1051 Workbook

MATH1051
CALCULUS
AND
LINEAR ALGEBRA I
Semester 1, 2015
Lecture Workbook
How to use this workbook

This book should be taken to lectures, tutorials and lab sessions. In the lectures, you will be expected
to fill in the blanks in any incomplete definitions, theorems etc, and any examples which appear. The
lecturer will make it clear when they are covering something that you need to complete.
The completed workbook will act as a study guide designed to assist you in working through assignments
and preparing for the final exam. For this reason, it is very important to attend lectures.
The text for the calculus part of the course is Calculus by James Stewart, 7th edition. We often refer
to the text in this workbook, and many of the definitions, theorems, and examples come from the text.
(References are also provided for the 6th edition). It is strongly advised that you purchase a copy.
Although there is no set text for the linear algebra part of the course, there are several books in the
library which cover the material well (see the course website).
c
Mathematics,
School of Physical Sciences, The University of Queensland, Brisbane QLD 4072, Australia
Edited by Joseph Grotowski, Phil Isaac, Michael Jennings, Birgit Loch, Barbara Maenhaut, Victor Scharaschkin, Mary Waterhouse,
Poh Wah Hillock, February 2015.
CONTENTS
Contents
0 MATLAB
0.1
Introduction to MATLAB . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
0.2
Plotting Functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
0.3
Vectors & Matrices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
0.4
Systems of Linear Equations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
0.5
Eigenvalues & Eigenvectors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
0.6
Plotting points . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
0.7
Introduction to M-files . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
10
0.8
Writing Functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
14
0.9
Loops . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
15
0.10 Useful Commands for MATH1051
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
17
0.11 Appendix . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
18
1 Complex Numbers
20
1.1
Notation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
20
1.2
Complex Numbers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
20
1.3
Polar form . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
22
1.4
Eulers formula . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
24
2 Vectors
26
2.1
Review . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
26
2.2
Position Vectors
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
27
2.3
Definition: Norm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
27
2.4
Vector Addition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
27
2.5
Scalar multiplication . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
28
2.6
Unit Vectors
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
28
2.7
Vectors in 3 dimensions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
29
2.8
Row and column vectors
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
31
2.9
Dot Product
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
34
2.10 The Projection Formula . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
36
2.11 Vectors in Rn . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
38
2.12 Properties of Dot Product . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
39
CONTENTS
3 Matrices and Linear Transformations
42
3.1
2 2 matrices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
42
3.2
The effect of matrix multiplication on vectors . . . . . . . . . . . . . . . . . . . . . . . . .
42
3.3
Composing Transformations & Matrix Multiplication . . . . . . . . . . . . . . . . . . . . .
43
3.4
Definition: Matrix . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
44
3.5
Equality . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
44
3.6
Addition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
44
3.7
Scalar Multiplication . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
45
3.8
Matrix Multiplication . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
45
3.9
Transposition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
47
4 Gaussian elimination
50
4.1
Simultaneous equations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
50
4.2
Gaussian elimination . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
50
4.3
The general solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
60
4.4
Gauss-Jordan elimination . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
62
5 Inverses
65
5.1
The Identity Matrix . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
65
5.2
Definition: Inverse . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
66
5.3
Inverses via GaussJordan. Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
71
6 Determinants
76
6.1
Definition: Determinant . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
76
6.2
Properties of Determinants . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
77
6.3
Connection with inverses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
82
7 Vector Products in 3-Space
85
7.1
Definition: Cross product . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
85
7.2
Application: area of a triangle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
88
7.3
Scalar Triple Product . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
89
7.4
Geometrical Interpretation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
90
8 Eigenvalues and eigenvectors
94
8.1
Ranking Webpages . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
94
8.2
Geometry of eigenvectors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
96
8.3
Eigenvalues & Vectors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
98
8.4
How to find eigenvalues . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
99
CONTENTS
9 Vector Spaces
104
9.1
Linear combinations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104
9.2
Linear independence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106
9.3
How to test for linear independence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106
9.4
Invertible matrices and linear independence . . . . . . . . . . . . . . . . . . . . . . . . . . 109
9.5
Vector spaces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111
9.6
Eigenspace . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116
9.7
The span of a set of vectors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117
9.8
Bases
9.9
Dimension . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 120
9.10 Components . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 122

9.11 Further properties of bases
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123
10 Numbers
126
10.1 Number systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126

10.2 Real number line and ordering on R . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126
10.3 Definition: Intervals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126
10.4 Absolute value . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128
11 Functions
131
11.1 Notation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 131

