as maximum strategy i.e. X chooses the highest of these minimum
pay offs
Using the same reasoning from the point of view of y
If Y plays with his 1st column, then X will play his 3rd row to win 4 points
If Y plays with his 2nd column, then X will play his 1st row to lose 1 point
If Y plays with his 3rd column, then X will play his 1st row to win 4 points
If Y plays with his 4th column, then X will play his 1st row to win 2 points
Thus player Y will make the best of the situation by playing his 2nd column which is a Minimax
strategy
This game is also a game of pure strategy and the value of the game is 1(win of 1 point per
game to y) using matrix notation, the solution is shown below
Player Y
3 -1 4 2
Player X -1 -3 -7 0
4 -7 3 -9
4
-1 4
Row Minimum
7
9
column maximum
In this case value of the game is 1

Minimum of the column maximums is 1
Maximum of the row is also 1
i.e.
Xs strategy is maximim strategy
Ys strategy is Minimax strategy
Saddle Point
The saddle point in a pay off matrix is one which is the smallest value in its row and the largest
value in its column. It is also known as equilibrium point in the theory of games.
Saddle point also gives the value of such a game. In a game having a saddle point, the optimum
strategy for both players is to pay the row or column containing the saddle point.
Note: if in a game there is no saddle point the players will resort to what is known as mixed
strategies.
Mixed Strategies
Example
Find the optimum strategies and the value of the game from the following pay off matrix
concerning two person game
Player Y
1 4
Player X
5 3
In this game there is no saddle point
Let Q be the proportion of time player X spends playing his 1st row and 1-Q be the proportion
of time player X spends playing his 2nd row
240
Decesion Theory
Similarly
Let R be the proportion of time player Y spends playing his 1st column and 1-R be the
proportion of time player Y spends playing his second row
The following matrix shows this strategy
Player Y
R 1 R
Q 1 4
1 Q 5 3
Player X
Xs strategy
X will like to divide his play between his rows in such a way that his expected winning or loses
when Y plays the 1st column will be equal to his expected winning or losses when y plays the
second column
Points
1
5
Column 1
Proportion played
Q
1-Q
Expected winnings
Q
5(1-Q)
Column 2
Proportion played
Q
1-Q
Expected winnings
4Q
3(1-Q)
Total = Q + 5(1 Q)
Points
4
3
Total = 4Q + 3(1 Q)
Therefore Q + 5(1-Q) = 4Q +3(1-Q)
Giving Q = 2
and
(1-Q) = 3
5
5
This means that player X should play his first row 2
th
of the time and his second row 3
th
of
the time
Using the same reasoning
1R + 4(1-R) =
5R +3(1-R)
and
(1-R) = 4
Giving R = 1
5
5
This means that player Y should divide his time between his first column and second column in
the ratio 1:4
Player Y
1
5
Player X
2
5
3
5
4
5
1 4
5 3
Short cut method of determining mixed matrices
Player Y
1 4
Player X
5 3
240
Lesson Eight
241
Step I
Subtract the smaller pay off in each row from the larger one and smaller pay off in each column
from the larger one
1 4 4 -1 = 3
5 3 5 - 3 = 2
5 1 = 4 4 3 = 1
Step II
Interchange each of these pairs of subtracted numbers found in step I
1 4 2
5 3 3
1 4
Thus player X plays his two rows in the ratio 2: 3
And player Y plays his columns in the ratio 1:4
This is the same result as calculated before
To determine the value of the game in mixed strategies

In a simple 2 x 2 game without a saddle point, each players strategy consists of two probabilities
denoting the portion of the time he spends on each of his rows or columns. Since each player
plays a random pattern the probabilities are listed under
Pay off
1
Strategies which produce this pay off

Row I column I
Row I column II
Row II column I
Row II column II
Expected value (or value of the game)

Pay off
Probability p(x)
1
2
25
4
8
25
5
3
25
3
12
25
Joint probability
2 1 =2
5
5
25
2 4 =8
5
5
25
3 1 =3
5
5
25
3 4 = 12
5
5
25
Expected value x (p(x)
2
25
32
25
15
25
36
25
x p(x) = 85/25 = 17/5 = 3.4

3.4 is the value of the game
Dominance
Dominated strategy is useful for reducing the size of the payoff table
Rule of dominance
242
Decesion Theory
i.
ii.
If all the elements in a column are greater than or equal to the corresponding
elements in another column, then the column is dominated
Similarly if all the elements in a row are less than or equal to the corresponding
elements in another row, then the row is dominated
Dominated rows and columns may be deleted which reduces the size of the game
NB always look for dominance and saddle points when solving a game
Example
Determine the optimum strategies and the value of the game from the following 2xm pay off
matrix game for X and Y
Y
6 3 1 0 3
X
3 2 4 2 1
In this columns I, II, and IV are dominated by columns III and V hence Y will not play these
columns
So the game is reduced to 22 matrix, hence this game can be solved using methods already
discussed
Y
1 3
X
4 1
GRAPHICAL METHOD
Graphical methods can be used in games with no saddle points and having pay off m2 or 2n
matrix
The aim is to substitute a much simpler 22 matrix for the original m2 or 2m matrix
Example I
Determine the optimum strategies and the value of the game from the following pay off matrix
game.
Y
6 3 1 0 3
X
3 2 4 2 1
Draw two vertical axes and plot two pay offs corresponding to each of the five columns. The
pay off numbers in the first row are plotted on axis I and those in second row on axis II
242
Lesson Eight
243
Axis I
Axis II
2 K
6 A
-1
L -1
B 3
-2
-2
-3
-3
-4
-5
-5
-1
-1
-6
-6
-2
-7
-7
-3
-8
-8
L -4
-9
-9
-2
-3 K
-4
Example I
-4
Example II
Thus the two pay off number 6 and 3 in the first column are shown respectively by point A on
axis I and point B on axis II
On the two intersecting lines at the very bottom thicken them from below to the point of
intersection i.e. highest point on the boundary.
The thick lines on the graph KT and LT meet at T
The two lines passing through T identify the two critical moves of Y which combined with X
yield the following 2 2 matrix
Y
1 3
X
4 1
The value of the game and the optimum strategies can be calculated using the methods
described earlier
Example II
Determine the optimum strategies and the value of the game from the following pay off matrix
concerning two person 4 2 game
Y
6
3
X 2
2
4
244
Decesion Theory
The method is similar to the previous example, except we thicken the line segments which binds
the figure from the top and taken the lowest point on the boundary
The segments KP, PM and ML drawn in thick lines bind the figure from the top and their
lowest intersection M through which the two lines pass defines the following 2 2 matrix
relevant to our purpose
Y
3 4
X
7 1
The optimal strategies and the value of the game can now be calculated
Non Zero Sum Games

Until recently there was no satisfactory theory either to explain how people should play non zero
games or to describe how they actually play such games
Nigel Howard (1966) developed a method which describes how most people play non zero sum
games involving any number of persons
Example
Each individual farmer can maximize his own income by maximizing the amount of crops that
he produces. When all farmers follow this policy the supply exceeds demand and the prices fall.
On the other hand they can agree to reduce the production and keep the prices high
This creates a dilemma to the farmer
This is an example of a non zero sum game
Similarly marketing problems are non zero sum games as elements of advertising come in. in
such cases the market may be split in proportion to the money spent on advertising multiplied
by an effectieness factor
Prisoners Dilemma
It is a type of non zero sum game and derives its name from the following story
The district attorney has two bank robbers in separate cells and offers each a chance of
confession. If one confesses and the other does not then the confessor gets two years and the
other one ten years. If both confess they will get eight years each. If both refuse to confess there
is only evidence to ensure convictions on a lesser charge and each will receive 5 years
Another example
The table below is a pay off matrix for two large companies A and B. initially they both have the
same prices. Each consider cutting their prices to gain market share and hence improve profit
Corporation B
maintain prices
Maintain prices
3,3 status quo
Corporation A
Decrease prices
4, 1, A gains market
share and profit
Decrease prices
1 , 4 B gets market share and
profit
(2,2) Both retain market share
but lose profit
The entries in the pay off matrix indicate the order of preference of the players i.e. first A then
B.
We may suppose that if both player study the situation, they will both decide to play row I
column I(3,3).
244
Lesson Eight
245
However
Suppose As reasoning is as follows
If B plays column I then I should play row 2 because I will increase my gain to 4
In the same way Bs reasoning may be as follows
If A plays row I then I should play column 2 to get pay off 4 per play
If both play 2(row 2 column 2) each two receives a pay off of 2 only
In the long run pay off forms a new equilibrium point because if either party departs from it
without the other doing so he will be worse off before he departed from it
Game theory seems to indicate that they should play (2,2) because it is an equilibrium point but
this is not intuitively satisfying. On the other hand (3,3) is satisfying but does not appear to
provide stability. Hence the dilemma.
Theory of Metagames
This theory appears to describe how most people play non zero sum games involving a number
of persons
Prisoners dilemma is an example of this. The aim is to identify points at which players actually
tend to stabilize their play in non zero sum games.
This theory not only identifies equilibrium points missed by traditional game theory in games
that have one or more such points but also does so in games in which traditional theory finds no
such point
Its main aim is that each player is trying to maximize the minimum gain of his opponent
ADVANTAGES AND LIMITATIONS OF GAME THEORY
Advantage
Game theory helps us to learn how to approach and understand a conflict situation and to
improve the decision making process
LIMITATIONS
1. Businessmen do not have all the knowledge required by the theory of games. Most often
they do not know all the strategies available to them nor do they know all the strategies
available to their rivals
2. there is a great deal of uncertainty. Hence we usually restrict ourselves to those games with
known outcomes
3. The implications of the Minimax strategy is that the businessman minimizes the chance of
maximum loss. For an ambitious business man, this strategy is very conservative
4. the techniques of solving games involving mixed strategies where pay off matrices are rather
large is very complicated
5. in non zero sum games, mathematical solutions are not always possible. For example a
reduction in the price of a commodity may increase overall demand. It is also not necessary
that demand units will shift from one firm to another
246
Decesion Theory

QUESTION ONE
An Oil Company has recently acquired rights in a certain area to conduct surveys and test
drillings to lead to lifting oil where it is found in commercially exploitable quantities. The area is
already considered to have good potential for finding oil in potentially exploitable quantities. At
the outset the company has the choice to conduct further geological tests or to carry out a
drilling programme immediately. On the known conditions, the company estimates that there is
a 70:30 chance of further tests showing a success.
Whether the tests show the possibility of ultimate success or not, or even if no tests are
undertaken at all, the company could still pursue its drilling programme or alternatively sell its
rights to drill in the area. Thereafter, however, if it carries out its drilling programme, the
likelihood of final success or failure is considered dependent on the foregoing stages. Thus:
If successful tests have been carried out, the expectation of success in drilling is given as 80:20
If test indicate failure then the expectation of success in drilling is given as 20:80
If no tests have been carried out at all then the expectation of success in drilling is given as 55:45
Costs and revenues have been estimated for all possible outcomes and the net present value
(NPV) of each is given below.
Outcome
SUCCESS
With prior tests
Without prior tests
FAILURE
With prior tests
Without prior tests
SALE OF EXPLOITATION RIGHTS
Prior test show success
Prior test show failure
Without prior tests
NPV (m)
100
120
-50
-40
65
15
45
Required.
a) Draw up a decision tree diagram
b) Advise the company on the best course of action
QUESTION TWO
1.
A construction company has a 1 million contract to complete a building by 31 March

1995, but is experiencing delays due to the complex design. The managers have to make a
decision now whether to continue as at present, or to employ specialist-engineering
consultants at a cost of 200000.
If the company continues as at present, it estimates there is only a 30% chance of
completing the building on time, and that the delay could be one two or three months,
with equal probability. If the building is late, there are penalties for each months delay (or
part of a month).
246
Lesson Eight
247
The managers believe that if they employ specialist-engineering consultants, their chances
of finishing the building on time will be trebled. But if the building is still late, it would
only be one or two months late, with equal probability.
Required
a)
To draw a tree diagram to represent this decision problem, using squares for
decision points, circles for random outcomes, and including probabilities
revenues and penalties;
b)
To analyse the tree using expected value techniques:
c)
To write a short report for the managers, with reasons and comments,
recommending which decision to make.
QUESTION THREE
Define minimax and maximax decision rules
QUESTION FOUR
A has two ammunition stores, one of which is twice as valuable as the other. B is an attacker
who can destroy an undefended store but he can only attack one of them. A can only
successfully defend one of them.
What would A do so as to maximize his return from the situation no matter what B may do?
QUESTION FIVE
Determine the optimum strategies and the value of the game for the following pay off matrix.
X
1
-2
2
Y
2
1
0
-1
1
1
Compare your answers with those given in lesson 9 of the study pack
248
Decesion Theory
LESSON EIGHT
Operation Research
Linear programming
Transport and Assignments
Network Analysis
8.1
Linear programming
Linear programming is a technique of decision making used by managers to allocate limited
resources eg machinery, raw materials and labor in order to minimize costs or maximize
production. Decision variable are the amounts of each product to be made in a given time
period. Linear programming assumes that the variable has a linear relationship.
Application of linear programming
Production department to decide the quantity of pots to be produced subject to limited
resources (constraints) eg labour, power, machine hours, raw materials etc.
Marketing department: Allocation of salesmen to different sales regions subject to their expected
performance.
Human Resource: Scheduling personnels work hours and job description to either maximize
production or minimize cost.
Steps in solving linear programming problems(problem formulation)
1. Identify variables (eg product x and product y)
2. Identify the objective (To maximize contribution or to minimize cost), and write down its
mathematical presentation in terms of variables.
3. Identify the constraints (ie the limited resources shared among the variables), and write
down its mathematical representation in terms of variables.
4. write down the objectives and the constraints in terms of the variables.
These steps apply regardless of the number of the variables.
NOTE: If only two variables are involved, a graphical solution can be used otherwise for
multivariable problems, an algebraic method is applied to find the solution.
Example 1:
Long Castling Breweries manufactures two brands of beer, Benko lager and Benoni lager. Benko
has a contribution of Sh.4 per unit and Benoni has a contribution of Sh.3 per unit. Benko
requires 30 machine minutes and 30 labor minutes to manufacture a unit. Total available
machine hours per day are 12hrs whereas total available labour hours per day are 14hrs.
Required:
1. Formulate linear programming model.
2. How much of each brand should Long Castling produce if it wishes to maximize its daily
contribution assuming that all the lager produced is sold.
248
Lesson Eight
249
Solution:
1.
Formulating a linear programming model
Step 1: Identifying variables:
The variables here are the number of units of Benko and Benoni lager produced by Long
castling breweries per day; we can represent them as:
X1= a unit of Benko lager.
X2= a unit of Benoni lager.
Step 2: Identify the objective:
Definition: An objective is the desired result i.e. optimization of a function dependent on
decision variable and subject to some constraints.
The objective of Long Castling Breweries is to maximize daily contribution. Objective function
is the formula that will give us the total contribution in a day for both Benko lager and Benoni
lager.
The information above can thus be represented in a tabular form as:
(PER DAY)
Machine hours
Labor hours
Contribution
PRODUCT
X1
0.5
0.5
4
X2
0.33
0.5
3
Maximum available
hours/day
12
14
Objective function = 4X1 + 3X2

The objective is to maximize 4X1 + 3X2
Step 3: Identifying constraints (constraints formulation)

Definition: Constraints are circumstances that govern achievement of an objective.
Limitations must be quantified mathematically and they must be linear.
For Long castling breweries we have limited machine hours (12hrs/day), which must be shared
among production of Benko and Benoni lagers.
Therefore production must be such that the numbers of machine hours required is less than or
equal to 12 hours per day.
0.5X1+0.33X2 12hrs
Similarly for labour hours we have:

0.5X1+0.5X2 14hrs
Non-Negativity: It is logical assumption to assume that the company cannot manufacture

negative amounts of a product, thus it can only manufacture either zero product or more.
Therefore we have:
X10
or
X1, X20
X20
Thus the complete linear programming model is;
Maximize 4X1 + 3X2
Subject to the constraints;
0.5X1 + 0.33X2 12
0.5X1 + 0.5X2 14
X1, X2 0
250
Decesion Theory
Solving linear programming problems

The question requires us to optimize (in our case, maximize) the objective (the contribution
function), or in simple terms we are required to solve the linear programming model.
Solving linear programming model entails finding the values of variables that satisfy all
inequalities simultaneously and optimize the objective.
Graphical solution
This method is used to solve LP models in case where only two variables are involved. For more
than two variables (multivariable) then the simplex technique (algebraic method may be used).
Now in solving the problem above we first draw the axis, taking X1 to be the y axis and X2 to be
the X axis.
X1
0
X2
Next we plot the scales on each axis to approximate the scales to use them, we consider each
constraint equation. We get the value of one of variables putting the other variable to be zero
and by substituting the inequality or with equality sign (=).
For:
0.5X1 + 0.33X2 12
When X1 = 0
0.5(0) + 0.33X2 = 12
0.33X2 = 12
X2 = 12/0.33 = 36
Therefore point to plot is (36, 0)
Implying that when X2 = 36 then X1 = 0
When X2 = 0
0.5X1 + 0.33(0) = 12
0.5X1 = 12
X1 = 12/0.5 = 24
Therefore the point is (0, 24)
For
0.5X1 + 0.5X2 14
When X1 = 0
0.5(0) + 0.5X2 = 14
0.5X2 = 14
X2 = 14/0.5 = 28
250
Lesson Eight
251
When X2 = 0
0.5X1 + 0.5(0) = 14
0.5X1 = 14
X1 = 14/0.5 = 28
Comparing these values we see that X2 ranges between 0 28, therefore we can have the graph
plotted as:
Next draw each limitation (constraint) as separate line on the graph.