11.2 Definition: Function, domain, range . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 131
11.3 Graphs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132
11.4 Convention (domain) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133
11.5 Caution (non-functions) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133
11.6 Vertical line test . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 134
11.7 Exponential functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 135
11.8 Trigonometric functions (sin, cos, tan) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 137
11.9 Composition of functions
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 142
11.10One-to-one (1-1) functions - Stewart, 7ed. pp. 384-387; 6ed. pp. 385-388 . . . . . . . . . . 142
11.11Inverse Functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 144
11.12Logarithms - Stewart, 7ed. p. 404; 6ed. p. 405
. . . . . . . . . . . . . . . . . . . . . . . . 146
11.13Natural logarithm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 146

11.14Inverse trigonometric functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 150
CONTENTS
12 Limits
156
12.1 Definition: Limit - Stewart, 7ed. p. 50; 6ed. p. 66 . . . . . . . . . . . . . . . . . . . . . . . 156

12.2 One-sided limits . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 160
12.3 Theorem: Squeeze principle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 162
12.4 Limits as x approaches infinity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 163
12.5 Some important limits . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 166
13 Sequences
170
13.1 Formal Definition: Sequence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 170

13.2 Representations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 170
13.3 Application: Rabbit population - Stewart, 7ed. p. 726 ex. 83; 6ed. p. 722 ex. 71 . . . . . . 172
13.4 Limits . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 173
13.5 Theorem: Limit laws . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 174
13.6 Theorem: Squeeze . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 175
13.7 Useful sequences to remember . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 175
13.8 Application: Logistic sequence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 176
14 Continuity
179
14.1 Definition: Continuity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 179

14.2 Definition: Continuity on intervals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 181
14.3 Properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 182
14.4 Intermediate Value Theorem (IVT) - Stewart, 7ed. p. 89; 6ed. p. 104 . . . . . . . . . . . . 183
14.5 Application of IVT (bisection method) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 183
15 Derivatives
187
15.1 Tangents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 187

15.2 Definition: Derivative, differentiability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 188
15.3 Theorem: Differentiability implies continuity . . . . . . . . . . . . . . . . . . . . . . . . . 190
15.4 Rates of change - Stewart, 7ed. p. 108; 6ed. p. 117 . . . . . . . . . . . . . . . . . . . . . . 190
15.5 Higher derivatives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 192
15.6 Rules for differentiation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 193
15.7 The chain rule - Stewart, 7ed. p. 148; 6ed. p. 156 . . . . . . . . . . . . . . . . . . . . . . . 194
15.8 Derivative of inverse function . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 196
15.9 LH
opitals rule - Stewart, 7ed. p. 470; 6ed. p. 472 . . . . . . . . . . . . . . . . . . . . . . 198
15.10Theorem: Continuous extension . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 199
15.11The Mean Value Theorem (MVT) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 200
15.12Definition: Increasing/Decreasing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 201
CONTENTS
15.13Increasing/Decreasing test - Stewart, 7ed. p. 214; 6ed. p. 221 . . . . . . . . . . . . . . . . 201

15.14Definition: Local maxima and minima . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 202
15.15First derivative test . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 204
15.16Second derivative test . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 205
15.17Extreme value theorem - Stewart, 7ed. p. 199; 6ed. p. 206 . . . . . . . . . . . . . . . . . . 209
15.18Optimization problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 210
16 Series
217
16.1 Infinite sums (notation) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 217

16.2 Motivation
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 217
16.3 Definition: Convergence - Stewart, 7ed. p. 729; 6ed. p. 724 . . . . . . . . . . . . . . . . . 218

16.4 Theorem: p test . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 219
16.5 Theorem: nth term test - Stewart, 7ed. p. 733; 6ed. p. 728 . . . . . . . . . . . . . . . . . 220
16.6 Test for divergence (nth term test) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 220
16.7 Proof that the Harmonic Series Diverges . . . . . . . . . . . . . . . . . . . . . . . . . . . . 221
16.8 Example: Geometric series . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 221
16.9 Application: Bouncing ball - Stewart, 7ed. p. 736 qn 72; 6ed. p. 731 qn 58 . . . . . . . . . 223
16.10Comparison tests . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 225
16.11Theorem: Comparison test - Stewart, 7ed. p. 746; 6ed. p. 741 . . . . . . . . . . . . . . . . 225
16.12Alternating series . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 226
16.13Theorem: Alternating series test - Stewart, 7ed. p. 751; 6ed. p. 746 . . . . . . . . . . . . . 227
16.14Definition: Absolute/conditional convergence . . . . . . . . . . . . . . . . . . . . . . . . . 229
16.15Theorem: Ratio test - Stewart, 7ed. p. 758; 6ed. p. 752 . . . . . . . . . . . . . . . . . . . 231
17 Power series and Taylor series
234
17.1 Definition: Power series . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 234