For 0.5X1 + 0.33X2 12
The two points that represent this line are (36, 0) and (0, 24). This is plotted as a straight line
from 36 on X2 axis to 24 on X1 axis.
252
Decesion Theory
Now including the Non-Negativity constraints since no negative product can be produced;
X1 0; x2 0
252
Lesson Eight
253
We must now consider how to choose the production which will maximize contribution. This
we do by plotting a line representing the objective function (4x1 + 3x2).
First choose a convenient point inside the feasible region
eg X2(10) +3X1(20) = 40 +60
= Sh 100
All of the other product mixes that give a contribution of Sh.100 lies on the line:
100 = 4X1 + 3X2 .....................................................................................(i)
<<This line is called a contribution line>>
Picking another point, say X2 =10 ad X1 = 20
Its contribution value is SH 110, thus give a contribution line of
110 = 4X1 + 3X2 ........................................................................................(ii)
Plotting these two contribution lines to our graph we get two parallel lines.
Until we reach the last feasible solution(s) before the line moves entirely out of the feasible
region.
254
Decesion Theory
Point X is the last feasible solution. Coordinates of this point give a combination of the two
lagers production volumes that fetches the highest contribution.
Coordinates of point X can be read from the graph, but for precision they are calculated by
solving simultaneously the equations of the two lines that intersect at point X.
The two constraints are called binding or limiting constraints. They are the resources being fully
used thus preventing daily contribution from increasing further.
Therefore to get point x we solve:
0.5X1 + 0.33X2 = 12 .(i)
0.5X1 + 0.5X2 = 14 .(ii)
Since X is the intersection of these two constraints, solving by deducting (i) from (ii) we get
0.17X2=2
X2 =11.76
And substituting X2 = 11.76 to equation (i) we get
X1 = 16.24
Therefore 11.76 units of Benko lager and 16.24 units of Benoni lager need to be produced for
maximum contribution.
Contribution = 4(16.24) + 3(11.76) = 100.24
Assumption made in linear programming

Assumptions that are made to solve these types of problems are that:.
Proportionality: all activities in linear programming problems are proportional to the
level of decision variables.
254
Lesson Eight
255
Divisibility: the solution to a linear programming problem does not have to be an

integer but for strictly whole number solutions, use integer programming.
Non-negativity: no decision variable can be negative.
Additivity: the total of all activities in linear programming problems are assumed to
equal to the sum of individual activities.
Special cases in linear programming

Infeasibility: This is when all constraints dont satisfy a particular point thus there is no
feasible solution.
Redundancy: A constraint is considered redundant if it does not affect the feasible
region. This happens in cases of excess resources since it does not limit attainment of
the objective.
Multiple optimal solutions: This occurs when the objective function has the same slope
as a binding constraint.
Minimizing problem
Example
A manufacturing company has acquired new machine for producing product P at a rate of 25
units per hour with a 98% rate of efficiency. The company requires to produce atleast 1800 units
of P per day. The 10 old machines that the company has, produce 15 units of P with a 95%
efficiency.
The cost of operating the new machine is Sh. 4 per hour and Sh. 3 per hour for the old ones.
The cost incurred due to inefficiency is Sh. 2 per unit; It is government policy that at least 2 of
the new machines must be indulged into production.
The company wishes to optimally allocate the machines in order to minimize the total
manufacturing cost if the total available hours for production in a day are 8 hours.
Solution
In minimizing problems we use (greater or equal to) type inequality.
Step 1: Identifying variables.
Since the problem requires us to appropriately allocate the machines in order to minimize the
costs thus our variables are the new machines and the old machines, we can let;
X1 =new machines
X2=old machines
Step 2: Identify objectives:
The objective is to minimize manufacturing costs.
Total manufacturing cost per machine=Operating cost + (inefficiency rate number of units
cost of loss)
Therefore cost for new machines
Cost = 4+(0.02 25 2) = Sh. 5 per hour
=5 8= Sh. 40 per day]
Similarly for old machines
Cost = 3+(0.05152) = Sh. 4.5 per hour
=4.58=Sh. 36 per day
Therefore the objective function is to minimize 40X1 + 36X2
Step 3: Identifying constraint functions.
X1 8
X2 10
256
Decesion Theory
(25 8) X1 +(8 15) X2 1800 (This can be simplified further as)

200 X1 +120 X2 1800
(dividing through by 40)
5X1 +3 X2 45
X1 2
X1, X2 0
Thus the LP model is:
Minimize
40X1 + 36X2
Subject to:
X1 8
X2 10
5X1 +3 X2 45
X1 2
X1, X2 0
Plotting this on a graph we get:
X2
8
X1
The line X1 2 does not affect the feasible region (doesnt cause reduction of the feasible
region), this constraint doesnt limit attainment of the objective, thus its known as a redundant
constraint.
Now picking a convenient point inside the feasible region, say (6, 10)
We get a total cost of
600 = (6(40)+10(36))
Thus the objective function line of 40X1 + 36X2 = 600
Moving this line parallel toward the origin to locate the last apex before the line completely fall
off the feasible region, we get:
256
Lesson Eight
257
Point X is the point of optimal solution

The binding constraints here are:
X1 8
And
5X1 +3 X2 45
Solving this to get coordinates of point X
5X1 +3 X2 = 45 (when X1 = 8)
40 + 3 X2 = 45
X2 = 5/3
Thus the solution is
X1 = 8
X2 = 5/3
Shadow or dual prices

Definition: A shadow price or a dual price is the amount increase (or decrease) of the objective
function when one more (or one less) of the binding constraints is made available.
Consider example 1.
Maximize 4X1 + 3 X2
Subject to:
0.5X1 + 0.33X2 12 (Machine hours)
0.5X1 + 0.5X2 14 (labor hours)
Starting with machine hours; lets assume that one more machine hour is available (with labor
hours remaining constant)
We get:
0.5X1 + 0.33X2 = 13
0.5X1 + 0.5X2 = 14
Solving this simultaneously we get the values of X1 and X2 as
0.17 X2 = 1
X2 = 5.88
X1 = 22.12
258
Decesion Theory
Thus the contribution is

4(22.12) + 3(5.88) = Sh.106.12
Comparing this with its original contribution of Sh.100.24 (see example 1) we see increasing
machine hours by one unit has increased contribution by Sh.5.88, which is the shadow price per
machine hour.
Note: This figure is also arrived at if we assume that machine hours are reduced by 1 unit ie 121.
Similarly assuming that one more labor hour is made available, then contribution change is:
0.5X1 + 0.33X2 = 12
0.5X1 + 0.5X2 = 15
Solving this simultaneously gives:
0.17 X2 =3
X2 = 17.65
X1 = 12.35
Which give a contribution of:
4(12.35) + 3(17.65) = Sh.102.35
The contribution change is Sh.2.11 which is the shadow price per labor hour.
Note:
The shadow prices apply in so far as the constraint is binding for example if more and more
labor hours are available it will reach a point where labor hours are no longer scarce thus labor
hours cease to be a binding constraint and its shadow price becomes a zero.(All non-binding
constraints have zero shadow price). Logically its senseless to pay more to increase a resource,
which is already abundant.
Interpretation of shadow prices
A shadow price of a binding constraint indicates to management how much extra contribution
will be gained by increasing a unit of the scarce resource.
In the example above Sh.2.11 is the shadow price for labor hours. This implies that management
is ready to pay up to Sh.2.11 extra per hour for the extra hours i.e. say an employee is paid sh.5
per hour and one day he works for two hours extra (overtime), the management is prepared to
pay up to sh.7.11 per hour for the two hours overtime worked.
Sensitivity Analysis
Definition: Sensitivity analysis is the test of how certain changes in resources affect the optimal
solution.
In sensitivity analysis we consider the effect of additional limiting or non-limiting constraints.
We already know that adding more non-limiting constraints does not change the optimal
solution.
We also know that adding more binding constraints affects the objective function.
It is very important for the management to know how much of a limit resource can be made
available until it has no effect on the objective function (ie ceases to be a binding resource)
SIMPLEX METHOD
When analyzing linear programming problems with three or more variables the graphical
method becomes enadequate, in such cases we employ simplex method . Simplex method is an
algebraic procedure for solving systems of equations requiring optimization of the objective
function..
This method can be applied to any number of variables, the more they are the more complex it
becomes to workout a solution on paper. Computer programs e.g. Tora are used to solve the
most intricate problems.
258
Lesson Eight
259
The first step in simplex method is conversion of inequalities to linear equations
Example
Consider the linear problem.
Maximize
45 x1 + 80 x2
5 x1 + 20 x2 400
10 x1 + 15 x2 450
Subject to
x1 0, x2 0
Solution.
1. to convert this problem to a system of linear equation, we introduce slack variables to
each constraint.
Z = 45 x1 + 80 x2
5 x1 + 20 x2 + x3 = 400
10 x1 + 15 x2 + x4 = 450
Subject to
x1 , x2 , x3 , x4 0
where the structural variables x3 and x4 are slack variables
2. we then place this information in a tabular form known as a tableu
Initial tableu
Products
Slack Variables
Solution
Solution
Quantity
Variable
x
x
x
x
1
x3
20
400
x4
10
15
450
45
80
The tableu shows that
x3 = 400
x4 = 450
Z =0
(Total contribution)
It also show that unused capacity is at maximum i.e. the value for the slack x3 and x4 is 0.
Afer several operations and when an optimal solution has been attained, these values will change
to give an optimal feasible solution.
3. Select the column with the highest value of Z (i.e. 80), then devide the positive numbers
in that column (i.e. the x2 column) into the quantity column.
i.e. 400 20 = 20
450 15 = 30
chose the rowgiving the lowest result (in our case the row with x3 gives 20) and mark the element
falling on the intersection of the selected row and selected column (i.e. 20: selected element)
4. Devide all the elements in the selected row by the value of the selected and change the
solution variable to the heading of the identified column (from x3 to x2)
260
Decesion Theory
Thus the tableu will appear as follows.

Products
Slack Variables
Solution
Variable
x
x
x
x
1
Solution
Quantity
x2
1
4
1
20
20
x4
10
15
450
45
80
5. next we conduct row operations that aim to reduce elements falling in the same column
as the previously marked element to zero. These row operations may sometime
necessitate multiplying or deviding the selected row with an arbitrary number.
Therefore:
Row 2 15Row 1
15
new
x4 10 15 0 1 = 450
x2
15
4
x4
6 14
15
15
20
0*
3
4
0 = 300
1 = 150
note that the aim was to attain the zero.
80
new
Z 45 80 0 0 =
0
x2 20 80 4 0 = 1600
Z
25 0*
4 0 = 1600
on replacing the new rows to the tableu we get

2nd tableu.
Solution
Products
Slack Variables
Variable
x
x
x
x
1
Solution
Quantity
x2
1
4
1
20
20
x4
6 14
3
4
150
25
-4
-1600
Since in the Z row under products column we still have values greator than zero, we
conduct another operation.
Taking the column with a Z value of 25, we repeat the process in the same manner.
20 14 = 80
150 6 14 = 24
thus we pick the x4 row and mark the element 6 14 , the row solution variable is changed to
x1 and we devide the row by 6 14 to convert the marked element to 1.
Therefore;
x1
3
25
4
25
= 24
Next we do the row operations
260
Lesson Eight
( 4 )
x2
1
4
1
20
x1
1
4
3
100
new x
2
261
Z
x1
25
new Z
0* 1
0 = 20
4
100
251
1
50
= 6
= 24
25 0 4 0 = 1600
25 0 3 4 = 600
0*
0 7 4 = 2200
the tableu becomes, 3rd Tableu

Solution
Products
Variable
x
x
1
x3
x4
Solution
Quantity
x2
1
50
251
24
x1
3
25
4
25
24
-7
-4
-2200
Slack Variables
This is the final tableu since the Z row has no values greator than zero thus we have the
optimal solution.
Interpretation
to maximise Z we need to produce 24 units x2 and 24 units of x1, we obtain these
values from the solutions quantity column
thus, Z = 24(45) + 24(80) = 3000
we have zero slack (unused quantities of constraints).
Assume tableu 2 is the final tableu and let the constraint with variable x3 be labour
hours and x4 be raw materials, the slack wouldhve been interpreted as:
o 150 units of raw materials were unused
o to maximize Z we produce 20units of x2 and none of x1.
The values represnts in Z row under slack vaiable column represents shadow
prizes. Thus the shadow prize for the first constraint with x3 is 7 and the shadow
prize for the second constraint with the vaiable x4 is 4.
INTERPRETATION OF COMPUTER GENERATED SOLUTION

Example
Maximize
25 x1 + 20 x2 + 24 x2
where : x1 = Xtragrow, x2 = Youngrow, x3 = Zupergrow
Subject to
0.3 x1
+ 0.2 x3 500
0.5 x2 + 0.4 x3 1000
0.2 x1 + 0.1x2 + 0.1x3 800

0.4 x2 + 0.3 x3 600
x1
1500
x1 0, x2 0, x3 0
262
Decesion Theory
The computer generated solution for this problem is as follows;

Objective value = 71666.7
Variable
Value
Obj. Coeff
X1: Xtragrow
1666.7
25
X2: Youngrow
1500
20
X3: Zupergrow
0
24
Constraint
1(<)
2(<)
3(<)
4(<)
5(>)
Variable
X1: Xtragrow
X2: Youngrow
X3: Zupergrow
Constraint
1(<)
2(<)
3(<)
4(<)
5(>)
RHS
500
1000
800
600
1500
Slack-/Surplus+
0
250316.670
166.7+
Current obj coeff
Min obj coeff
25
13.50
20
9.78
24
-Infinity
Current RHS
500
1000
800
600
1500
Obj Value Contribution

4166.7
30000
0
Min RHS
450
750
483.3
0
-Infinity
Max Obj Coeff

Infinity
Infinity
31.67
Reduced cost
0
0
7.67
Max RHS
975
Infinity
Infinity
800
1666.7
Dual price
83.33
0
0
50
0
Required
Interpret the data generated by the the computer.
Solution.
Table 1:
Objective value, is the solution to objective function (e.g the solution to this example is 71,436)
The four columns of table 1 are to be interpreted as follows;
Variable: these are the variables of the model. In our example we have x1 = Xtragrow, x2 =
Youngrow and x3 = Zupergrow
Value: this is value that the variables assume at optimal solution (to optimize the
objective function one needs to produce this amounts of the variables). In our example we
are required to produce 1,666.67 of x1 and 1,750 of x2 and none of x3
Objective coefficient: these are the coefficients of the objective function
Objective value contribution: this is the value contributed by each variable to the
objective function (for x1=251,666.67), the total of this is equal to our objective value
(i.e 41,666.67+35,000=76,666.67).
The 3 columns of the second part of table1 can be interpreted as follows;
Constraints: this is constraints of the model representing the limited resources.
RHS: the Right hand side value is the limiting value of the constraint. E.g for the first
constraint the maximum amount of material A is 500 tons.
Slack/surplus: at optimal production not all the materials for some of the constrants will
be fully utilized, slack is the amount of material that is left over after production. For
constraint 1 and 4 no material remained, this also implies that these are the binding constraints i.e their
adjustment directly affects the objective solution
262
Lesson Eight
263
This is the analysis of the effect of adjusting variables or constraint, whether te objective
solution will be affected. How much of the objective coefficient (or the maximum available
amount of a constraint) can be reduced or increased without affecting the objective solution.
The columns of this table can be interpreted as follows;
Variable: as explained above
Current objective coefficient: this is the value of coefficients of the objective function
Minimum objective coefficient: this is how low the coefficient can be reduced without
affecting the optimal basis. The coefficient for x1 can be reduced from 25 to 13.50 (the prize for
Xtragrow fall to as low as $13.50 from $25) but the optimal solution will remain the same
Maximum objective coefficient: this is how high the coefficient can be increased without
affecting the optimal basis
Reduced cost: this is amount by which the coefficient of the variable has to be adjusted
with for it to become a basic variable (included to the objective optimal solution). X1
and x2 have 0 reduced costs implying that they already make part of the optimal solution, x3 will
require to be increased by 7.67 for it to make part of the basic variable.
The second part of the table is interpreted as follows;
Constraints: as described above
Current Rihgt Hand Side: the limiting value of the constraint. E.g for the first constraint the
maximum amount of material A is 500 tons.
Minimum RHS: the lowest the available amount of the constraint can be reduced
without affecting the optimal basis
Maximum RHS: the highest the available amount of the constraint can be increased
without affecting the optimal solution
Dual price: this is amount increase to the objective contribution due to a unit increase
of the available constraint. Since there are only 500 tons of material A if management decides to
increase it by a unit to 501tons then the objective optimal solution will be 76,666.67+83.33.
TRANSPORTATION
A transportation problem deals with a number of sources of supply (e.g a manufacturing
company, warehouse) and a number of destinations (e,g shops, houses). The usual objective is
minimizing transportation costs of supplying items from a set of source points to a set of
destinations.
A major characteristic of this problem is the linearity requirement, i.e. transport cost fom one
point to another must be clearly defined, if it will cost sh.50 to transport a bag from a warehouse
to shop A then it will cost sh.250 to transport 5 bags.
Assumptions
The model assumes a homogeneous commodity, one type of commodity
Total supply is equal to total demand
Example 1
64 chambers a computer support firm has three branches at different parts of the city, it receives
orders for a total of 15 desktop computers from four customers. In total in the three branches
264
Decesion Theory
there are 15 machines available. The management wish to minimise delivery costs by dispatching
the computers from the appropriate branch for each customer.
Details of the availabilities, 'requirements, and transport costs per computer are given in the
following table.
Table 1
Available
Customer Customer Customer Customer

A
B
C
D
Computers
3
3
4
5
Branch X.
2
Sh.13
11
15
20
Branch Y
6
Sh.17
14
12
13
Branch Z
7
Sh.18
18
15
12
Total
15
Total
15
transportation
cost
per unit
Solution
Step 1 Make an initial feasible allocation of deliveries by selecting the cheapest route first, and
allocate as many as possible then the next cheapest and so on. The result of such an
allocation is as follows.
Table 2
A
3
Available
X
Y
Z
2 Units
6 Units
7 Units
1(4)
2(5)
Requirement
B
C
3
4
2(1)
1(3)
4(2)
D
5
5(2)
Note: the number in the table represent deliveries of computers and the number in the brackets (1), (2), etc
represent the sequence in which they are inserted, lowest cost first i.e.
Sh.
Total cost 22
1. 2 units X B sh.11/unit
2. 4 units Y C sh.12/unit
Total cost 48
5 units Z D sh.12/unit
Totals cost 60
3. The next lowest cost move which is feasible i.e. doesnt exceed row or column totals is 1
unit Y B sh.14/unit
14
4. similarly the next lowest feasible allocation 1 unit Y A
sh.17/unit
17
5. finally to fulfill the row /column totals 2 units Z A sh.18/unit
__36
197
Step 2. Check solution obtained to see if it represents the minimum cost possible. This is done
by calculating shadow costs (i.e. an imputed cost of not using a particular route) and
comparing these with the real transport costs to see whether a change of allocation is
desirable.
This is done as follows:
Calculate a nominal 'dispatch' and 'reception' cost for each occupied cell by making an
assumption that the transport cost per unit is capable of being split between dispatch and
reception costs thus:
264
Lesson Eight
265
D(X) + R(B) = 11
D(Y) + R(A) = 17
D(y) + R(B) = 14
D(Y) + R(C) = 12
D(Z) + R(A) = 18
D(Z) + R(D) = 12
Where D(X), D(Y) and D(Z) represent Dispatch cost from depots X, Y and Z, and R(A) R(B),
R(C) and R(D) represent Reception costs at customers A, B, C, D.
By convention the first depot is assigned the value of zero i.e. D(X) = 0 and this value is
substituted in the first equation and then all the other values can be obtained thus
R(A) = 14
R(B) = 11
R(C) = 9
R(D) =
8
D(X) = 0
D(Y) = 3
D(Z) = 4
Using these values the shadow costs of the unoccupied cells can be calculated. The unoccupied
cells are X : A, X : C, X : D, Y : D, Z : B, Z : C.
D(X) + R(A)
D(X) + R(C)
D(X) + R(D)
D(Y) + R(D)
D(Z) + R(B)
D(Z) + R(C)
=
=
=
=
=
=
0
0
0
3
4
4
+
+
+
+
+
+
14
9
8
8
11
9
Shadow
costs
14
9
8
11
15
13
=
=
=
=
=
=
These computed 'shadow costs' are compared with the actual transport costs (from Tab- I),
Where the actual costs are less than shadow costs, overall costs can be reduced by allocating units
into that cell.
CellX:A
X:C
X:D
Y: D
Z:B
Z:C
Actual
cost
13
15
20
13
18
15
Shadow
cost
14
9
8
11
15
13
+ Cost increase
- Cost reduction
=
-1
=
+6
=
+ 12
=
+2
=
+3
=
+2
The meaning of this is that total costs could be reduced by sh.1 for every unit that can be
transferred into cell X : A. As there is a cost reduction that can be made the solution , Table 2
is not optimum.
Step 3: Make the maximum possible allocation of deliveries into the cell where actual costs are
less than shadow costs using occupied cells i.e.
Cell X : A from Step 2, The number that can be allocated is governed by the need to keep within
the row and column totals. This is done as follows:
266
Decesion Theory
Table 3
X
Y
Z
Available
A
3
+
12
2 Units
6 Units
7 Units
Requirement
B
C
3
4
21+
4
D
5
Table 3 is a reproduction of Table 2 with a number of + and - inserted. These were inserted for
the following reasons.
Cell X : A + indicates a transfer in as indicated in Step 2
Cell X : B - indicates a transfer out to maintain Row X total.
Cell Y : B + indicates a transfer in to maintain Column B total
Cell Y : A - indicates a transfer out to maintain Row Y and Column A totals.
The maximum number than can be transferred into Cell X : A is the lowest number in the
Minus cells i.e. cells Y : A, and X : B which is 1 unit.
Therefore 1 unit is transferred in the + and - sequence described above resulting in the following
table
Table 4
Available
X
Y
Z
2 Units
6 Units
7 Units
A
3
1
Requirement
B
C
3
4
1
2
4
D
5
The total cost of this solution is
Cell X:A
Cell X:B
Cell Y:B
Cell Y:C
Cell Z:A
Cell Z:D
1 unit @ sh.13
1 Unit @ sh.11
2 Units @ sh.14
4 Units @ sh.12
2 Units @ sh.18
5 Units @ sh.12
Sh.
= 13
= 11
= 28
= 48
= 36
= 60
196
The new total cost is sh.1 less than the total cost established in Step 1. This is the result expected
because it was calculated in Step 2 that sh.1 would be saved for every unit we were able to
transfer to Cell X : A and we were able to" transfer 1 unit only.
Notes: Always commence the + and - sequence with a + in the cell indicated by the (actual cost shadow cost) calculation. Then put a - in the occupied cell in the same row which has an
occupied cell in its column. Proceed until a - appears in the same column as the original +.
Step 4. Repeat Step 2 i.e. check that solution represents minimum cost. Each of the processes in
266
Lesson Eight
267
Step 2 are repeated using the latest solution (Table 4) as a basis, thus: Nominal dispatch
and reception costs for each occupied cell.
D(X) + R(A) = 13
D(X) + R(B) = 11
D(y) + R(B) = 14
D(Y) + R(C) = 12
DZ) + R(A) = 18
D(Z) + R(D) = 12
On setting D(X) to be 0, the rest of the values are found to be

R(A) = 13
D(X) = 0
R(B) = 11
D(Y) = 3
R(C) = 9
D(Z) = 5
R(D) = 7
Using these values the shadow costs of the unoccupied cells are calculated. The unoccupied cells
are X:C , X:D, Y:A, Y:D, Z:B, and Z:C
Therefore;
D(X) + R(C) = 9
D(X) + R(D) = 7
D(Y) + R(A) = 16
D(Y) + R(D) = 10
D(Z) + R(B) = 16
D(Z) + R(C) = 14
The computed shadow costs are compared with actual costs to see if any reduction in cost is
possible.
Cell X :C
X:D
Y:A
Y:D
Z:B
Z:C
cost
15
20
17
13
18
15
+ Cost
increase
cost - Cost reduction
9=
+6
7=
+13
16 =
+1
10 =
+3
16 =
+2
14 =
+1
Shadow
Actual
-
It will be seen that all the answers are positive, therefore no further cost reduction is possible
and optimum solution has been reached.
Thus the optimal solution is represented by table 4
UNEQUAL SUPPLY AND DEMAND QUANTITIES

Consider the following example.
Example 2
Wanjiru books supplies in a firm dealing with import of books and it has three stores
strategically situated around the country. Yesterday the company received orders to supply 100
books from 4 schools, of the books ordered the firm has 110 books in stock. The firm wishes to
268
Decesion Theory
minimize cost and its seeking your advice, advise the firm.
Below is a table of availability and requirement;
Available
Store I
Store II
Store III
Total
Books
40
20
50
110
Sch. A
25
Sh.3
Sh.1
Sh.4
Required
Sch. B Sch. C
25
42
16
9
9
3
5
2
Sch. D
8
8
5
Total
100
transport
costs per
Book
Solution
Step 1: add a dummy destination to table 5 with zero transport costs and requirements equal to
the surplus availability.
Available
Store I
Store II
Store III
Total
Books
40
20
50
110
Sch. A
25
Sh.3
Sh.1
Sh.4
Sch. B
25
16
9
5
Required
Sch. C Sch. D Dummy
Total
10
42
8
100
0
9
transport
costs per
3
8
0
Book
2
5
0
Step 2. Now that the quantity available equals the quantity required (because of insertion of the
dummy) the solution can proceed in exactly the same manner described in the first
example. First set up an initial feasible solution
Available
I
II
III
40
20
50
A
25
5(4)
20(1)
B
25
17(6)
8(5)
Requirement
C
D
42
8
8(3)
Dummy
10
10(7)
42(2)
The numbers in the table represent the allocations made and the numbers in brackets represent
the sequence they were inserted based on lowest cost and the necessity to maintain row/column
totals. The residue of 10 was allocated to the dummy. The cost of this allocation is
Sh.
Sh.
IA
5 units @ 3
15
IB
17 units @ 16
272
ID
8units @ 2
16
IDummy
10 units @ zero cost
IIA
20 units @ 1
20
IIIB
8 units @ 5
40
IIIC
42 units @ 2
84
447
268
Lesson Eight
269
Step 3. Check solution to see if it represents the minimum cost possible in the same manner as
previously described i.e.
Dispatch & Reception Costs of used routes:
D(I) + R(A)
=3
D(I) + R(B)
= 16
D(I) + R(D)
=2
D(I) + R(Dummy) = 12
D(II) + R(A)
=1
D(III) + R(B)
=5
D(III) + R(C)
=2
Setting D(I) at zero the following values are be obtained

R(A)
R(B)
R(C)
R(D)
R(Dummy)
=3
=16
=13
=2
=0
D(I)
=0
D(I)
=-2
D(III) =-11
Using these values the shadow costs of the unused routes can be calculated .The unused routes
are I:C,II:B,II:C,II:D,II:Dummy,III:D,and Dummy
D (I)
D (II).
D (II).
D (II)
D (II)
D (III)
D (III)
D (III)
+
+
+
+
+
+
+
+
R(C)
R (B)
R(C)
R (D)
R (Dummy)
R (A)
R (D)
R (Dummy)
=
=
=
=
=
=
=
=
Shadow
Costs
0+13
-2+16
-2+13
-2+ 2
-2+0
-11+3
-11+2
-11+0
=13
=14
=11
=0
=-2
=-8
=-9
=-11
The shadow costs are then deducted from actual costs

It will be seen that total cost can be reduced by 8 per unit for every unit that can be transferred
into Cell II:C
Step4.Make the maximum possible allocation of deliveries into Cell II:C.This is done by
inserting a sequence of +and -,maintaining row and column totals.
270
Available
Decesion Theory
I
II
III
40
20
50
A
25
5+
20-
Requirements
B
25178+
C
42
D
8
8
Dummy
10
10
42-
The maximum transferable number is the lowest number in the minus cell, i.e. 17. after the
transfer is made we get;
Available
I
II
III
40
20
50
A
25
22
3
B
250
25
C
42
D
8
8
Dummy
10
10
17
25
Step 3 is repeated again to check if the cost is minimum after setting D(I) = 0.
In our case after deducting shadow costs from actual costs we find that there are no more
negative numbers thus we deduce from the last table that the minimum transportation cost is,
(223) + (82) + (100) + (31) + (173) + (255) + (252) = Sh.311
Maximization using Transportation

Transportation problems are usually minimizing problems, on occasions problems are framed so
that the objective is to make the allocations from sources to destinations in a manner which
maximizes contribution or profit. These problems are dealt with similar to minimizing problems
but the reverse of it. i.e.
a) Make initial feasible allocation on basis of maximum contribution first, then next highest
and so on.
b) For optimum, the differences between actual and shadow contributions for the unused
routes should be all negative. If not, make allocation into cell with the largest positive difference.
c) In case there are more items available than are required, a dummy destination with zero
contribution should be introduced and the maximizing procedure in a). followed
8.2
Assignment Models
The following example will be used as a basis of the step-by-step explanation.
Example 1
A company employs services engineers based at various locations throughout the country to
service and repair their equipment installed in customers premises. Four requests for services
have been received and the company finds that four engineers are available. The distances each
of the engineers is from the various customers, is given in the following table and the company
wishes to assign engineers to customers to minimise the total distances to be travelled.
Customers
Alf
Bill
Charlie
Dave
W
25
38
15
26
X
18
15
17
28
Y
23
53
41
36
Z
14
23
30
29
270
Lesson Eight
271
Step 1. Reduce each column by the smallest figure in that column. The smallest figures are 15, 15,
23 and 14 and deducting these values from each element in the columns produces the following
table.
Table 2
A
B
C
D
W
10
23
0
11
X
3
0
2
13
Y
0
30
18
13
Z
0
9
16
15
Step 2 Reduce each row by the smallest figure in that row.

The smallest figures are 0, 0, 0 and 11 and deducting these values gives the following table.
A
B
C
D
W
10
23
0
0
Table 3
X
3
0
2
2
Y
0
30
18
2
Z
0
9
16
4
Note: Where the smallest value in a row is zero (i.e. as in rows A, B and C above) the row is, of
course, unchanged.
Step 3 Cover all the zero in the table 3 by the minimum possible number of lines. The lines may be
horizontal or vertical.
Table 4
A
B
C
D
W
10
23
0
0
X
3
0
2
2
Y
0
30
18
2
Z
0
9
16
4
Note: Line 3, covering Row B, could equally well have been drawn covering column X.
Step 4.Compare the number of lines with the number of assignments to be made (in this example
there are 3 lines and 4 assignments).If the number of line equals the number of assignments to be
made go to step 6.
If the number of lines is less than the number of assignments to be made (i.e. as in this example
which has three lines and four assignments) then
a) Find the smallest uncovered element from step 3, called X (in Table 4 this value is 2).
b) Subtract X to every element in the matrix.
c) Add back to every element covered by a line. If an element is covered by two lines, for
example, cell A: W in Table 4, X is added twice.
Note: The effect of these steps is that X is subtracted from all covered by one line remain
unchanged, and elements covered by two lines are increased by X.
272
Decesion Theory
Note: The effect of these steps is that X is subtracted from all uncovered elements, elements
covered by one line remains unchanged, and elements covered by two lines are increased by
X.
Carrying out this procedure on Table 4 produces the following results:
In Table 4 the smallest elements is 2. New table is
Table 5
A
B
C
D
12
25
0
0
3
0
0
0
0
30
16
0
0
9
14
2
Note: It will be seen that cells A: W and B: W have been increased by 2; cells A : X, A : Y,A :Z, B
:X,B:Y, B:Z, C:W and D:W are unchanged, and all other cells have been reduced by 2.
Step 5. Repeat steps 3 and step 4 until the number of lines covering the zero equals the number
of assignments without any further repetition, thus:
Table 6
A
B
C
D
12
25
0
0
3
0
0
0
0
30
16
0
0
9
14
2
Line 1
Line 2
Line 3
Line 4
Step 6 when the number of lines equals the number of assignments to be made, use the
following rules:
a) Assign to any zero which is unique to both a column and a row.
b) Assign to any zero which is unique to a column or a row.
c) Ignoring assignments already made repeat rule (b) until all assignments are
made.
Carrying out this procedure for our example results in the following:
a) (Zero unique to both a column and a row). None in this example.
b) (Zero unique column or row). Assign B to X and A to Z. The position is
now as follows.
Table 7
A
B
C
D
Row
Row
0
0
Satisfied
Satisfied
Column Satisfied
Column Satisfied
16
0
Column satisfied
Column satisfied
Column Satisfied
Column Satisfied
272
Lesson Eight
273
c) Repeating rule (b) results in assigning D to Y and C to W.

Notes:
a) Should the final assignment not be to a zero, then more lines than necessary were used in
step 3.
b) If a block of 4 or more zeros is left for the final assignment, then a choice of assignment
exits with the same mileage.
Step 7 Calculate the total mileage of the final assignment.
A to Z
Mileage
B to X
C to W
D toY
14
15
15
36
80 Miles
The assignment technique for maximising

A maximising assignment problem typically involves making assignments so as to maximise
contribution. To maximise only one step 1 from above differs-the columns are reduced by the
largest number in each column. From then on the same rules apply that are used for minimising.
Maximising example
Example 2
The previous example No.1 will be used with the changed assumptions that the figures
relate to contribution and not mileage and that it is required to maximise contribution
.The solution would be reached as follows.(In each case the step number corresponds to
the solution given for Example No 1.)
Original data
Table 8
A
B
C
D
W
25
38
15
26
X
18
15
17
28
Y
23
53
41
36
Z
14
23
30
29
Contributions
to be gained
Step 1: Reduce each column by the largest figure in that column and ignore the resulting signs.
Table 9
A
B
C
D
W
13
0
23
12
X
10
13
11
0
Y
30
0
12
17
Z
16
7
0
1
274
Decesion Theory
Step 2. Reduce each row by smallest figures in that row.

Table 10
A
B
C
D
W
3
0
23
12
X
0
13
11
0
Y
20
0
12
17
Z
6
7
0
1
Step 3.Cover zeros by minimum possible number of lines.

Table 11
A
B
C
D
W
3
0
23
12
X
0
13
11
0
Y
20
0
12
17
Z
6
7
0
1
Step 4. If a number of lines equals the number of assignments to be made go to step 6.If less, (as
in this example), carry out the uncovered element procedure previously described. This results
in the following table:
Table 12
A
B
C
D
W
0
0
20
9
X
0
16
11
0
Y
17
0
9
14
Z
6
10
0
1
Table 13
A
B
C
D
W
0
0
20
9
X
0
16
11
0
Y
17
0
9
14
Z
6
10
0
1
Step 6. Make assignment in accordance with the rules previously described which result in the
following assignment:
C to Z
D to X
A to W
B to Y
274
Lesson Eight
275
Step 7.Calculate contribution to be gained from the assignments.

C to Z
D to X
A to W
B to Y
Total
30
28
25
53
136
Notes:
a) It will be apparent that maximising assignment problems can be solved in virtually the
same manner as minimising problems.
b) The solution methods given are suitable for any size of matrix. If a problem is as small
as the illustration used in this chapter, it can probably be solved merely by inspection.
Unequal sources and destinations

5. To solve assignments problems in the manner described the matrix must be square, i.e. the
supply must equal the requirements. Where the supply and requirements are not equal, an
artificial source or destination must be created to square the matrix. The
cost/mileage/contributions etc for the fictitious column or row should be zero throughout.
Solution method
Having made the sources equal the destinations, the solutions method will be as normal, treating
the fictitious elements as though they were real. The solution method will automatically assign a
source or destination to the fictitious row or column and the resulting assignment will incur zero
cost or gain zero contribution.
Points to note
a) The assignment technique can be used for allocating type of problems, e.g. taxis to
customers, jobs to personnel.
b) Most practical problems of size illustrated could be solved fairly readily using
nothing more than commonsense. However, the technique illustrated can be used
to solve much larger problems.
Exercises with answers
1. A foreman has four fitters and has been asked to deal with five jobs. The times for each job
are estimated as follows:
Fitters
Alf
Job 1
Job 2
Job 3
Job 4
Job 5
6
22
12
16
18
Bill
Charlie
Dave
12
18
16
8
14
20
15
18
12
10
12
20
15
20
17
Allocate the men to the jobs so as to minimise the total time taken and identify the job which
will not be dealt with.
276
2.
Decesion Theory
A company has four salesmen who have to visit four clients. The profits records from
previous visits are shown in the table and it is required to maximise profits by the best
assignments.
A
6
22
12
16
12
18
16
8
20
15
18
12
12
20
15
20
1. Dummy fitter inserted to square matrix

A
B
1
6
12
2
22
18
3
12
16
4
16
8
5
18
14
C
20
15
18
12
10
D
12
20
15
15
17
Customer 1
Answers to exercises
DUMMY
0
0
0
0
0
Reduce columns by the smallest element and cover by lines

0
10
16
6
10
12
10
8
0
6
5
8
2
0
8
3
8
5
0
0
0
0
4 lines so not optimum, smallest element 3

Therefore reduce uncovered elements by 3 and increase elements crossed by 2 lines by 3
0
13
3
10
12
4
7
5
0
6
10
2
5
2
0
0
5
0
8
5
3
0
0
3
3
5 Lines so optimum.
Assignments
B to 4
C to 5
A to 1
Dummy to 2
8.3
NETWORK ANALYSIS
This is a system of interrelationship between jobs and tasks for planning and control of
resources of a project by identifying critical path of the project.
276
Lesson Eight
277
Terminology
Activity. Task or job of work, which takes time and resources e.g building a bridge. It is
represented by an arrow which indicates where the task begins and ends
Event (node). This is a point in time and it indicates the start or finish of an activity e.g in building
a bridge, rails installed. It is represented by a circle.
Dummy activity. An activity that doesnt consume time or resources, it is merely to show logical
dependencies between activities so as abide by rules of drawing a network, it is
represented by a dotted arrow
Network. This is a combination of activities and events (including dummy activities)
Rules for Drawing a Network

a) A network should only have one start point and one finish point (start event and finish
event )
b) All activities must have at least one preceding event (tail event) and at least one
succeeding event (head event), but an activity may not share the same tail event and head
event.
c) An activity can only start after its tail event has been reached
d) An event is only complete after all activities leading to it are complete.
e) Activities are identified by alphabetical or numeric codes i.e. A,B,C; 1,2,3 or
identification by head or tail events 1-2, 2-4, 3-4,1-4
f) Loops (a series of activities leading back to the same event) and danglers (activities
which do not link to the overall project) are not allowed
Loop
Dangling activity
Dummy Events
This is an event that does not consume time or resources, it is represented by dotted arrow.
Dummies are applied when two or more events occur concurrently and they share the same
head and tail events e.g. when a car goes to a garage tires are changed and break pads as well,
instead of representing this as;
A- Tires Changed
Car Arrives (CA)
Car ready (CR)

B- Break pads Changed
278
Decesion Theory
These events are represented as;
CA
CR
Example of a network.
Activities
1-2
- where 1 is the preceding event where as 2 is the succeeding event of the activity
1-3
2-4
2-5
3-5
4-5
4-6
5-6
6-7
4
6
1
3
8.4
Network Analysis-Time Analysis
Assessing the time
a) After drawing the outline of the network time durations of the activities are then
inserted.
a) Time estimates. The analysis of the projects time can be achieved by using :
i.
Single time estimates for each activity. These estimates would be based on the
judgment of the individual responsible or by technical calculations using data
from similar projects
Multiple time estimates for each activity. the most usual multiple time
ii.
estimates are three estimates for each activity , i.e. optimistic (O), Most Likely
(ML), and Pessimistic (P). These three estimates are combined to give an
expected time and the accepted time formula is:
Expected time =
O + P + 4ML
6
For example assume that the three estimates for an activity are
Optimistic
11 days
Most likely
15 days
Pessimistic
18 days
278
Lesson Eight
279
Expected time =
11 + 18 + 4 (15 )
6
= 14.8 days
b) Use of time estimates. as three time estimates are converted to a single time
estimate. There is no fundamental difference between the two methods as
regards the basic time analysis of a network. However, on completion of the
basic time analysis, projects with multiple time estimates can be further
analyzed to give an estimate of the probability of completing the project by a
scheduled date.
c) Time units. Time estimates may be given in any unit, i.e. minutes , hours, days
depending on the project. All times estimates within a project must be in the
same units otherwise confusion is bound to occur.
Basic time analysis critical path

b) The critical path of a network gives the shortest time in which the whole project can
be completed. It is the chain of activities with the longest duration times. There may
be more than one critical path which may run through a dummy.
Earliest start times (EST) Forward pass, Once the activities have been timed we can
assess the total project time by calculating the ESTs for each activity. The EST is
the earliest possible time at which a succeeding activity can start.
Assume the following network has been drawn and the activity times estimated in
days.
2
B
D
4
2
0
A
1
C
3
E
1
F
2
The ESTs can be inserted as follows.
EST
2
3
D
4
2
0
0
A
1
1
1
C
3
3
4
E
1
4
7
F
2
5
9
The method used to insert the ESTs is also known as the forward pass, this is obtained
by;
EST = The greater of [EST (tail event) + Activity duration]
280
Decesion Theory
a) Start from the start event giving it 0 values,

b) For the rest of the events EST is obtained by summing the EST of the tail event and the
activity duration
c) Where two or more routes converge into an activity, calculate individual EST per route
and then select the longest route (time)
d) The EST of the finish event is the shortest time the whole project can be completed.
Latest Start Times (LST) Backward pass. this is the latest possible time with which a
preceding activity can finish without increasing the project duration. After this operation
the critical path will be clearly defined.
From our example this is done as follows;
2
3 3
B
2
0
0 0
A
1
1
1 1
C
3
3
4 6
LST
D
4
E
1
4
7 7
F
2
5
9
LST = Lowest of [LST (head event) activity duration]

a) Starting at the finish event, insert the LST (i.e. 9 for our example) ,and work backwards
through the network.
b) deduct each activity duration from the previously calculated LST (i.e. head LST).
c) Where the tails of activities join an event, the lowest number is taken as the LST for that
event
Critical Path. . This is the chain of activities in a network with the longest duration Assessment
of the resultant network shows that one path through the network (A, B, D, F) has
EST's and LST's which are identical this is the critical path.
The critical path can be indicated on the network either by a different colour or by
two small transverse lines across the arrows along the path, thus in our example we
have;
2
3 3
B
2
0
0 0
A
1
1
1 1
C
3
3
4 6
D
4
E
1
4
7 7
F
2
5
9
280
Lesson Eight
281
Activities along the critical path are vital activities which must be completed by their
EST's/LST's otherwise the project will be delayed.
Non critical activities (in the example above, C and E) have spare time or float
available. C and/ or E could take up to an additional 2 days in total without
delaying the project duration. If it is required to reduce the overall project duration
then the time of one or more of the activities on the critical path must be reduced
perhaps by using more labour, or better equipment to reducing job times.
FLOAT
Float or spare time can only be associated with activities which are non-critical. By definition,
activities on the critical path cannot have float. There are three types of float, Total Float, Free
Float and Independent Float. To illustrate these types of float we use the following example.
5
10 20
B
10
6
40 50
Section of the network

a) Total float. Amount of time by which a path of activities could be delayed without
affecting the overall project duration. The path in this example consists of one activity
only i.e. B
Total Float = Latest Finish time (LFT) - Earliest Start time(EST) time Activity
Duration
Total Float = 50 - 10 - 10
= 30 days
b) Free float Amount of time an activity can be delayed without affecting the
commencement of a subsequent activity at its earliest start time, but may affect float of
a previous activity.
Free Float = Earliest Finish Time(EFT) - EST - Activity Duration
c)
Free Float = 40-10-10