17.2 Theorem: Defines radius of convergence - Stewart, 7ed. p. 767; 6ed. p. 761 . . . . . . . . 236
17.3 Taylor series . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 237
17.4 Theorem: Formula for Taylor series . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 238
17.5 Binomial series . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 243
18 Integration
248
18.1 Definition: antiderivative - Stewart, 7ed. p. 269; 6ed. p. 274 . . . . . . . . . . . . . . . . . 248

18.2 Definition: Indefinite integral . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 249
18.3 Area under a curve - Stewart, 7ed. pp. 284-291; 6ed. pp. 289-296 . . . . . . . . . . . . . . 249
18.4 Fundamental Theorem of Calculus . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 252
18.5 Approximate integration (trapezoidal rule)
. . . . . . . . . . . . . . . . . . . . . . . . . . 256
CONTENTS
18.6 Improper integrals - Stewart, 7ed. pp. 543-551; 6ed. pp. 544-551 . . . . . . . . . . . . . . 257
18.7 Techniques of integration - Stewart, 7ed. ch. 7; 6ed. ch. 8 . . . . . . . . . . . . . . . . . . 259
18.8 Integrals involving ln function . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 264
18.9 Integration of rational functions by partial fractions . . . . . . . . . . . . . . . . . . . . . 264
18.10Volumes of revolution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 269
19 Appendix - Practice Problems
273
19.1 Numbers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 273

19.2 Functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 273
19.3 Limits . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 274
19.4 Continuity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 275
19.5 Derivatives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 276
19.6 Integration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 277
19.7 Sequences . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 279
19.8 Series . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 280
19.9 Power series and Taylor series . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 281
19.10Vectors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 282
19.11Matrix Algebra . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 283
20 Appendix - Mathematical notation
287
21 Appendix - Basics
288
21.1 Powers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 288