= 20 days
Independent float. Amount of time an activity can be delayed when all preceding activities
are completed as late as possible and all succeeding activities commenced as early a
possible. Independent float therefore does not affect the float of either preceding or
subsequent activities.
Independent float = EFT- Latest Start time (EST) - Activity Duration

Independent float = 40 - 20 - 10
= 10 days
282
Decesion Theory
Note:
for examination purposes, float always refers to total float
The total float can be calculated separately for each activity but it is often useful to find the total float over
chains of non-critical activities between critical events
Example.
The following represents activities of a network.
Activity
Preceding Activity
Duration Days
A
4
B
A
7
C
A
5
D
A
6
E
B
2
F
C
3
G
E
5
H
B,F
11
I
G,H
7
J
C
4
K
D
3
L
I,J,K
4
Required:
a) Draw the network diagram and find the critical path
b) Calculate the floats of the network in question
Solution. (a)
E
2
4
13
G
5
23
H
11
11
5
12
I 7
B 7
F
1
0
A
4
2
4
C
5
30
J
4
6
9
10
L
4
34
7
10
282
Lesson Eight
283
First we draw the network structure ensuring it fits the data above
We then label all activities from 1 to 12 and indicate activity duration
Conduct a forward pass operation (to obtain the diagram above)
Operate backward pass to establish the critical path, thus we have
G
5
13 18
23 23
H
11 15
11
I 7
12 12
B 7
F 3
0 0
A
4
30 30
C
5
J
4
L
4
34 34
10 27
Therefore we get the critical path to be, A- C- F- H- I- L
b) The floats of the network,
Activity
*A
B
*C
D
E
*F
G
*H
*I
J
K
*L
EST
0
4
4
4
11
9
13
12
23
9
10
30
LST
EFT
0
4
4
4
15
9
21
12
23
9
22
30
4
11
9
10
13
15
23
23
30
30
30
34
Activity
Duration
LFT
D
4
15
9
22
21
15
23
23
30
30
30
34
4
7
5
6
2
3
5
11
7
4
3
4
Total Float
Free Float
LFT -EST- D
EFT-EST-D
Independent
Float
EFT-LST-D
4
12
8
5
17
17
-
5
17
17
-
17
5
-
284
Decesion Theory
The total float on the non-critical chains are;

Non-critical
chain
B,E,G
B,Dummy
D,K
J
Time required
(sum of duration)
14
7
9
4
Time available
(LFT of last activity-EST of 1st activity)
19
8
26
21
Total Float over

chain
5
1
17
17
Slack
This is the difference between the EST and LST for each event. Strictly it does not apply to
activities but on occasions the terms are confused in examination questions and unless the
context makes it abundantly clear that event slack is required, it is likely that some form of
activity float is required. Events on the critical path have zero slack.
8.5
Cost Scheduling
This is done by calculating the cost of various project durations, cost analysis seeks to find the
cheapest way of reducing the overall cost duration of a project by increasing labour hours,
equipment e.t.c.
Terminologies
Normal cost. The costs associated with a normal time estimate for an activity. Often the normal
time estimate is set at the point where resources (labour, equipment, etc.) are used
in the most efficient manner.
Crash cost. The costs associated with the minimum possible time for an activity. Crash costs,
because of extra wages, overtime premiums, extra facility costs are always higher
than normal costs.
Crash time. The minimum possible time that an activity is planned to take. . The minimum time is
invariably brought about by the application of extra resources, e.g. more labour or
machinery.
Cost slope. This is the average cost of shortening an activity by one time unit (day, week, month as
appropriate). The cost slope is generally assumed to be linear and is calculated as
follows:
Cost slope = Crash cost Normal cost

Normal time Crash time
Example
A project has the following activities and costs. You are required to prepare the least cost
schedules for all possible durations from normal time normal cost to crash time crash cost.
284
Lesson Eight
Activity
285
Preceding
Activity
A
A
B,C
A
B
C
D
E
Duration
days
4
8
5
9
5
Crash
time
3
5
3
7
3
1
4
Cost
(Shs).
360
300
170
220
200
60
70
50
40
80
3
14 14
C
4
0
0
420
510
270
300
360
D
9
Crash cost Cost slope
B
8
2
9
Project duration and costs

(a)
Normal duration = 14 days

Critical path = A,C,E
Project cost (cost of all activities in normal time) = Shs. 1,250.
(b)
Reduce by 1 day the activity on the critical path with the lowest cost slope. Thus we
reduce C at extra cost of Shs. 50.
Now
Project duration = 13 days
Project cost
= Shs. 1,300
Note: that all activities are now critical.
(c)
Further reducing the critical path by 1 day will require that more than one activity is
affected because there exist several critical paths.
Reduce by 1 day
A and B
D and E
B, C and D
A and E
Extra cost
60 + 70 = 130
40 + 80 = 120
70 + 50 + 40 = 160
60 + 80 = 140
Activities critical
All
All
All
A, D, B, E
From this we realize that reducing D and E is the cheapest.

However closer examination of the fourth alternative reveals that C is now non-critical
and has 1 day float. Since we earlier reduced C for Shs. 50, if we reduce A and E and
increase C by a day which will save Shs. 50.
286
Decesion Theory
Then the net cost for 12 day duration = 1,300 + (140 50) = 1,390.
The network becomes
1
3
D
3
3 (crash)
12
12
5
C
E
4
0
0
B
0
(d)
Next we reduce D & E

Project duration = 11 days
Project cost = 1,510
Critical activities = All
(e)
Final reduction possible is by reducing B, C & D for Shs. 160 the network then
becomes.
1
3
A
D
3
7 (Crash)
10
10
3 (crash)
4
C
E
3 (crash)
0
0
B
0
2
7
Duration = 10 days
Cost = Shs. 1,670
Critical activities = All.
Note: only critical activities affect project duration.
: Always look for a possibility of increasing the duration of a previously
crashed activity.
286
Lesson Eight
287
SCHEDULING RESOURCES AND GANTT CHART

Apart from time, cost network analysis also help in controlling and planning of resources.
Example
A project has the following activity durations and resource requirements.
Activity
A
B
C
D
E
F
Preceding activity
C
B
D
Duration (days)
6
3
2
2
1
1
Resource requirement (man power)

3
2
2
1
2
1
Required
i)
What is the networks critical path
Draw a gantt chart diagram indicating activity times, using their estimate.
ii)
Show resource requirement on a day to day basis assuming all events commence at their
iii)
estimates.
Assuming that only six employees are available, how will the activities be planned for?
iv)
Solution
i)
Activities
A
B
C
D
E
F
Duration
6
3
2
2
1
1
EST
0
0
0
2
3
4
LST
0
0
0
3
5
5
Man power
3
2
2
1
2
1
288
Decesion Theory
ii)
A gantt chart or a bar chart. This is a diagram indicating a resource scaled

network.
iii)
Resource requirements on a day to day basis.
iv)
When on 6 manpower resources are available then we adjust the activities to

accommodate this and still end at the given critical time duration i.e.
288
Lesson Eight
289
Node Networks
This network also known as a procedure diagram is represented with the same information as a
network diagram.
Its characteristics are;
Activities are shown in boxes instead of arrows
i)
Events are not represented.
ii)
iii)
The arrows linking boxes indicate the sequence precedence of activities.
Dummies arent necessary.
iv)
E.g.
Would appear as
290
Decesion Theory
A full activity node network is represented as;
This is represented as;
Note:
i)
EST and LST are calculated by the same process we learnt earlier.
EFT and LFT are calculated by adding the activity time duration to EST and LST
ii)
respectively.
iii)
Critical path is similarly identified by identifying equal EST and LST throughout the
path.
290
Lesson Eight
291
QUESTION ONE
Regal Investments has just received instructions from a client to invest in two shares; one an
airline share, the other an insurance share. The total maximum appreciation in share value over
the next year is to be maximized subject to the following restrictions:
the total investment shall not exceed Sh.100,000
at most Sh.40,000 is to be invested in the insurance shares
quarterly dividends must total at least Sh.2,600
The airline share is currently selling for Sh.40 per share and its quarterly dividend is Sh.1per
share. The insurance share is currently selling for Sh.50 per share and the quarterly dividend is
Sh.1.50 per share. Regals analysts predict that over the next year, the value of the airline share
will increase by Sh.2 per share and the value of the insurance share will increase by Sh.3 per
share.
A computer software provided the following part solution output:
Objective Function Value = 5,400
Airline shares
Insurance shares
Constraint
Total investment
Investment in insurance
Dividends
Variable Number
1,500
800
Reduced cost
0.000
0.000
Slack/Surplus
0.000
0.000
100.000
Dual prices
0.050
0.010
0.000
292
Decesion Theory
Objective Coefficient Ranges

Variable
Airline share
Insurance share
Lower limit
2.500
0.000
Current value
3.000
2.000
Upper limit
No upper limit
2.400
Right-hand Side Ranges

Constraint
Total investment
Dividends
Lower limit
96,000.00
20,000.00
No lower limit
Current value
100,000
40,000
2,600
Upper limit
No upper limit
100,000.00
2,700.00
Required:
a)
b)
c)
d)
Formulate the above problem.

Explain what reduced cost and dual prices columns above mean.
How should the clients money be invested to satisfy the restrictions?
Suppose Regals estimate of the airline shares appreciation is an error, within what limits
must the actual appreciation lie for the answer in (c) above to remain optimal?
(Q 6 Dec 2001)
QUESTION TWO
a)
A baker makes two products; large loaves and small round loaves. He can sell up to 280 of
the large loaves and up to 400 small round loaves per day. Each large loaf occupies 0.01m3
of shelf space, each small loaf occupies 0.008m3 of space, and there is 4m3 of shelf space
available. There are 8 hours available each night for baking, and he can produce large loaves
at the rate of 40 per hour, and small loaves at the rate of 80 per hour. The profit on each
large loaf is Sh.5.00 and Sh.3.00 profit on the small round loaf.
Required:
In order to maximize profits, how many large and small round loaves should he produce?
b)
Summarize the procedure for solving the kind of quantitative technique you have used to
solve part (a) above.
(Q 6 June 2001)
QUESTION THREE
a)
A small company will be introducing a new line of lightweight bicycle frames to be made
from special aluminium alloy and steel alloy. The frames will be produced in two models,
deluxe and professional. The anticipated unit profits are currently Sh.1,000 for a deluxe
frame and Sh.1,500 for a professional frame. The number of kilogrammes of each alloy
needed per frame is summarized in the table below. A supplier delivers 100 kilogrammes of
the aluminium alloy and 80 kilogrammes of the steel alloy weekly.
Deluxe
Professional
Aluminium alloy
2
4
Steel alloy
3
2
Required:
i)
Determine the optimal weekly production schedule.
292
Lesson Eight
ii)
b)
293
Within what limits must the unit profits lie for each of the frames for this solution to
remain optimal?
Explain the limitations of the technique you have used to solve part (a) above.
(Q 6 Dec 2000)
QUESTION FOUR
a)
Define the following terms as used in linear programming:

i) Feasible solution
ii) Transportation problem
iii) Assignment problem
b)
The TamuTamu products company ltd is considering an expansion into five new sales
districts. The company has been able to hire four new experienced salespersons. Upon
analysing the new salespersons past experience in combination with a personality test
which was given to them, the company assigned a rating to each of the salespersons for
each of the districts .These ratings are as follows:
c)
Salespersons
A
B
C
D
Districts
1
2
92
90
84
88
90
90
78
94
3
94
96
93
89
4
91
82
86
84
5
83
81
93
88
The company knows that with four salespersons, only four of the five potential districts can
be covered.
Required:
i) The four districts that the salespersons should be assigned to in order to maximize the
total of the ratings
(Q 6 June 2002)
ii) Maximum total rating.
QUESTION FIVE
a)
Explain the value of sensitivity analysis in linear programming problems and show how
dual values are useful in identifying the price worth paying to relax constraints.
b)
J.A Computers is a small manufacturer of personal computers. It concentrates on

production of three models- a Desktop 386, a Desktop 286, and a Laptop 486, each
containing one CPU Chip. Due to its limited assembly facilities JA Computers are unable to
produce more than 500 desktop models or more than 250 Laptop models per month. It has
one hundred and twenty 80386 chips (these are used in Desktop-386) and four hundred
80286 chips (used in desktop 286 and Laptop 486) for the month. The Desktop 386 model
requires five hours of production time, the Desktop 286 model requires four hours of
production time, and the Laptop 486 requires three hours of production time. J.A
Computers have 2000 hours of production time available for the coming month. The
company estimates that the profit on Desktop 386 is Sh. 5,000. for a desktop 286 the profit
is Sh.3,400 and Sh.3,000 profit for a laptop 486.
294
Decesion Theory
Required:
Formulate this problem as a profit maximization problem and mention the basic
assumptions that are inherent in such models.
c)
An extract of the output from a computer package for this problem is given below:
Output solution
X1=120, X2 = 200, X3 = 200
Dual values
Constraints 3 150
Constraints 4 90
Constraints 5 20
Sensitivity analysis of objective function coefficients:
Variable
X1
X2
X3
Lower
limit
100
150
127.5
Original
value
250
170
150
Upper
limit
No limit
200
170
Lower
limit
320
200
80
350
1950
Original
value
500
250
120
400
2000
Upper
limit
No limit
No limit
130
412.5
2180
Sensitivity analysis on R.H.S ranges.
Constraints
1
2
3
4
5
X1=Monthly production level for Desktop 386.

X2 =Monthly production level for Desktop 286.
X3=Monthly production level for Laptop 486.
Required:
i) Interpret the output clearly, including optimum product mix, monthly profit, unused
resources and dual values
ii) Explain the purpose of upper limits and lower limits for the variables X1,X2,X3 and
constraints 1 to 5.
iii) Calculate the increase in profit if the company is able to produce a further 10 CPU
80386 chips.
(Q7 July 2000 Pilot paper)
QUESTION SIX
Preface Retailers is a high-technology retailer and mail order business. In order to improve its
process the company decides to install a new microcomputer system to manage its entire
operation (i.e. payroll, accounts, inventory).
Terminals for each of its many stores will be networked for fast, dependable service. The
specific activities that Preface will need to accomplish before the system is up and running are
listed below. The table also includes the necessary increased staffing to undertake the project.
294
Lesson Eight
295
Activity
A. Build insulated enclosure
B. Decide on computer system
C. Electrical wiring of room
D. Order and collect computer
E. Install air conditioning
F. Install computer
G. Staff testing
H. Install software
I. Staff training
Preceding
Activities
A
B
A
D, E
B
C, F
G, H
Duration
(Days)
4
1
3
2
4
2
5
2
3
Increased
Staff
1
3
2
1
2
2
1
1
1
Required:
a)
b)
c)
Draw a network diagram for the project and determine the critical path and its
duration.
Assuming that all activities start as soon as possible, draw a progress chart for the
project, showing the times at which each activity takes place and the manpower
requirements.
The union has decided that any staff employed on the project must be paid for the
duration of the project whether they work or not, at a rate of 500 per day.
Assuming that the same staff is employed on the different activities, determine the
work schedule that will minimise labour costs though not necessarily the project
time. What is the cost associated with this schedule?
Comment on the validity of the assumption.
296
Decesion Theory
COMPREHENSIVE ASSIGNMENT FOUR
Instructions:

SECTION I
QUESTION ONE
a)
i)
ii)
iii)
iv)
Define the following terms as used in game theory:

Dominance.
Saddle point.
Mixed strategy.
Value of the game
(2 marks)
(2 marks)
(2 marks)
(2 marks)
b) Consider the two person zero sum game between players A and B given the following payoff table:
Player A Strategies
1
2
Player B Strategies
1
2
2
4
3
2
-1
6
2
3
Required:
i) Using the maximin and minimax values, is it possible to determine the value of the game?
Give reasons.
(3 marks)
ii) Use graphical methods to determine optimal mixed strategy for player A and determine
the value of the game.
(9 marks)
(Total: 20 marks)
QUESTION TWO
Central and Eastern Industries is planning to introduce a new mobile phone service. To do so,
the following activities are necessary:
Preceding
Expected
Standard
Activity
Activity
Time (weeks) Deviation (weeks)
A
6
1.0
B
3
0.5
C
A
5
1.0
D
A
4
1.0
E
A
3
0.5
F
C
3
0.5
G
D
5
1.0
H
B.D.E
5
1.0
I
H
2
0.5
J
F.G.I
3
1.0
The costs of the project are estimated to be Sh.10 million. If the projects is completed within 24
weeks the expected net revenue will be about Sh.100 million but if the deadline of 24 weeks is
296
Lesson Eight
297
not met, the service will fail to penetrate the market and a net revenue of Sh.2 million is
expected.
Required:
a) Determine how long the project would take.
b) If the start of activities B, E and G are respectively delayed by 3,2 and 2 weeks,
how would this affect the total project time?
c) Determine a 95% confidence interval for the expected time of the project
and explain your answer. Ignore the delays referred to in (b) above.
d) Determine the expected profit on this project. Again ignore the delays referred
in (b) above
(Total:
(8 marks)
(6 marks)
(3 marks)
(3 marks)
20 marks)
QUESTION THREE
a) Explain the value of sensitivity analysis in linear programming problems and show how dual
values are useful in identifying the price worth paying to relax constrains.
(4marks)
b) J.A Computers is a small manufacturer of personal computers. It concentrates production
on three models, a Desktop 286,and a laptop 486,each containing one CPU Chip. Due to
its limited assembly facilities J.A Computers are unable to produce more than 500 desktop
model or more than 250 Laptop models per month. It has one hundred and twenty 80386
chips (these are used in Desktop 386) and four hundred 80286 chips(used in desktop 286
model requires four hours of production time, and the laptop 486 requires three hours of
production time. J.A Computers have 2000hours of production time available for the
coming month. The company estimates that the profit on a desktop 386 is Sh.5000, for a
desktop 286 the profit is Sh.3400 and Sh.3000 profit for a Laptop 486.
Required:
Formulate this problem as a profit maximization problem and mention the basic
assumptions that are inherent in such models.
(7 marks)
a) An extract of the output from a computer package for this problem is given below:
Output solution
X1 = 120, X2 = 200, X3 = 200
Dual values
Constraints 3
Constraints 4
Constraints 5
150
90
20
Sensitivity analysis of objective function coefficients

Variable
Lower limit
Original value
100
250
X1
150
170
X2
X3
127.5
150
Sensitivity analysis on R. H. S ranges
Constraints
Lower limit
1
320
2
200
3
80
4
350
Original value
500
250
120
400
Upper limit
No limit
200
170
Upper limit
No limit
No limit
130
412.5
298
Decesion Theory
5
X1
X2
X3
1950
=
=
=
2000
2180
Monthly level for Desktop 386.

Monthly production level for Desktop 286
monthly production level for Laptop 486
Required:
i) Interpret the output clearly, including optimum product mix, monthly profit,
unused resources and dual values.
ii) Explain the purpose of upper limits for the variables X1, X2 ,X3 and
constraints 1 to 5.
iii) Calculate the increase in profit if the company is able to produce a further
10 CPU 80836 chips.
(Total:
(3 marks)
(3 marks)
(3 marks)
20 marks)
QUESTION FOUR
a) Give two applications of simulation in business.
(2 marks)
b) Collins Simiyu recently acquired a piece of land in Kitale. A property development company
has offered him Sh.300,000 for the piece of land. He has to make a decision on whether to
cultivate the piece of land or to sell it to the property development company. if he decides
to cultivate the land, there is a probability of getting a high, medium or low harvest. The
expected net income for each of the above states of harvest is shown below:
State of harvest
High
Medium
Low
Net income (Sh.)

500,000
100,000
(200,000)
From past experience, there is a 10 per cent probability that the harvest will be low, a 30 pr cent
probability that the harvest will be medium and a 60 per cent probability that the harvest will be
high. Collins Simiyu can engage an agricultural expert to carry out a survey on the productivity
of the land, which will cost him Sh.30,000. The agricultural expert gives the following
information as to the reliability of such surveys (prior probabilities).
Results of survey
Accurate
Not accurate
State of harvest
High
Medium
0.35
0.10
0.10
0.25
0.20
0.60
Low
0.05
0.15
0.20
Total
0.5
0.5
1.0
Required:
i) Construct a decision tree for the above problem.
ii) The expected monetary value for each decision
iii) The decision that you would recommend
(6 marks)
(10 marks)
(2 marks)
(Total: 20 marks)
QUESTION FIVE
a) Explain the difference between assignment and transportation problems.
(4 marks)
b) State the assumptions made in solving a transportation problem.
(4 marks)
298
Lesson Eight
299
c) Umoja Engineering Works Ltd. Has a network of branches all over Kenya. The branches
are used to service, repair and install equipment for their clients. Currently, the Nairobi
branch has four clients who require installation of equipment. Each client requires the
services of one engineer.
There are four engineers who are not engaged at the moment and can be assigned any one of the
tasks. However, these engineers have to travel from different locations and the Nairobi branch
has to meet their travel and subsistence allowances. The allowances vary from one engineer to
another and according to the client the engineer has been assigned to work for.
The table below shows the costs (in thousands of shillings) associated with each engineer.
Engineer
A
B
C
D
1
37.0
57.0
22.0
39.0
2
27.0
22.0
25.0
42.0
Client
3
34.0
79.0
61.0
54.0
4
21.0
34.0
45.0
43.0
Required:
i) The assignments to be made in order to minimize the total cost of the
engineers.
ii) The minimum cost of using engineers.
(10 marks)
(12 marks)
(Total: 20 marks)
SECTION II
QUESTION SIX
a) Define the following terms as used in the network analysis:
i)
ii)
iii)
iv)
v)
Crash time
Optimistic time
Forward pass
Dummy activity
Slack
(2 marks)
(2 marks)
(2 marks)
(2 marks)
(2 marks)
b) James Mutiso is a computer engineer in an information technology firm. The firm has
decided to install a new system to be used by the firms help desk. James Mutiso has
identified some activities required to complete the installation.
The table below provides a summary of the activities durations and the required number of
technicians:
Activity
1-2
1-3
2-4
2-5
3-4
3-6
4-5
5-6
6-7
Duration (Weeks)
3
1
3
2
2
4
2
2
2
Required number of technicians

2
4
4
2
4
4
2
2
2
300
Decesion Theory
Required:
i) Draw a gantt chart for the project.
(6 marks)
ii) Mr. Mutiso would like to reschedule activities so that not more than 6 technicians
are required each week.
Determine if this is possible and how it can be achieved by rescheduling
the activities.
(4 marks)
(Total: 20 marks)
QUESTION SEVEN
Regal investments has just received instructions from a client to invest in two shares; one an
airline share, the other an insurance share. The total maximum appreciation in share value over
the next year is to be maximized subject to the following restrictions:
The total investment shall not exceed Sh.100,000
At most Sh.40,000 is to be invested in the insurance shares
Quarterly dividends must total at least Sh.2,600
The airline share is currently selling for Sh.40 per share and its quarterly dividend is Sh.1 per
share. The insurance share is currently selling for Sh.50 per share and the quarterly dividend is
Sh.1.50 per share. Regals analysts predict that over the next year, the value o f the airline share
will increase by Sh.2 per share and the value of the insurance share will increase by Sh.3 per
share.
A computer software provided the following part solution output:
Objective Function Value = 5,400
Variable
Airline shares
Insurance shares
Number
1,500
800
Reduced cost
0.000
0.000
Constraint
Total investment
Dividends
Slack/Surplus
0.000
0.000
100,000
Dual prices
0.050
0.010
0.000
Objective coefficient Ranges

Variable
Insurance share
Airline share
Lower limit
2.5000
0.000
Current value
3.000
2.000
Upper limit
No upper limit
2.400
Right-hand Side Ranges

Constraint
Total investment
Dividends
Lower limit
96,000.00
20,000.00
No lower limit
Current value
100,000
40,000
2,600
Upper limit
No upper limit
100,000.00
2,700.00
Required:
i) Formulate the above problem.
ii) Explain what reduced cost and dual prices columns above mean.
iii) How should the clients money be invested to satisfy the restrictions?
iv) Suppose Regals estimate of the airline shares appreciation is in error, within
(7 marks)
(5 marks)
(4 marks)
300
Lesson Eight
301
what limits must the actual appreciation lie for the answer in (c) above to
remain optimal?
(4 marks)
(Total: 20 marks)
QUESTION EIGHT
The linear programming model and output model below represent a problem whose solution
will tell a road side kiosk owner how many of the four different types of household goods to
stock in order to maximize profits. It is assumed that every item stocked will be sold. The
variables measure the packets of Unga, Spaghetti, Rice and Sugar to stock respectively. The
constraints measure storage space in units, special display racks, demand and a marketing
restriction, respectively.
Maximize Z = 4X1+ 6X2+ 5X3+ 3.5X4
Subject to :
2X1 + 3X2+ 3X3+ X4 120
1.5X1 + 2X2
2X2 + X3 + X4
X2 + X3
Where X1
X2
X3
X4
=
=
=
=
(1)
54
72
12
(2)
(3)
(4)
packets of Unga
packets of Spaghetti
packets of Rice
packets of Sugar
Optimal solution
Variable
X1
X2
X3
X4
Value
12.00
0.00
12.00
60.00
Reduced cost
0.50
-
Constraint
1
2
3
4
Stack/surplus
-
Dual/shadow price
2.00
1.50
-2.50
Objective Coefficient Ranges

Variable
Lower limit
X1
1.50
No limit
X2
4.50
X3
X4
3.00
Right Hand Side Ranges
Constraint
Lower limit
1
96
2
18
3
24
4
0
Current value
Current value
-
Upper limit
5.00
6.50
7.50
No limit
Upper limit
168
No limit
96
24
302
Decesion Theory
Required:
a) Determine the retailers optimal profit level.
b) Determine and interpret the missing values under:
i)
ii)
Reduced cost column.

Slack/surplus column, indicate whether the value is a slack or a surplus.
iii)
iv)
v)
Dual/shadow price column.

Current value column under objective coefficient ranges.
Current value column under right hand side ranges.
c) Interpret the value 0.50 under the reduced cost column and values; 2.00, 1.50
and -2.50 under the dual/shadow price column
d) Determine whether the current, optimal solution would change if the
current profit of packets of Unga is increased by Sh.2.00.
e) Determine by how much the amount of space would increase before there
is a change in the dual/shadow price.
f) The above problem could have been solved manually. Explain how the
optimal solution can be determined using the manual approach.
(Total:
(2 marks)
(2 marks)
(2 marks)
(2 marks)
(2 marks)
(2 marks)
(2 marks)
(2 marks)
(2 marks)
(2 marks)
20 marks)
302
Revision Aid
303
LESSON NINE
Revision Aid
SOLUTIONS
LESSON ONE
QUESTION ONE
1. (3, 12)
2. (12, 0)
4. ( 8 ,
3. (10, 4)
8)
QUESTION TWO
a) x = 0, y = 4; b) x = 3, y = 3; c) x = 10.5, y = 36
QUESTION THREE
Equillibrium will be attained
Population =
Pnorth 23.33
=
Psouth 46.67
QUESTION FOUR
1.
60

The vector for final demand = 60

60
The input/output coefficient matrix (also called the matrix for technical coefficients). Call it
A
80
320
80
A = 320
80
320
100
400
200
400
100
400
100
300
60
300
100
300
14
= 1
4
1 4
1
1
1
4
2
4
1
5
1
3
1
let X be the total output vector.

Thus;
X = AX + Y
X = (I A )-1Y
304
Lesson Nine
1 0 0 1 4
I A = 0 1 0 1 4
0 0 1 1 4
17
60
240 13
60
Inverse (I - A)-1=
23 3
16
1
4
5
12
1
4
13
60
7
30
5
16
1
1
1
4
2
4
3 4 1 4 13
1 = 1
1
1 5
5
2
4
1
1 4 1 4 2 3
3
1
68 60 52
= 1 52 100 56
23
45 60 75
68 60 52 60
10,800
1
1
Therefore X =
52 100 56 60 = 17, 480
23
23
45 60 75 60
10,800
469.57
= 542.60 in Shs
469.57
QUESTION FIVE
i)
Express each percentage as a decimal and the matrix equation becomes,
0.4 0.6 t1 b1
0.5 0.25 t = b
2 2
ii)
Put t1= 400 and t2= 700 and the matrix equation becomes.
0.4 0.6 400 b1

0.5 0.25 700 = b
2
b1 = (0.4 400) + (0.6 700)
= 580 kilos
That is
b2 = (0.5 400) + (0.25 700)

= 375 kilos
iii)
To establish the value of t1 and t2 which correspond to b1 and b2 it is necessary to form

the inverse of the matrix.
0.4 0.6
0.5 0.25
which is
Revision Aid
305
1.25 3.0
2.5 2.0
the equation to determine t1 and t2 now becomes,
t1 1.25 3.0 600

t = 2.5 2.0 450
2
Thus;
t1 = - 1.25600 + 3.0 450
= 600 kilos
and
t2= - 2.5600 - 2 .0 450
= 600 kilos
QUESTION SIX
i)
2
A3 =
3
-3
-3
10
-2
A2 =
ii)
F(A)
10
-2
-3
15
14
26
39
-51
=
3
-3
A3 3A2 2A + 41
14
26
-3
15
10
-2
-3
-2
39
-51
-3
15
-3
14
26
30
39
-51
-9
45
-6
-16
28
42
-86
A-1 =
1_
-12
iii)
-2
=
-3
1
4
1
4
1
0
-3
+4
1
6
-1
6
306
Lesson Nine
QUESTION SEVEN
i)
12
10
15
15
20
18
16
24
22
=
=
=
348 pence
536 pence
660 pence
Cost of Standard set

Cost of Deluxe set
Cost of super set
348
=
536
660
=
=
=
3.48
5.36
6.60
ii)
10
15
16
3,000
100
15
20
24
10
15
50)15
20
16
24
4,300
80
or (100 80
1,480
50
620
Squares 3,000, triangles 4,300 hexagons 1,480, octagons 620

iii)
12
8
(3,000 4,300 1,480 620)
18
20
= 110,680 in pence
= 1,106.80
QUESTION EIGHT
a)
NP =
10
60
12
0
50
20
storage maintenance
2
3
2
Cost of storing in warehouse Y

Cost of storing in warehouse W
Cost of maintaining warehouse Y
Cost of maintaining warehouse W
0.5
1.5
0.5
= 156p
=160p
= 48p
= 40p
156
160
48
40
=
=
=
=
1.56
1.60
0.48
0.40
b)
A
storage maintenance
Revision Aid
Day I
307
Y
W
Day II
10
40
10
0
50
20
10
40
10
15
60
20
2
3
2
0.5
1.5
0.5
150
120
2
3
2
0.5
1.5
0.5
191
165
c)
10
150
45
C
191
60.5
165
52.5
+3
120
30
cost in
pence
storage maintenance
For total cost add Day I and Day II values

A
45
30
Cost in pence
60.5
52.5
cost in
pence
308
Lesson Nine
LESSON TWO
QUESTION ONE
a)
b)
2
x3
1
1 + 2x
c)
y=
d)
1
,
x
dy
1 3
= x 2
dx
2
QUESTION TWO
dc
= 2Q 30
dQ
Therefore Q = 15 at minimum
Note:
d 2c
= 2 which is positive indicating a minimum value
dQ 2
QUESTION THREE
n() = 250
P + 12 + 59 = 147 giving P = 76
Q + 59 + 37 = 102 giving Q = 6
i)
Those who did not vote

= 250 (76 + 12 + 14 + 59 + 6 + 37)
= 250 204 = 46
ii)
x = 76 + 12 + 14 = 102
Revision Aid
309
x = 12 + 59 + 6 = 77
z = 37 + 14 + 6 = 57
iii)
x won the election.
QUESTION FOUR
i)
To find maximum or minimum value we use differential calculus as follows:
x3
R = 14 + 81x
12
dR
3x 2
= 81
dx
12
d 2R
6x
x
= 0
=
2
dx
12
2
i.e. 81 x x=2 0
81
= 0 thus
Which gives x = 18 or = -18

x =18 or 18
Check for a maximum or a minimum
d2R
x
=
2
2
dx
d2R
= -9
dx 2
when x = 18,
which is negative
Therefore at x = 18, the value of R is maximum

Similarly at x = -18, the value of R is minimum.
Therefore, the number of units that maximize the revenue = 18 units
ii)
The maximum revenue is given by

R = 14 + 81 18
(18 )3
12
= Ksh.986
iii)
The price per unit to maximize the revenue is

986
18
= 45.78
i.e. Ksh.54.78
QUESTION FIVE
i)
Profit is maximized when
310
Lesson Nine
Marginal Revenue = Marginal cost

Note: Profit = Revenue Cost
=RC
grad of = Grad of R grad of C
= (marginal Revenue marginal Cost)
Grad of = 0 for maximum and minimum.
PRODUCTION DEPARTMENT
Marginal Revenue = Revenue per unit = 6 + 6 0.8
= Shs.10.8 per unit
Cost per unit =
Therefore
20,000
20,000
Q
Giving Q =
+ 6 + 0.0002 (marginal cost)
+ 6.0002 = 10.8
20,000
4.7998
= 4,166.84
= 4,167 units
ii)
SALES DEPARTMENT
Profit = Revenue Cost
Revenue = Sales price x Quantity = p(40,000 2,000p)
= 40,000p 2,000p2
Cost
=
=
=
=
Profit =
2 X q + 6,000 + 6q + 80% of 6q
2q + 6p + 4.8q + 6,000
12.8q + 6,000
12.8 (40,000 2,000p) + 6,000
(40,000p 2,000p2) [6,000 + 12.8(40,000 2,000p)]
= 40,000p 2,000p2 6,000 512,000 + 25,600p

= -2,000p2 + 65,600p 518,000
d
dp
-4,000p + 65,600
Revision Aid
d 2
dp 2
P=
311
65,600
400
-4,000p + 65,600
When
d
dp
= 0
16.4
Selling price = Shs.16.4

iii)
Quantity 4,167 (as in i)

Selling price Shs 16.4 (as in ii)
Firms Profit =
Total Revenue Total Cost
Revenue
= 4,167 16.4 = Shs.68,338.8
Cost of production
Production Department
Sales Department
Total
20,000 + 6.0002 4,167

6,000 + 2 4,167
= 26,000 + 25,002.833 + 8,334

= 59,336.83
= Shs.59,336.85
Profit
= 68,338.8 59,336.85
= Shs.9,002
iv)
Quantity and sales price that maximizes the shops profit.

Revenue
(40,000 2,000p)p
[Q = 40,000 2,000p]
Cost
(40,000 2,000p) (6 + 0.0002 + 2) + 20,000 + 6,000
Profit
(40,000p 2,000p2) [(40,000 2,000p) (8.0002) + 26,000]
40,000p 2,000p2 320,008 + 16,000.4p - 26,000
-2,000p2 + 56,000.4p 340,008
For maximum profit

d
dp
d
dp
=0
and
d 2
is ve
dp 2
-4,000p + 56,000.4
312
Lesson Nine
Therefore
d 2
dp 2
-4,000p + 56,000.4 = 0 giving p = 1.0001 = Shs.14

= -4,000
which is ve
for maximum profit =
Quantity
-2,000 (14) + 56,000.4 14 340,008
-392,000 + 784,005.6 340,008
47,998
40,000 2,000 14
12,000 units.
QUESTION SIX
a) Quadratic functions in decision making.
Due to economies of scale, the cost of production is usuall dependent upon volume of sales.
Total revenue = sale price per unit quantity sold (say x)
Sales price is usually a function of x; f(x).
; which may be a quadratic function
Hence total revenue = f(x) x
Using calculus techniques (i.e. maxima and minima) we can calculate the optimum value
of x, which might give maximum profits or minimum costs.
b) Demand function p = 400 q
(price p in shillings and q in Kg)
Average total cost of prodcing the quantity

Hence total cost
100
+ 100 5q + q 2
q
1000
=
+ 100 5q + q 2 q = 1000 + 100q 5q 2 + q 3
q
Revenue = Price Quantity = p q = (400 q) q = 400q q2

Profit = (400q q2) (1000 + 100q 5q2 + q3)
= 300q + 4q2 q3 1000
i.
1000
Fixed cost
=
(100 is fixed cost)
q
Quantity produced and sold
= average fixed cost/unit
as quantity sold gradually increases, average fixed cost per unit
decreases
ii.
for maximum profit
Revision Aid
313
= 300 q + 4 q 2 q 3 1000
d
2
dq
= 300 + 8q 3q (puting
we get q =
d
=0 )
dq
8 60.53
11.422 Kg (note that the negative part of q is invalid)
6
d 2
= 8 6q
dq 2
for q = 11.422 ;
d 2
is negative, it gives the mximumvalue of .
dq 2
maximum profit = 300(11.42) + 4(11.42) 2 (11.42)3 1000

= Sh.1458.31
iii.
p x 13 + 2 x + y = 0 (Product x demand function)
p y 13 + x + 2 y = 0 (Product y demand function)
p x = 13 2 x y and p y = 13 x 2 y
total revenue for x = (13 2 x y ) x = 13 x 2 x 2 xy
total revenue for y = (13 x 2 y ) y = 13 y xy 2 y 2
total revenue for x and y combined = (13 x 2 x 2 xy ) + 13 y xy 2 y 2
= 13 x + 13 y 2 x 2 2 y 2 2 xy
total cost = x + y
Profit = Total revenue - total cost
= (13 x + 13 y 2 x 2 2 y 2 2 xy ) ( x + y )
= 12 x + 12 y 2 x 2 2 y 2 2 xy
px 13 + 2 x + y = 0 (Product
Using partial differentiation and standard calculus techniques for maxima and minima
= 12 4 x 2 y,
x
2
= 4 (-ve),
x 2
= 12 2 x 4 y
y
2
= 4 (-ve)
y 2
2
= 2
xy
314
Lesson Nine
Rules :
For maximum or minimum values
=0,
=0
x
y
2 2 2
For maximum value 2 2
x y xy
In this case, 12 2 x 4 y = 0 ; 12 4 x 2 y = 0
which gives; x = 2 units , y = 2 units
other conditions are a lso certified.
Price P x = 13 4 2 = 7
Price P y = 13 2 4 = 7
QUESTION SEVEN
i.
The required matrix equation is

5 8 x 640
4 12 y 820
Note that 640, 000 has been written as 640 (in thousands)
5 8
multiplying both sides with an inverse matrix of
we get;
4 12
1 12 8 640
1 12 8 5 8 x
60 32 4 5 4 12 y 60 32 4 5 820
1 28 0 x 1 1120
28 0 28 y 28 1540
x 40
=
y 55
thus x = 40, y = 55
ii.
1. Marginal productivity
60
+ 10 (this is the rate of change)
x2
total productivity P is given by the function
60
P = 2 + 10 dx
x
60
=
+ 10 x + C where C is a constant
x
Revision Aid
315
For the values; x = 5 and P = 62

we have; 62 =
60
+ 10(5) + C
5
thus C = 24
60
+ 10 x + 24
x
when x = 10
Hence P =
P=
60
+ 10(10) + 24 = 118 furnaces
10
2. Marginal productivity is the increase in output of electric furnaces per week if the
capitalization is increased by Sh.1 million.
316
Lesson Nine
LESSON THREE
QUESTION ONE
a) Discrete data have distinct values with no intermediate points, whereas continuous data can have
any values over a range either a whole number or any fraction.
b) Dispersion is the variation or scatter of a set of values.
Standard deviation is represented by;
s=
( x x)
n 1
( x x)
Where s: is for a sample and : is for a population

c) This is the coefficient of dispersion of a distribution that is used in comparing dispersion between
distributions. It is given by;
coefficient of variation =
100%
.
d) See Text
QUESTION TWO
a)
Payment in days
Mid-point
x
7
12
17
22
27
32
37
42
59
10 14
5 19
20 24
25 29
30 -34
35 39
40 44
Arithmetic Mean =
Number of customers
f
4
10
17
20
22
16
8
3
100
fx = 2, 405 = 24.05days
f 100
f
28
120
289
440
594
512
296
123
2,405
fx2
196
1,440
4,913
9,680
16,038
16,384
10,952
5,292
64,895
Revision Aid
317
b)
fx
f
Standard deviation =
64,895 2, 405
100
100
fx
= 8.4
c) Histogram to show payment record of 100 customers
Number of
Customers
25
22
20
20
17
15
16
10
10
8
4
3
4.5
9.5
14.5
19.5
24.5
29.5
34.5.
39.5
44.5
Days taken to settle debt
d) Out of 100, 16 i.e. in the class 30 to 34 days and 8 lie in the class 35 to 39 days. Therefore, the
best estimate that an unpaid invoice chosen at random will be between 30 and 39 days old is
2,478
64
0.24
QUESTION THREE
a) The smallest value in the distribution is 105, the largest in the distribution is 142. The range to
be spanned is 142 105, i.e. 37. The following grouping is a suggestion.
The classes should be of equal width.
Group
105 but less than
110
110
115
115
120
120
125
125
130
130
135
135
140
140
145
Tally
II
IIII
IIII
IIII III
IIII IIII
IIII
IIII
II
Frequency
2
5
4
8
10
5
4
2
40
318
Lesson Nine
b) Calculate the cumulative frequency.
Group
105 but less than
110
115
120
125
130
135
140
110
115
120
125
130
135
140
145
Summary Median value

First quartile
Third quartile
=
=
=
Frequency Cumulative
Frequency
2
2
5
7
4
11
8
19
10
29
5
34
4
38
2
40
125 p
119 p
131 p
The semi-interquartile range is given by the formula

value)
(third quartile value first quartile
Thus the semi-interquartile range
(131 119)
12
6p
=
=
=
c)
Group
105 but less than
110
115
120
125
130
135
140
110
115
120
125
130
135
140
145
Mid-point
107.5
112.5
117.5
122.5
127.5
132.5
137.5
142.5
f
2
5
4
8
10
5
4
2
40
Calculate the mean value first using the

Formula
x =
Thus
x = 5,000 40
fx
f
x = 125p
Continuing now with x = 125 we can calculate the standard deviation
fx
215.0
562.5
470.0
980.0
1,275.0
662.5
550.0
285.0
5,000
Revision Aid
319
Group
105 but less than
110
115
120
125
130
135
140
110
115
120
125
130
135
140
145
(x - x )
f (x- x )2
107.5
112.5
117.5
122.5
127.5
132.5
137.5
1425
-17.5
-12
.-7.55
-2.5
2.5
7.5
12.5
17.5
2
5
4
8
10
5
4
2
40
612.50
781.25
225.00
50.000
62.50
281.25
625.00
612.50
3,250.00
The standard deviation is given by the formula
=
2
f xx
f
3250
40
= 81.25
d)
i)
This distribution is very nearly normal and so consequently the mean at 125 and the median at
just over 125 are close to one another.
ii)
The semi-interquartile range and the standard deviation both measure dispersion. The semiinterquartile range, in this case 6p, gives the dispersion around the median. The standard
deviation measures the dispersion around the mean, in this case 9p, for the whole distribution
QUESTION FOUR
Use the same method as in question 77 to find mean and standard deviation or use the formulae
Mean
f xx
fx
f
2
Supermarket A
599
Mean
Standard Dev
30.49
30.5
12
49.92
Supermarket B
320
Lesson Nine
Mean
613
Standard Dev
12
51.1
34.25
51.08
Coefficient of Variation Supermarket A

=
=
30.5
49.92
100
61.1%
Coefficient of Variation Supermarket B

=
34.25 100
51.1
=
67.02
Hence variability of supermarket B is relatively greater than supermarket A
QUESTION FIVE
The objective of this question is to test the candidates knowledge of:
Use of base and current weighting index numbers and contrast their construction.
a) To establish the base weighted indices, the weights are the quantities used in 1981. the
following tabulation leads to the solution.
For 1981
Competent
Weight
A
B
C
D
Price ()
3
4
1
7
3.63
2.11
10.03
4.01
Price X weight ()
10.89
8.44
10.03
28.07
57.4
For 1982
Competent
Weight
A
B
C
D
Price ()
3
4
1
7
4.00
3.10
10.36
5.23
Price X weight ()
12.00
12.40
10.36
36.61
71.37
For 1983
Competent
A
B
C
D
Weight
Price ()
3
4
1
7
4.49
3.26
12.05
5.21
Price X weight ()
13.47
13.04
12.05
36.47
75.03
Revision Aid
321
The base weighted price indices are therefore for

1981,
100
71.37
1982,
57.43
75.03
1983
57.43
100
124.27
100
130.65
b) The current weighted indices use the weights of the components in 1982 to establish the
1982 index, then the weights of the components in 1983 for the 1983 index. The following
tabulation leads to the solution.
The 1981 index as before 100
For 1982
Component
Weight
1982
1981
Price ()
A
B
C
D
2
5
1
6
3.63
2.11
10.03
4.01
1981
Price
Weight ()
7.26
10.55
10.03
24.06
51.90
1982
Price ()
4.00
3.10
10.36
5.23
1982
Price
Weight ()
8.00
15.50
10.36
31.38
65.24
The 1982 current weighted price index is therefore

65.24
51.90
100 that is 125.70
Component
Weight
1981
Price ()
A
B
C
D
2
6
1
5
3.63
2.11
10.03
4.01
1981
Price
Weight ()
7.26
12.66
10.03
20.05
50.00
1982
Price ()
4.49
3.26
12.05
5.21
1982
Price
Weight ()
8.98
19.56
12.05
26.05
66.64
The 1983 current weighted price index is therefore

66.64
50.00
100 that is 133.28
c) Laspeyres price indices use weights at the base period, whereas Paasche price indices use
weights from the current period.
322
Lesson Nine
The weights at the base period will always be available whereas the current weights may not
always be available. In application like Retail or Consumer Price Index establishing the
current weights will be much more difficult and expensive than establishing the current
prices. These arguments give a preference for the Laspeyres type of index number.
It can be argued that the use of current weights reflects the present situation more
accurately, giving a preference for the Paasche type of index number. However, in
computing the series of index numbers, it is clearly demonstrated in part (a) that the
denominator need only be calculated once in the series for base weighted indices, whereas a
recomputation is needed for current weighted indices. This leads to a favouring of the
Laspeyres type index.
QUESTION SIX
a)
i)
The Laspeyres Price index is often summarized by the formula
PQ
PQ
n
100
That is the weighting factor in the calculation is the quantity at the base period.
For 1983
Component
Quantity
Price (1982)
Quantity 1982
Price1982
Price 1983
Quantity1982
Price1982
A
B
C
D
E
10
6
5
9
50
3.12
11.49
1.40
2.15
0.32
31.20
68.94
7.00
19.35
16.00
142.49
3.17
11.58
1.35
2.14
0.32
31.70
69.48
6.75
19.26
16.00
143.19
Thus the Laspeyres Price Index for 1983

=
143.19
100
142.49
100.49
For 1984
Component
A
B
C
D
Quantity
1982
10
6
5
9
Price
1984
3.2
11.67
1.31
2.63
Quantity 1982
Price 1984
32.00
70.20
6.55
23.67
Revision Aid
323
E
50
0.32
16.00
148.24
Thus the Laspeyres Price Index for 1984

148.24
142.49
100
104.04
The Paasche Price Index is often summarized by the formula
ii)
PQ
PQ
n
100
That is the weighting factor in the calculation is the quantity at the current year.
For 1983
Component
Quantity
Price (1982)
Quantity 1982
Price1982
Price 1983
Quantity1982
Price1982
A
B
C
D
E
12
7
8
9
53
3.12
11.49
1.40
2.15
0.32
37.44
80.43
11.20
19.35
16.96
165.38
3.17
11.58
1.35
2.14
0.32
38.04
81.06
10.80
19.26
16.96
166.12
Thus the Paasche Price Index for 1983

=
166.12
100
165.38
100.45
For 1984
Component
Quantity
Price (1982)
Quantity 1982
Price1982
Price 1983
Quantity1982
Price1982
A
B
C
D
E
14
5
9
10
57
3.12
11.49
1.40
2.15
0.32
43.68
57.45
12.60
21.50
18.24
153.47
3.20
11.67
1.31
2.63
0.32
44.80
58.35
11.79
26.30
18.24
159.48
Thus the Paasche Price Index for 1984
324
Lesson Nine
=
=
iii)
159.48
153.47
103.92
100
The following is a comparison between the index numbers.
Year
1983
1984
Laspeyres
100.49
104.03
Paasche
100.45
103.92
There is little to choose between the two measures in 1983 as the weightings, that is, the
quantity for 1982 and 1983 are close. The situation is different in 1984 where the weightings
have increased in three cases by a significant amount over the 1982 figures resulting in a
large increase in the Laspeyres index than the Paasche index.
b) An index of industrial production would probably by calculated on a month by month basis
by central government, indicating in percentage terms by how much production in the
industrial sector has either grown or declined over the previous month or year. The full
index would be made up from components which apply to particular sectors and so it is
possible for an employer to measure increase or decrease in production in that sector and
increase in production may thus be rewarded and a decrease would be looked upon with less
favour.
An index of retail prices would also probably be calculated on a month by month basis by
central government, and is an indication in percentage terms of the increase or decrease in
retail prices. This measure is often used to quantify inflation. A trade union may therefore
argue its case for an increase in pay to be greater than or equal to the rate of inflation to
keep up with the cost of living and not decrease the living standard s of its members. An
index of wages may also be published by central government, and may be available for
different industrial sectors.
QUESTION SEVEN
Note: The production for 1980 and 1990 is given in 1,000 boxes. As long as units are kept the
same throughout the problem, the rates will not change.
Produce
Cabbages
Tomatoes
Onions
Spinach
Production
1,000 boxes
1980 Qo
48,600
22,000
47,040
43,110
Price per
box
Po(Shs)Pn
1980
1990
1990 Qn
62,000
37,440
61,430
55,720
a) Mean Relative Index of prices
100
220
180
130
150
310
200
170
5.3279
4
Pn
Po
P o Qo
1.5
1.4091
1.1111
1.3077
5.3279
4,860,000
4,840,000
8467,200
5,604,300
23,71,500
133.20
pn qo
7,290,000
6,820,000
9,408,000
7,328,700
30,846,700
Revision Aid
325
Hence there is a 33.2% increase in average price of the four horticultural products from
1980 to 1990.
b) Laspeyres Price Index
30,846,700
23,771,500
100
129.76
% increase 29.76%
c) Paasches Price Index
pnqn 100
pOqn
And
d) Marshall Hedge-worth index =
p n qn
9,300,000
11,606,400
12,286,000
9,472,400
42,664,800
Cabbages
Tomatoes
Onions
Spinach
e) Paasches index
% increase
f)
Pn(qn + qn) 100

PO (qn + q n )
p o qn
6,200
8,236,800
11,057,400
7,43,600
32,737,800
42,664,800
32,737,800
100
pn + qn
110,600
59,400
108,470
98,830
pn (qn + pn)
16,590,000
18,426,400
21,694,000
16,801,100
73,511,500
130.32
30.32%
Marshall Hedge-worth Index =

% increase
g) Fishers Price Index =
73,511,500
56,509,300
100
130.09
130.4
30%
(129.76 130.32)1/2
This gives a percentage increase
= 30%
pn(pn + qn)
11,060,000
13,076,800
19,524,600
12,847,900
56,509,300
326
Lesson Nine
LESSON FOUR
Question 1 and 2
See text
QUESTION THREE
r2x1 = 0.78
This is the coefficient of determination of miles traveled to cost and means that 78% of total
cost is attributable to mileage
r2x1 = 0.16 i.e. 16% of cost is accounted for by the type of journey
R2 = 0.88 is the overall coefficient of determination and indicates that the multiple regression
equation accounts for 88% of the total variation in costs.
The coefficient in the equation are:
a = 86
b1 = 0.37
b2 = 0.08
= fixed costs
= amount per mile
= influence of the type of journey
QUESTION FOUR
Forecasts produced by
3 monthly
Moving average
450
437
417
397
383
377
380
407
443
470
6 monthly moving average
12 monthly moving average
423
410
397
388
395
410
425
424
Revision Aid
327
QUESTION FIVE
a)
Year
Qtr
1990/91
b)
1
2
3
4
1
2
3
4
1
2
3
4
1
2
3
49
37
58
67
50
38
59
68
51
40
60
70
50
42
61
Sum of
four qtrs
Sum of two
qtrs
Trend
Actual
minus
trend
Actual
/
trend
211
212
213
214
215
216
218
219
221
220
222
223
423
425
427
429
431
434
437
440
441
442
445
52.875
53.125
53.375
53.625
53.875
54.25
54.625
55
55.125
55.25
55.625
5.125
13.875
-3.375
-15.625
5.125
13.75
-3.625
-15
4.875
14.75
-5.625
1.097
1.261
0.937
0.709
1.095
1.253
-.934
0.727
1.088
1.267
0.899
Either
Year
1990/91
1991/1992
1992/93
1993/94
Total
Average
Seasonal variation
Q1
-3.375
-3.625
-5.625
-12.625
-4.208
-4
Additive model
Quarter
Q2
Q3
5.125
-15.625
5.125
-15
4.875
______
_____
-30.625
15.125
-15.313
5.042
-15
5
Q4
13.875
13.75
14.75
______
42.375
14.125
14
= -0.354
=0
Or
Multiplicative model
Quarter
Q2
Q3
1.097
0.709
1.095
0.727
1.088
Year
Q1
1990/91
1991/1992
0.937
1992/93
0.934
1993/94
0.899
Total
2.770
0.718
Average
0.923
0.718
Adjustment factor*
1.0015
1.0015
Seasonal variation
0.924
0.719
* adjustment factor = 4/(3.994) = 1.0015
3.280
1.093
1.0015
1.095
c) An explanation of the forecasting method:

Plot a graph of the trend.
i)
Q4
1.261
1.253
1.267
3.781
1.260
1.0015
1.262
= 3.994
= 4.000
328
Lesson Nine
ii)
By eye, establish an appropriate forecast of the trend for the last qurter of 1993/94 and
the first three quarters of 1994/5. (Note: linear regression or the high/low method may
be appropriate methods to establish the forecast.)
Adjust the forecast trend for these quarters for the seasonal variations:
iii)
Additive model
Estimated data value = forecast trend value + appropriate seasonal variation value.
Multiplicative method
Multiply each point by the appropriate seasonal factor.
QUESTION SIX
a)
A = y bx
b=
)(
x=x yy
2
xx
)
A
x=
2
8
6
8
10
4
4
2
6
10
60
6
6
6
6
6
6
6
6
6
6
x- x
-4
2
0
2
4
-2
-2
-4
0
4
60
60
132
100
120
150
84
90
68
104
140
1,048
104.8
104.8
104.8
104.8
104.8
104.8
104.8
104.8
104.8
104.8
y=
10
Therefore b =
B
y
804
80
1,048
10
y- y
-44.8
27.2
-4.8
15.2
45.2
-20.8
-14.8
-36.8
-0.8
35.2
= 104.8
= 10.05
a = 104.8 (10.05 6) = 44.5

where y = total cost (x 10)
x = age in years
b) Maintenance costs using formula from (a):
Age of vehicles
(years)
1
2
Cost
( = 10)
54.55
64.60
AB
179.2
54.4
0
30.4
180.8
41.6
29.6
147.2
0
140.8
804.0
(x - x )2
16
4
0
4
16
4
4
0
16
16
80
Revision Aid
329
3
4
5
6
7
8
9
10
74.65
84.70
94.75
104.80
114.85
124.90
134.95
145.00
c) 12-year-old vehicle would have an estimated maintenance cost of:

44.5 + (10.05 12) = 165.1 ( x 10), or x 10), or 1,651.00
This forecast is an extrapolation beyond the data and consequently is less sound.
QUESTION SEVEN
a) Two possible reasons for the large variation in output each month are:
Seasonal variation
Production problems in some months
b) Graph showing relationship between output and costs.
The graph shows there is a strong positive relationship between output and costs. This means
that output may be used to predict costs.
330
Lesson Nine
x
16
20
23
25
25
19
16
12
19
25
28
12
240
b=
a=
y
170
240
260
200
280
230
200
160
240
290
350
200
2,920
xy
2,720
4,800
5,980
7,500
7,000
4,370
3,200
1,920
4,560
7,250
9,800
2,400
61,500
12 61,500 - 240 2,920

=
12 5.110 - 240 2
2,920
12
10
x2
256
400
529
625
625
361
256
144
361
625
784
144
5,110
37,200
3,720
y2
28,900
57,600
67,600
90,000
78,400
52,900
40,000
25,600
57,600
84,100
122,500
40,000
745,200
=
10
240
= 43.333
12
y = 43.333 = 10x
This means that when output is zero, costs are zero 43.333, and that for every one unit increase
in output, costs will rise by 10. This assumes linearity.
QUESTION EIGHT
i.
The tabulation of the trend pattern is as follows and has been computed using the formula
Trend = Sales Seasonal deviation
Year
Quarter
Sales
1983
2
3
4
1
2
3
4
1
2
3
4
1
2
3
4
1
000
360
530
354
304
430
750
395
340
500
660
509
374
590
710
521
440
1984
1985
1986
1987
Seasonal
deviation
000
Trend
-42
-128
-37
276
-93
-145
12
153
-15
-165
43
153
396
432
467
474
488
485
488
507
524
539
547
557
000
Revision Aid
ii.
331
The seasonal variations are established from the following table

Year
1983
1984
1985
1986
Total
Average
Adjusted
average
Quarter
3
-128
-145
-165
-438
-146
-147
-37
12
43
18
6
5
276
153
153
582
194
193
4
-42
-93
-15
-150
-50
-51
Total 4
Total 0
Adjustment to each average 4/4 that is 1

To the nearest integer the seasonal variations are
Quarter
iii.
-147
193
-51
Forecast sales 1987 quarter II =

590 + 5 ( 000)
= 595 ( 000)
QUESTION NINE
The object of this question was to test the candidates knowledge of the time series and the
ability to present data on a labeled diagram
Year
Quarter
Costs
1980
1981
IV
I
II
III
IV
I
II
III
IV
I
II
III
IV
I
II
III
1560
1730
1554
1504
1630
1950
1595
1540
1700
1860
1709
1574
1790
1910
1721
1640
1982
1983
1984
Four quarter
total
Four quarter total in

pairs
Trend
Deviation
6348
6418
6638
6679
6715
6785
6695
6809
6843
6933
6983
6995
7061
12766
13056
13317
13394
13500
13480
13504
13652
13776
13916
13978
14056
1595.77
1632.00
1664.63
1674.25
1687.50
1685.00
1688.00
1706.50
1722.00
1739.50
1747.25
1757.00
-41.75
-128.00
-34.63
275.75
-92.50
-145.00
12.00
153.50
-13.00
-165.50
42.75
153.00
332
Lesson Nine
a) The calculations for the trend figures and the deviations are summarized in the above table
The seasonal effect is removed from the data by first totaling four quarter figures, then
totaling the four quarter figures in pairs and finally dividing by eight and centering the trend
figure at the middle point
Deviation is the difference between the costs and the corresponding trend figure
b) The seasonal deviation is another calculation produced from a table
Year
1983
1984
1985
1986
Total
Average
Adjusted
average
Seasonal
deviation
I
275.75
153.50
153.00
582.25
194.08
192.70
193.00
Deviations
II
-41.75
-92.50
-13.00
Seasonal
Quarter
III
-128.00
-145.00
-165.50
-147.25
-49.08
-50.46
-438.50
-146.17
-147.55
20.12
6.71
5.33
-50.00
-148.00
IV
-34.63
12.00
42.75
Total = 5.54
Total = 0.02
The adjustment is obtained by reducing each figure by 1.38 (that is 5.54 4)

c) From the trend the forecast trend value of Quarter IV of 1984 is 1790. the seasonal
deviation is 5, hence the forecast heating costs for quarter IV for 1984 are 1795
It can be seen that the trend is a little variable and the seasonal deviations are not very
regular, possibly due to the weather.
Another possible reason for the irregular fluctuation in the heating cost trend line is an
uneven increase in the price of the heating medium (for example, oil or gas or
electricity.)
Revision Aid
333
LESSON FIVE
QUESTION ONE
a) The Poisson distribution is appropriate because there is a small probability of an event
occurring these are discrete values and the average of these events (i.e. m = np) is below
10 (50 0.1 = 5)
b) The Poisson formula is
x e
P( x) =
!
In this exmple m = np = 50(0.02) =1
P ( x 2) = P(x=0) +P ( x=1 ) + p (x=2 )
=1
e e 2e
=
+
+
0!
1!
2!
27181 27181 12 27181

+
+
1
1
2 1
= 0.3679 + 0.3679 + 0.1840 = 0.92

c) Similarly , the following can be drawn up
P
0
0.02
0.05
0.10
0.15
Pa
1.000
0.920
0.544
0.125
0.020
QUESTION TWO
(a)
(i)
Probability = 0.7 0.3 0.7 = 0.147
(ii)
Probability =3C2 (0.7)2 (0.3) = 0.441;
(iii)
The answer to (ii) is 3 the answer to (i)
This is because either A or B could be detected, A and C, or B and C. Each of these situations is
equally probable, with the probability given in part (i).
(b) (i) The probability that the defect is undetected by the inspection procedure is
0.3.
334
Lesson Nine
The probability that it is undetected by the secondary check is 0.4.
The probability that it is undetected by both is 0.3 0.4 = 0.12
(ii)
Call the two faults A and B.

We can have the possibilities:
A is found by inspection procedure, B not found by either procedure.

B is found by inspection procedure, A not found by either procedure.
A is found by secondary check, A and B not found by inspection procedure.
B is found by secondary check, A and B not found by inspection procedure.
Adding together the probabilities corresponding these four possibilities, the required probability
is
(0.70.30.4) + (0.30.70.4) + (0.30.30.60.4) + (0.30.30.40.6)
=0.0840 + 0.0840 + 0.0216 + 0.0216 + 0.1680 + 0.0432
=0.2112
(iii)
Probability that a fault is detected by the inspection procedure is 0.7.

Probability that a fault is detected by secondary check is 0.3 0.6 =0.18
Therefore the proportion of faults detected by the inspection procedure and secondary check,
respectively, is
0.7
0.18
and
0.7 + 0.18
0.7 + 0.18
that is
70
18
and
88
88
QUESTION THREE
Let F represent a unit which has been found to be faulty.
Let P(S1) = probability that a unit chosen at random comes from S1
P(S1) = 0.40
P(S2) = 0.35
P(S3) =
0.25
1.00
The percentages of faulty unit are as follows:
Revision Aid
335
P( F S1 ) = 0.02
P( F S2 ) = 0.03
P( F S3 ) = 0.04
The required probability may be expressed as;
P(S1 F)
the unknown probability is P(F) to be slotted into the formula
Note that
P( S1 F) =
P( S2 F) =
P( S3 F) =
P(S1 ) P( F S1 )
P(F)
P(S2 ) P( F S2 )
P(F)
P(S3 ) P( F S3 )
P(F)
the faulty part can only have come from S1 or S2 or S3 and so
P(S1 F)+P(S2 F)+P(S3 F)=1.0

since P(F) is a denominator and the sum equals unity then the expression
P(S1 ) P( F S1 )+P(S2 ) P( F S2 )+P(S3 ) P( F S3 )

must be equal to P(F)
P(F)= (0.4 0.02)+(0.35 0.03)+(0.25 0.04)
thus
= 0.0080 + 0.0105 + 0.0100

= 0.0285
Substitution into P(S1 F) =
P(S1 ) P( F S1 )
P(F)
P( S1 F) =
Gives
0.4 0.02
0.0285
= 0.2807
QUESTION FOUR
P ( 0 ) = 6C0 ( 0.3) ( 0.7 )

0
P (1) = 6C1 ( 0.3 ) ( 0.7 )

1
P ( 2 ) = 6C2 ( 0.3) ( 0.7 )

2
P ( 3) = 6C3 ( 0.3) ( 0.7 )

3
336
Lesson Nine
P ( 4 ) = 6C4 ( 0.3 ) ( 0.7 )

4
P ( 5 ) = 6C5 ( 0.3) ( 0.7 )

5
P ( 6 ) = 6C6 ( 0.3) ( 0.7 )

6
QUESTION FIVE
P = 0.6 q = (1-p) = 1-.06 = 0.4
=
pq
n
0.6 0.4
200
z=
= 0.035
0.65 0.60
= 1.43
0.035
Which gives 0.4236 (42.36%)
This means there is a (0.5-0.4236) 0.0764(7.64%) chance of 65% or more passing the first
attempt. This is graphically shown below.
Probability
density
42.36%
7.64%
0.60
0.65
X
proportions
i.e. the mean
QUESTION SIX
a) The left-hand tail of the distribution below 900 hours represents the number of lamps that
will fail before 900 hours. Accordingly, if the probability of the distribution above 900 hours
is found and deducted from 0.5, the required number can be found thus:
z=
1000 900
= 1.33 The probability of which is 0.4082
75
Revision Aid
337
Lamps failing before 900 hours =5000(0.5- 0.4082)=459

b)
z=
1000 950
= 0.67
75
The probability of which is 0.2486
Lamps failing between 1000 and 950 hours = 5000 (0.2486)=1243

c)
z=
1000 925
=1
75
The probability of which is 0.3413
Proportion failing between before 925 hours = 0.5 0.3413 = 15.87%

d) A probability of 0.3(0.5-02) is found in the tables with a Z score of 0.84
Thus 0.84=
z=
1000-916
84
=1 s.d.=
=100
s.d.
0.84
338
Lesson Nine
LESSON SIX
QUESTION ONE
The sample mean is 150 grams so that the estimate of the population mean is 150 grams.
i.e. x = 150 grams = = 150 grams
Where x means W best estimate of X.
When n= 625
Standard error of the mean
s
30
=
=1.2grams
n 625
When n =1225
30
sx =
=0.857 grams
1225
QUESTION TWO
Correction factor =
N-n
800 80
=
= 0.9493
N-1
800 1
Approximation to correction factor= 1-
n
80
= 1
= 0.9486
N
800
It will be seen that to three significant figures, it is accurate enough for all practical purposes, the
two formulae give the same result, i.e. 0.949.
Standard deviation error of the means
n
=
n
N
6
0.949
80
= 0.637 grams
Note: The standard error without correction is
6
= 0.671 grams
80
Thus the precision of the sample estimate, measured by the standard error, is determined not
only by the absolute size of the sample but also to some extent by the proportion of the
population sampled.
Revision Aid
339
QUESTION THREE
The Central Limit Theorem states that the means of samples (and the medians and standard
deviations) tend to be normally distributed almost regardless of the shape of the original
population.
QUESTION FOUR
See text
QUESTION FIVE
H0 :2 =1
1 (one-tail test)
H1 : 2 >
P1 =
200
= 0.2
1000
p2 =
240
=0.22
1091
Pooled sample proportion
p=
200+240
= 0.21
1000+1091
and q = 1- p = 1- 0.21 = 0.79
S P1 P2 =
0.21 0.79
1000
0.21 0.79
1091
= 0.0178
z=
0.20-0.22
=1.12
0.0178
The critical value of one-tailed test t the 5% level is 1.64 so that as the calculated value is
lower than this value we conclude there is insufficient evidence to reject the null
hypothesis.
H 0 :1 - 2 =0
QUESTION SIX
H1:1 - 2 =0
It will be seen that this is a two-tailed test.

The common standard deviation is calculated first.
(n1 -1)s12 + (n 2 1)s 22

sp=
n1 +n 2 2
(12 1)632 + (9 1)762
12 + 9 2
= 68.77
=
340
Lesson Nine
s x1 =
s x2 =
SP
n1
68.77
= 19.85
12
sp
68.77
=
= 22.92
n2
9
s(x1 -x 2 )= s 2x1 +s 2x2 =30.32

and, finally the t score can be calculated
t=
x1 -x 2
1060 970
=
= 2.97
s(x1 -x 2 )
30.32
At 5% level with (n1+n2-2)= (12+9-2) = 19 degrees of freedom value is 2.093.

Since the calculated value is greater than 1.96, the null hypothesis can be rejected at the 5% level,
i.e. we conclude that there is a significant difference between the mean monthly incomes.
QUESTION SEVEN
We are testing whether the observed number of defects fits a binomial distribution, thus;
H0: the observed number of defects conforms to a binomial distribution of the form
(p+q) 5 where p=0.18
H1 :that the observations do not conform.
The observed frequencies are already given so its only necessary to calculate the frequencies
expected from a binomial distribution to the power 5 i.e. (p+q)5, where p=0.18 and q=1-0.18=
0.82.
The probabilities of the various values of p and q can be found from binomial probability tables
if available. Alternatively they can be calculated from the binomial expansion. I.e.
P5 +5 (p4 q)+ 10(p3 q2)+ 10(p2 q3)+ 5(p q4)+q5
This shows the probabilities for 5, 4, 3, 2, 1 and 0 defectives and when p = 0.18 (the probability
of a bulb being defective) and q=0.82 (the probability of not being defective) the probabilities
range from 0.0002 (i.e. 0.185) for five defectives to 0.3711 (0.825) for no defectives. When the
probabilities are known they are multiplied by 100 boxes to find the expected frequencies which
are used in the normal X2 procedures.
The table below summarises the calculations:
Revision Aid
Defectives
0
1
2
3
4
5
341
No. of boxes
Observed
40
37
17
5
1
0
Binomial
Probabilities
0.3711
0.4069
0.1786
0.0392
0.0040
0.0002
Expected
Frequency
37.11
40.69
17.86
3.92
0.4
0.02
(O - E) 2
8.35
13.62
0.74
2.76
(O - E) 2
E
0.22
0.33
0.04
0.64
1.23
Note: Because of the very small values of the expected frequencies for 3, 4 and 5 defectives they
have been combined into one but it makes little difference to the results if they are not
combined.
The calculated 2 value is compared with the 2 value for the appropriate degrees of freedom.
Because the last three classes have been combined there are four classes remaining, i.e. for 1, 2
and the combined class for 3-5 rejects thus
V =n 2 = 4 - 2=2.
The X2 value for two degrees of freedom at the 5% level is 5.991 and, as the calculated value of
1.23 is well below this, we accept the null hypothesis and conclude that the observed values fit a
binomial distribution to the power 5 when p = 0.18.
Note: If the last three classes had not been combined, the calculated X2 score would have been
1.8 and there would have been 4 degrees of freedom. At 4 degrees of freedom the score is 9.488,
so the conclusion would be the same, i.e. we accept the Null Hypothesis.
QUESTION EIGHT
See text
QUESTION NINE
1.
a)
S.e. =
pq
n
0.1 0.9
80
= 0.033
Where p =proportion late

Actual p =
6
80
= 0.075
0.1 0.075
= 0.75 standard errors
0.033
This is less than the standard errors at the 5% level 1.96 so we conclude there is no
significant improvement in deliveries.
342
Lesson Nine
b)0.75 standard errors from the mean would cover 0.2734 i.e. 54.68%, say 55% of the
population, so the MDs claim could be accepted at any level of confidence
2.
Calculated mean of sample = 4.68 gms

Calculated sample s.d. = 1.968
best estimate of population s.d.=
n
10
= 1.986
= 2.093
n-1
9
std. error of mean=
2.093
= 0.66
10
There are 10-1 =9 d.f and it is a one-tailed test. The 5%value for a one-tailed test is
1.833.
The sample mean should be within 1.833*0.66 gms = 1.21 gms.
The actual difference is 4.68-3.8 gms= 0.88 gms so the figures do not support the
Purchasing Managers assertion.
Revision Aid
343
LESSON SEVEN
QUESTION ONE
a)
NPV
45
Sell immediately
120
Succeed 0.55
48
Fail 0.45
-40
53.5
Drill immediately
70
1
Tests
53.5
65
Sell
Succeed 0.7
Drill
70
Succeed 0.8 100
Fail 0.2
Fail 0.3
15
Sell
-50
15
3
100
Drill
-20 Succeed 0.2

4
Fail 0.8
-50
b).
Drilling Immediately:
( 0.55 120+0.45 ( -40 ) ) m = ( 66-18 ) m = 48m
Tests:
Successful and drill, 0.8 100+0.20 ( -50 ) m = ( 80-10 ) m = 48m
Fail and drill, 20 100+0.8 ( -50 ) m = ( 20-40 ) m = 20m

There fore EV = ( 0.7 70+0.3 15 ) m = ( 49+4.5 ) m = 53.5m
Overall EV = 53.5
Thus course of action is: first carry out tests, if successful proceed to drill if tests fail, sell the
exploitation rights.
344
Lesson Nine
QUESTION TWO
a) tree diagram
Profits before construction
costs (000)
On time (p = 0.3)
Continue as present
Employ consultants
(Cost = 200,000)
1,000
1 month delay (p = 0.23)
1,000 100 = 900
2 months delay (p = 0.23)
1,000 200 = 800
1,000 300 = 700
On time
1,000 200 = 800
1 month delay (p = 0.05)
1,000 200 - 100 = 700
1,000 200 - 200 = 600
b) EV calculations
EV(present arrangement) = (0.31000) + (0.23900) + (0.23800) + (0.23700)
= 860
EV (with consultants) = 785
c) on strict EV calculations the firm should not use consultants, however management may
consider to use consultants in order to improve the chances of completing on time thus
safeguarding their reputation.
QUESTION THREE
See text
QUESTION FOUR
Let the value of the small store be = 1
And let the value of the large store be = 2
If both survive, A loses nothing, if only larger store survives, A loses 1 and if smaller store
survives, A will lose 2.
Revision Aid
345
GAIN MATRIX FOR A

Defend the smaller store
I
Defend the larger store
II
DEFENDER
Attack the smaller store

I
Both survive
0
The smaller store
destroyed
-1
B
I
II
Row
Minimum
-2
-2
-1
-1
II
Column
Maximum
There is no saddle point.
Hence this is a problem of mixed strategy.

Using method given
I
I
0
II
-2
0-(-2)=2
II
-1
0-(-1)=1
0-(-1)=1
0-(-2)=2
The final strategy is given by the matrix
I
0
II
-2
1/3
II
-1
2/3
2/3
1/3
A plays his first row 1/3 rd of the time (randomly)

A plays his second row 2/3 rd of the time.
Similarly
B plays his first column 2/3 rd of the time
B plays his second column 1/3 rd of the time.
Attack the larger store

II
The larger store destroyed
-2
Both survive
0
346
Lesson Nine
The values of the game
=01/32/3 +(-2)(1/31/3) +(-1)2/32/3 +01/32/3
=0-2/9-4/9+0
= -6/9
= -2/3
QUESTION FIVE
Y
1
-2
2
2
X
Column
Maximum
2
1
2
2
-1
-2
1
1
Row
Minimum
-1
-2
0
There is no saddle effect.

Let the probabilities of y be p, q, r.
Then the three payoffs to y corresponding to each of the three counter moves of his opponent x
must all be equal to the optimal value v of the game.
Ys payoffs against the three moves of x are
1p+2q+(-1r)
-2p+1q+1r
2p+0q+1r
We obtain three equations by equating each of the three payoffs to v
1p+2q-1r =
-2p+1q+1r =
2p+0q+1r =
Also p+q+r = 1 (Total probability)

1p+2q+1r=
1p+2q+1r=
p+q+r
=
-2p+1q+1r.. (1)
2p+0q+1r(2)
1
..(3)
Solving these equations simultaneously, we get

P=2/17,
q=8/17,
and
r=7/17
Similarly using the same reasoning as before

Let the three probabilities of x be p, q, r
We get 1p+2q+1r
Also p+q+r =
= 2p+1q+0r = -1p+1q+1r
1
Revision Aid
347
Solving them simultaneously, we get

P=3/17,
q=5/17,
Hence x should play his rows in the ratio

And
y should play his columns in the ratio
and
r= 9/17
3: 5: 9
2: 8: 7
(Note: Rows and columns should be played at random)

[(payoff * joint probability of the payoff]
2 3
8 3
7 3
i.e. 1 + 2 + (1)
17 17
17 17
17 17
+6 other values calculated in the same way as above
which amount to 11/17
Alternatively
Value of the game is
1 3 + (-2) 5 + 2 9 11
=
3+5+9
17
348
Lesson Nine
LESSON EIGHT
QUESTION ONE
a)
Let x1 and x2 be the number of shares invested in airline and insurance shares respectively.
Then the objective function will be as follows:
Objective function (maximize)
z = 2x1 + 3x2
Share appreciation.
Subject to the following restrictions:

1. 40x1 + 50x2 100,000 sh. Total investment.
sh. Investment in insurance.
2. 50x1 40,000
sh. Dividends.
3. x1 + 1.5x2 2600
Non-negativity of number.
4. x1, x2 0
b)
Reduced cost represents the amount that objective function coefficient of a non-basic
decision variable must improve (increase in this case) to be put into the basic. In this case
since reduced cost is equal to zero for both variables, it means they are in the basic.
Dual prices on the other hand mean the amount that share appreciation will improve in
case any of the limiting constraints is increased by one unit. This occurs for constraints that
the slack is exactly zero. In this case total investment and investment in insurance do have
positive dual prices while Dividends does have zero dual price (can not increase share
appreciation if increased by one unit).
c)
From the computer solution, the clients money should be invested as follows, to satisfy the
restrictions:
1500 shares to be invested in Airline shares, and
800 shares to be invested in insurance shares, to give an optimum quarterly share
appreciation of sh. 5400.
d)
For the optimal decision to remain the airline shares appreciation should not be lower than
2.5, but can be any higher amount. That is, the optimum solution is insensitive to increase
in the share appreciation.
QUESTION TWO
a)
Formulation of the problem

Take x1 to be the number of large loaves and x2 to be the number of small loaves.
Objective function
z = 5x1 + 3x2
Constraints.
Line
1
Line
2
Line
3
Line
4
Profit
x1 280
x2 400
10x1 + 8x2 4000
25x1 + 12.5x2 8000
5
x1, x2 0
Maximum number of large loaves.

Maximum number of small loaves.
Space (1000-3) m3
Hours (1000-3)
Non-negativity.
Revision Aid
349
Using graphical method.
Small Vs Large loaves

700
600
500
x2
400A
300
C
200
100
D
F
0
-100
100
200
300
400
x1
3
The feasible area is that enclosed by corner points ABCDEF

Corner points are where the optimum feasible solution exists. The upper points are to be
considered for maximization problem.
At corner point A, Line 2 and x1=0 intersect
So x2= 400
Putting these value in the objective function gives the following
z = 5x1 + 3x2=50 + 3400=1200
ProfitA
At corner point B, Line 2 and Line 3 intersect
x2= 400 and x1=
4000 8 400
= 80
10
Putting these value in the objective function gives the following

ProfitB
z = 5x1 + 3x2=580 + 3400=1600
At corner point C, Line 3 and Line 4 intersect
The equations for the lines are:
(1)
Line 3 10x1 + 8x2 = 4000
(2)
Line 4 25x1 + 12.5x2 = 8000
Multiplying equation (1) by 25 and equation (2) by 10 gives the following
(10x1 + 8x2 = 4000) 25
(25x1 + 12.5x2 = 8000) 10
250x1 + 200x2 = 100,000
(3)
500
350
Lesson Nine
250x1 + 125x2 = 80,000
(4)
Deducting equation (4) from Equation (3) gives:
75x2 =20,000
x2 =266.7
x1 =
4000 - 8 266.7
= 186.7
10
Putting these values in the objective function gives the following

ProfitC
z = 5x1 + 3x2=5 186.7 + 3 266.7=1733
At point D, Line 1 and Line 4 intersect.
x1 = 280
x2 =
8000 + 280 25
= 80
12.5
The profit at this point is then equal to:

z = 5x1 + 3x2=5280 + 380=1640
At point E, Line 1 and x2 = 0 intersect
x1 = 280
x2 = 0
z = 5x1 + 3x2=5280 + 30=1400
ProfitD
ProfitE
Comparing these profits, it is at point C that profit is maximized.

So the solution is that:
x1 = 186
No. of large loaves produced.
No. of small loaves produced.
x2 = 266
And the maximum profit is Profit c = Shs. 1,733 where x1 = 186.7, x2 = 266.7
NOTE: Two methods can be used to solve the problem. It is easily solved using the
graphical rather than the simplex method, since it is just two variables and sensitivity
analysis is not required.
b)
To solve this kind of problem (linear programming problem) the following procedure is
followed:
First, the problem has to be formulated. That is, the objective function and constraints
are determined.
Objective function is that which is to be optimised.
Constraints are the limitations in resources.
-
Secondly, the method of solving is determined. In this case, of a two-variable problem,

the better method to use is graphical method, rather than simplex method.
Thirdly, the constraints are taken as equalities and a line graph drawn. The unwanted
regions are shaded out. Resulting region indicates the feasible region. The optimum
point exists where there are corner points, which show extreme amounts. For
maximization it is the outer ones to the right and up. For minimization it is the lower
side.
Lastly, the profit is determined at those points where there is maximum profit, is the
point to be used.
NOTE: This part simply asks for the procedure followed.
Revision Aid
351
QUESTION THREE
a)
i)
Simplex method will be appropriate.

Formulation of problem.
Objective function.
Let x1 and x2 be the number of Deluxe and Professional bicycle frames produced
respectively per week.
z = 1000x1 + 1500x2
Constraints:
2x1 + 4x2 100
3x1 + 2x2 80
x1, x2 0
Profit sh.
Aluminum alloy
Steel alloy
In standard form:
0 = z 1000x1 1500x2 + 0s1 + 0s2
100 = 2x1 + 4x2 + s1 + 0s2
80 = 3x1 + 2x2 + 0s1 + s2
Table 1
s1
s2
z
Table 2
x2
s2
z
x1
2
3
-1000
x2
4
2
-1500
s1
1
0
0
s2
0
1
0
Solution
100
80
0
Ratio
25
40
1/4
2
-250
1
0
0
1/4
-1/2
375
0
1
0
25
30
37,500
50
15
Table 3
0
1
3/8
-1/4
x2
x1
1
0
-1/4
1/2
z
0
0
312.5
125
Stop here
The optimal weekly production schedule is as follows:
Deluxe bicycle Frame = 17.5 17
Professional bicycle Frame = 15
ii)
17.5
15
41,250
Let 1 be the change in profit from Deluxe bicycle frame.

2 be the change in profit from Professional bicycle frame. So
limit of profit.
C1 = 1000 + 1 and C2 = 1500 + 2
From the final table:
To avoid entry of
s1
312.5 1/41 > 0
1 < 1250
125 + 1/21 > 0
1 > -250
s2
From the two conditions:
-250 < 1 < 1250 and
352
Lesson Nine
750 < C1 < 2250
To avoid entry of
s1
312.5 + 3/82 > 0
s2
125 1/42 > 0
2 > -833.33
-2 > -500 2 < 500
So from the two conditions:

-833.33 < 2 < 500
And C2 varies as follows
666.7 < C2 < 2000
NOTE: This problem could be solved graphically with part (i) Easily determined. Part
(ii) Limits will be determined from equating slopes of the objective function which has
coefficients with constraints nearest to it.
For part (ii), accurate drawings will be required. Intuition will have to be followed and
there will be an assumption that fractions are possible.
b)
The technique is really involving.

Assumes fractions are possible, which is not really the case like here where we cannot make
a bicycle frame.
QUESTION FOUR
a)
A feasible solution is one that satisfies the objective function and given constraints
Transportation problem is a special linear programming problem where there a
number of sources and destinations and an optimum allocation plan is required. Total
demand equal total supply
iii) Assignment problem is a special kind of transportation problem where the number of
sources equals the number of destinations. That means for every demand there is one
supply.
This is a case of assignment problem.
Assignment problems usually require that the number of sources equal the number of
supply. Here there are 5 districts and only 4 salespersons. A dummy salesperson E is
introduced with zero ratings.
i)
ii)
b)
Sales persons
A
B
C
D
E
Districts
1
2
92
90
84
88
90
90
78
94
0
0
3
94
96
93
89
0
4
91
82
86
84
0
5
83
81
93
88
0
By following the Hungarian method:

Firstly:
For each row, the lowest rating is reduced from each rating in the particular row. This
results to a row reduced rating table. Then all the zeroes are to be crossed by the least
number of vertical and horizontal lines. If the number of lines equal the number of rows
(or columns = 5 in this case) then the final assignment has been determined. Otherwise the
following steps are followed.
Revision Aid
353
A
B
C
D
E
1
9
3
4
0
0
2
7
7
4
16
0
3
11
15
7
11
0
4
8
1
0
6
0
5
0
0
7
10
0
Secondly, for each column, the lowest rating is reduced from every rating in the particular
column. In this case the table will remain the same since the dummy salesperson has ratings
of zero for every district.
Thirdly a revision of the opportunity-rating table is done.
The smallest rating in the table not covered by the lines is taken (in this case it is one). This
is reduced from all the uncrossed ratings and added to the ratings at the intersection of the
crossings. Then all the zeroes are to be crossed by the least number of vertical and
horizontal lines. If the number of lines equal the number of rows (or columns = 5 in this
case) then the final assignment has been determined.
Otherwise the following steps are followed.
1
2
3
4
5
A
8
6
10
8
0
B
2
6
14
0
0
C
4
4
7
0
8
D
0
16
11
6
11
E
0
0
0
0
1
Third step is repeated as follows:
1
A
6
B
0
C
2
D
0
E
0
2
4
4
2
16
0
3
8
12
5
11
0
4
8
0
0
8
2
5
0
0
8
13
3
Still the optimal solution has not been reached. Third step is again repeated to give the
following table:
1
2
3
4
5
A
6
2
6
8
0
B
0
2
10
0
0
C
2
0
3
0
8
D
0
14
9
8
13
E
0
0
0
4
5
An optimal assignment can now be determined since the number of lines crossing the
ratings is equal to 5.
Lastly, the assignment procedure is that a row or column with only one zero is identified
and assigned. This row or column is now eliminated. The other zeroes are then assigned
until the last zero is assigned. This step-by-step assignment is shown on the following table
from the first one to the fifth one.
354
Lesson Nine
District
1
10
14
13
4
5
Sales person
0
1
D
E
0
0
The assignment is as follows

Salesperson
A
B
C
D
Total rating
District
5
4
2
1
Rating
83
82
90
78
333
The total rating is 333.
QUESTION FIVE
a)
Sensitivity analysis measures how sensitive a linear programming solution is to changes in

the values of parameters. These parameters include the coefficients of objective function,
limiting resources and non-limiting resources.
So sensitivity analysis involves changing any of these parameters and showing how the
linear programming problem is affected.
Dual values indicate the additional improvement of the solution due to additional unit of
limiting resource. In that way, the additional improvement of solution is the price worth
paying to release a constraint
b)
Let x1, x2 and x3 be the units of desktop 386, Desktop 286 and laptop 486
Maximize profit
Z = 5000x1 + 3400x2 + 3000x3
Subject to
limit of desktop models
x1+x2500
limit of laptop model
x3 250
x3 120
limit of 80386 chips
limit of 80286 chips
x2+x3400
5x1+4x2+3x22000 hours available
Assumptions
x1, x2, x3 0
Linearity/proportionality
Revision Aid
355
Divisibility
Deterministic
Additive
i)
The optimum product mix is that the numbers of units to produce are
Desktop 386-120
Desktop 286-200
Laptop 486-200
Maximum profit is
Z=5000120+3400200+3000200=Sh1,880,000
Unused resources include the following
JK computers can still produce 180 more desktop models (500-120-200) and 50 laptop
models (250-200)
For used up resources the prices to pay for any additional unit are as follows
Sh150 for 80386 chip
Sh90 for 80286 chip
Sh20 for any hour
ii)
The range for the variables x1, x2 and x3 are to indicate where the number of units can
change without affecting the basic solution
c)
The range for the constraints indicate the extent the resources can be changed without
altering the basic solution of the linear programming problem
iii) The dual value of 80386 chip is Sh 150. That is the addition increase in profit due to
increase of one chip. So if the company increases the number of chips by 10, the
additional profit will be 10150=Sh1,500.
QUESTION SEVEN
a) Network
EST
LST
1 6
10 10
8 8
12 12
15 15
4 4
The critical path is A, E, F, H, I.

The duration of the critical path is 15 days.
b) Assuming all the activities start as soon as possible, the following chart shows when
activities will start and finish.
356
Lesson Nine
0
10
11 12
13
14
15
Time scale
(days)
Critical
activities
Other
activities
G
C
Consequently the following resource allocation is required on a day to day basis.

Day
1
2, 3
4
5, 6
7
8
9, 10
11, 12
13, 14, 15
Activities
A, B
A, D, G
A, G
C, E, G
C, E
E
F
H
I
Number of increased staff

1+3
=4
1+1+1
=3
1+1
=2
2+2+1
=5
2+2
=4
2
=2
2
=2
1
=1
1
=1
Current costs with 5 staff = 15 days 5 500 = 37500

If activity C is delayed to start on day 7 only 4 staff are required and the project duration
unchanged. The cost is
15 days 4 500 = 30000
However if the duration is increased by starting B on Day 1 and delaying activities A, E, F, H, I
by one day and C until day 12 only a maximum of 3 staff are required.
Revision Aid
357
MOCK EXAMINATION
Instructions:

SECTION I
QUESTION ONE
a) Explain the importance of set theory in business
(4marks)
(8marks)
b) By use of matrix algebra, develop the leontief inverse matrix.
c) Digital ltd manufactures and sells floppy disks at Nairobi industrial area. The average
revenue (AR)(in thousands of shillings )of producing x floppy disks are given by the
following functions
ATC = 12 x 2 52 x + 50 + 50x
And
AR=800-2X2
Where: x is the number of floppy disks produced
Required:
i.
The profit function
The number of floppy disks required to maximize profit
ii.
iii.
The maximum profit
(3marks)
(3marks)
(2marks)
(Total: 20marks)
QUESTION TWO
a) State any five problems encountered in the construction of the consumer price index
(5marks)
b) An investment analyst gathered the following data on the 91-day Treasury bill rates for the
years 2003and 2004
Month
Treasury bill rates (%)
2003
2004
January
3.2
5.5
February
3.0
5.2
March
2.8
4.3
April
2.5
3.6
May
2.9
3.3
June
3.4
2.7
July
3.7
2.4
August
4.0
2.0
September
3.8
2.3
October
4.2
2.8
November
4.5
3.1
December
5.1
3.7
The analyst would like to determine if on average there was a significant change in the
Treasury bill rates over the two years.
358
Lesson Nine
Required:
i.
The mean and variance of the Treasury bill rates for each year
(10marks)
Determine if there is a significant difference in the average Treasury bill rates (use a
ii.
significant level of 1%).
(5marks)
Note: S2=
(n1 1)S12 + (n 2 1)S 22

n1 + n 2 2
QUESTION THREE
a) Describe the characteristics of the following statistical distributions
Binomial distribution
i.
Poisson distribution
ii.
(3marks)
(3marks)
b) High Grade Meat Ltd produces beef sausages And sells them to various supermarkets .In
order to satisfy the industrys requirements ,the firm may only produce 0.2percent of
sausages below a weight of 80 grammes .The sausage producing machine operates with a
standard deviation of 0.5 grammes .The weights of the sausages are normally distributed
The firms weekly output is 300,000sausages and the sausage ingredients cost shs5.00
per 100 grammes ,sausages with weights in excess of 82 grammes require additional
ingredients costing sh 2.50 per sausage
Required
i.
The mean weight at which the machine should be set
ii.
The firms weekly cost of production
(4marks)
(10marks)
(Total: 20marks)
QUESTION FOUR
a) A survey of undergraduate students at High Fliers University (HFU)showed the following
results regarding gender and the fields of specialization in their studies
Gender
Male
Female
Total
Field of specialization
Business
Science
Arts
100
250
100
50
100
200
300
200
300
Total
450
350
800
Required
i.
Determine if the field of specialization in the studies is dependent on gender (use
(10marks)
An earlier survey showed that the proportion of female students taking science was only
ii.
10%of the total student population taking science .Does the data above show any
significant improvement in the proportion of female students taking science (use a
(6marks)
b) Charles Nzioka who is a barber has found out that he can shave on average 4 customers per
hour .The arrival rate of customers averages 3customers per hour
Required
i.
The proportion of time that Charles Nzioka is idle
(1 mark)
The probability that a customer receives immediate service upon arrival (1mark)
ii.
Average number of customers in the queuing system
(1 mark)
iii.
Revision Aid
iv.
359
Average time a customer spends in the queuing system
(1 mark)
(Total: 20marks)
QUESTION FIVE
a.
Differentiate between the additive model and the multiplicative model as used in time series
analysis
(4marks)
b. The sales data of XYZ Ltd (in millions of shillings) for the year 2001 to 2004 inclusive are
given below.
Quarter
Year
1
2
3
4
2001
40
64
124
58
2002
42
84
150
62
2003
46
78
154
96
2004
54
78
184
106
Required
i.
The trend in the data using the least square method
(8marks)
The estimated sales for each quarter of the year 2004
(4marks)
ii.
The percentage variation of each quarters actual sales for the year 2004 (4marks)
iii.
(Total: 20marks)
SECTION II
QUESTION SIX
a. Give two applications of simulation in business
(2marks)
b. Collins Simiyu recently acquired a piece of land in Kitale .A property development company
has offered him 300,000 for the piece of land .He has to make a decision on whether to
cultivate the piece of land or to sell it to the property development company If he decides to
cultivate the land ,there is a probability of getting a high ,medium ,or low harvest .The
expected net income for each of the above states of harvest is shown below:
State of harvest
Net income (sh)
High
500,000
Medium
100,000
Low
(20,000)
From past experience there is a 10percent probability that the harvest will be low, a 30 per
cent probability that the harvest will be medium and a 60percent probability that the harvest
will be high .Colins Simiyu can engage an agricultural expert to carry out a survey on the
productivity of the land which will cost him sh30, 000. The agricultural expert gives the
following information as to the reliability of such surveys (prior probabilities)
Results of survey
Accurate
Not accurate
High
0.35
0.25
0.60
state of harvest
Medium
0.10
0.10
0.20
Low
0.05
0.15
0.20
Required
i.
Construct a decision tree for the above problem
The expected monetary value for each decision
ii.
The decision that you would recommend
iii.
Total
0.5
0.5
1.0
(6marks)
(10marks)
(2marks)
(Total: 20marks)
360
Lesson Nine
QUESTION SEVEN
a. Explain the difference between assignment and transportation problems (4marks)
b.
State the assumptions made in solving a transportation problem
(4marks)
Umoja Engineering Works Ltd has a network of branches all over Kenya .The
c.
branches are used to service, repair and install equipment for their clients .Currently,
the Nairobi branch has four clients who require installation of equipment .Each
client requires the services of one engineer
There are four engineers who are not engaged at the moment and can be assigned
any one of the tasks .However, these engineers have to travel from different
locations and the Nairobi branch has to meet their travel and subsistence
allowances. The allowances vary from one engineer to another and according to the
client the engineer has been assigned to work for.
The table below shows the costs (in thousand of shillings) associated with each
engineer
Client
1
2
3
4
Engineer
A
37.0
27.0
34.0
21.0
B
57.0
22.0
79.0
34.0
C
22.0
25.0
61.0
45.0
D
39.0
42.0
54.0
43.0
Required
i.
The assignments to be made in order to minimise the total cost of
engineers
(10marks)
The minimum cost of using the engineers
(2marks)
ii.
(Total 20marks)
QUESTION EIGHT
a.
b.
Define the following terms as used in network analysis:

i.
Crash time
Optimistic time
ii.
Forward pass
iii.
Dummy activity
iv.
Slack
v.
(2marks)
(2marks)
(2marks)
(2marks)
(2marks)
James Mutiso is a computer engineer in an information technology firm .The firm has
decided to install a new computer system to be used by the firms helpdesk .James
Mutiso has identified nine activities required to complete the installation.
The table below provides a summary of the activities durations and the required number
of technicians
Activity
1-2
1-3
2-4
2-5
3-4
3-6
4-5
5-6
6-7
Duration(weeks)
3
1
3
2
2
4
2
2
2
Required number of technicians

2
4
4
2
4
4
2
2
2
Revision Aid
361
Required:
i.
Draw a gantt chart for the project
(6 marks)
Mr. Mutiso would like to reschedule the activities so that not more than 6 technicians are
ii.
required each week
Determine if this is possible and how it can be achieved by rescheduling the activities.
(4marks)
(Total 20marks)

Documents - MX - QT Study Pack 2006 Wip PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Documents - MX - QT Study Pack 2006 Wip PDF

Uploaded by

Copyright:

Available Formats

ii

We gratefully acknowledge permission to quote from the past examination

STRATHMORE UNIVERSITY STUDY PACK

Instructions for Students

Instructions for Students

Quantitative techniques paper in section 4 of the CPA examination is an eight

1. Linear Algebra and Matrices.....................................................................................1

2. Sets Theory and Calculus........................................................................................45

3. Descriptive Statistics and Index Numbers ..............................................................81

4. Measures of Relationships and Forecasting..........................................................118

6. Sampling and Estimation ......................................................................................185

8. Operation Research ...............................................................................................248

9. Revision Aid .........................................................................................................303

STRATHMORE UNIVERSITY STUDY PACK

Quantitative techniques is a mathematical paper and fundamental to many proffessionalcourses

STRATHMORE UNIVERSITY STUDY PACK

Linear Algebra and Matrices

5x2 + 3x + 7 = 0 (quadratic equation)

Solution of Linear Equation

(by division rule)

Solution of quadratic equations

Linear Algebra and Matrices

Which is a pure quadratic equation

(5x 3) (3x +5) = 0 ..............................................(step 2)

Solution by Completing the Square

STRATHMORE UNIVERSITY STUDY PACK

Factorize the perfect square on the left hand side.

Find the square root of both sides

Linear Algebra and Matrices

STRATHMORE UNIVERSITY STUDY PACK

Rules for solving linear inequations

Linear Algebra and Matrices

Rule 2: Subtraction Rule

Rule 3: Multiplication rule

M/P > N/P: M1/P < N1/P

M/(-P) < N/(-P) : and M1/(-P) > N1/(-P)

STRATHMORE UNIVERSITY STUDY PACK

(by subtraction rule)

(by subtraction rule)

(by substraction rule)

Linear inequation in two variables: relations

Linear Algebra and Matrices

The above figure illustrates:

b) The elimination technique

STRATHMORE UNIVERSITY STUDY PACK

c) The substitution technique

The solution of this system can be obtained by

Linear Algebra and Matrices

STRATHMORE UNIVERSITY STUDY PACK

The following matrices are equal 2 2 3 = 2 2 3

Column Matrix or column vector

Linear Algebra and Matrices

Row matrix or row vector

STRATHMORE UNIVERSITY STUDY PACK

is a square matrix of order 2

B = n n is a square matrix of the order n

An identity of unity matrix

A null or zero matrix

Linear Algebra and Matrices

Multiplying a matrix by a number

STRATHMORE UNIVERSITY STUDY PACK

Let A = ( 4 5 6 ) and B = 200

then ( 4 5 6 ) 200 = 400 + 1,000 + 1,800 = Shs 3,200