21.2 Multiplication/Addition of real numbers . . . . . . . . . . . . . . . . . . . . . . . . . . . . 289
21.3 Fractions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 289
21.4 Solving quadratic equations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 290
21.5 Surds . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 292
0.1 Introduction to MATLAB
MATLAB
0.1
0.1.1
Introduction to MATLAB
Getting Started
Log in using your library username and password. Ask your tutor if you have a problem.
Open MATLAB just like any other Windows program.
StartPrograms MATLABMATLAB R2007a
Save all files to your student directory on H:/. Files saved onto individual computers or S:/ will
be deleted at the end of semester.
0.1.2
The MATLAB Program Window
The MATLAB program window consists of the command window, command history, current directory
and a menu bar for opening and editing M-files.
The command window is the main window we will use. Simple calculations can be performed in the
command window. Executing functions, defining variables, viewing answers and information about errors
which may have occurred also all happen in the command window.
A history of the commands entered into the command window can be found in the command history.
Selecting a command from this window will re-execute the command in the command window.
The current directory contains a list of files that MATLAB can utilize. This is also the folder in which
MATLAB will save your work. The current directory should be set to your student account on H:/.
0.1.3
Getting Help
There are a number of resources available should you need assistance when using MATLAB. The most
easily accessible is MATLABs own built-in Help system.
Clicking the blue question mark button in the toolbar will open the Help window, as will typing
doc in the command window. You can search for the topic in question or navigate using the panel
on the left.
If you know the name of the function or operation you need help with, typing help <function
name> in the command window will display a short description of the function and how to use it.
Typing doc <function name> will open the Help window and navigate to the page dedicated to
that function or operation.
If you have trouble finding what you need with MATLABs Help system, there is a great deal of documentation and useful articles available online. The Mathworks website
http://www.mathworks.com/products/matlab/
has a lot of helpful descriptions and solutions which go beyond what is available in the MATLAB Help,
as well as m-files written and uploaded by users. A Google search often turns up results as well.
Practical sessions will rely heavily on the following resources:
http://www-mdp.eng.cam.ac.uk/web/CD/engapps/octave/octavetut.pdf
http://www.cs.ubc.ca/ nando/cpsc530a/handouts/cam matlab.pdf
Of course, your tutors and lecturers will be able to help you out during class or consultation times, and
the First Year Learning Centre is open from 24pm in 67641 every day.
0.1.4
Variables & Basic Commands
Variables in MATLAB are names for mathematical objects (eg. numbers, vectors, matrices) which
are defined by the user, stored in MATLABs memory (for the current session), and can be used in
calculations. Variables can be defined (assigned values) by using = . For example, the commands
>> x = 3
x =
3
>> y = 1009.781
y =
1009.781
assign the value 3 to the variable x and the value 1009.781 to the variable y.
Variable names can be letters or words, and can contain numbers, so long as they begin with a letter
(eg. a1 is a valid variable name, but 1a is not). Variable names are case sensitive (eg. MATLAB treats
x and X as different variables).
Using the command clear followed by the name of a variable will remove that variable from MATLABs
memory. Using clear all will clear all variables at once. Note that it is not necessary to clear a variable
before redefining it. However, it is often a good idea to clear a variable after it is used to avoid mistakenly
using an incorrect value in future calculations.
MATLAB can be used in much the same way as a normal calculator for evaluating mathematical expressions and functions. Operations can be performed on numbers or variables (as long as the variables have
already been defined). Some operations, functions and constants are saved permanently in MATLAB.
We demonstrate a number of these here; a more thorough list can be found in the Appendix.
As with any computer language, commands in MATLAB must follow certain syntax rules. If you get an
error message after trying to execute a command, check that you have not made a mistake when typing
the command (eg. brackets, apostrophes, commas).
The MATLAB commands for basic mathematical operations are exactly what one would expect. Addition, subtraction, multiplication and division are all demonstrated below.
>> x = 10
x =
10
>> y = 5;
>> x+y
ans =
15
>> x-y
ans =
5
>> x*y
ans =
50
>> x/y
ans =
2
Note that after entering the command x = 10 MATLAB provided feedback, but after the command y =
5; no feedback was produced. Following a command with a semicolon will stop MATLAB from displaying
any output, though the command itself will still be executed.
Some other common operations (given below) include exponentiation, square root, log and sin. Note that
log is the natural logarithm (base e), not base 10.
>> x = 10
x =
10
>> y = 5
y =
5
>> x^y
ans =
100000
>> sqrt(x)
Tip 1: Entering a variable name as a command will return the value of that variable.
Tip 2: To remove a variable from MATLABs memory, use the command clear followed by the variable
name. To clear all variables at once, use clear all.
Tip 3: For help with any MATLAB command simply enter help followed by the command name (eg.
help acos).
Tip 4: By default MATLAB will display numbers to 4 decimal places. The command format long
changes this to 15 decimal places. To revert back to 4 decimal places, use format short.
Tip 5: Following any command with a semicolon will stop MATLAB from displaying any output (though
the operation will still be performed).
ans =
3.1623
>> log(x)
ans =
2.3026
>> sin(x)
ans =
-0.5440
>> exp(x)
ans =
2.2026e+004
The last answer above demonstrates MATLABs use of scientific notation: it is the number 2.2026 104 .
Do not confuse the e with Eulers number e = 2.718 . . .; to obtain this number, use the command
exp(1).
MATLAB has a number of other constants built-in, such as and i. For example, the commands
>> x = 3*pi
x =
9.4248
>> y = log(x)
y =
2.2433
>> clear all
assign the value 3 to the variable x, set y to be ln(x), and then clear MATLABs memory of all variable
values.
Check the Appendix for a more complete list of commonly used functions and constants.
Division of a non-zero number by 0 will return the result Inf or -Inf (representative of + and
respectively). Any other mathematically undefined operation will give the result NaN, which stands for
Not-a-Number. For example,
>> 0/0
ans =
NaN
In normal circumstances, Inf and NaN should be considered errors.
Tip 6: Pressing the up arrow in the command window will scroll up through previously executed
commands.
Tip 7: The result of the most recently executed command is stored as the variable ans.
Tip 8: The command who gives a list of all currently defined variables.
Tip 9: Some operations that might normally be considered undefined (eg. log(1) or arccos(2)) will
return complex answers in MATLAB. The reasons for this are not covered in MATH1051 (see courses on
complex analysis for details).
0.2 Plotting Functions
0.2
0.2.1
Plotting Functions
General ezplot Commands
MATLAB has a number of built-in commands to make plotting functions very easy. MATLAB plots
functions in a separate Figure window which will automatically open. To plot f (x) = sin(x) over
MATLABs default domain 2 < x < 2, use the command
>> ezplot(sin(x))
In general, to plot an arbitrary function f (x) type
>> ezplot(f(x))
Note that the symbols cannot be left out here.
MATLAB will endeavour to automatically assign a sensible range to the dependent variable if it is not
specified. The domain and range can, however, both be controlled by the user. The first command below
assigns the domain a x b, and the second assigns the same domain as well as range c f (x) d.
>> ezplot(f(x),[a,b])
>> ezplot(f(x),[a,b,c,d])
0.2.2
Labelling & Editing Figures
After a figure has been generated in MATLAB, it can be labelled and edited through the command
window (check MATLABs help file for instructions on how to do this) or through the menus in the figure
window itself. Axis labels, titles and legends can be added and modified through the Insert menu. To
edit the plots themselves, use View Property Editor.
0.3 Vectors & Matrices
0.3
0.3.1
Vectors & Matrices

Basic Vector & Matrix Operations
One of MATLABs strengths is its support for vectors and matrices. Performing operations with such
objects is very easy in MATLAB, which has a wide range of built-in functions for manipulating them.
The following commands demonstrate how to define the matrices and vectors

4

2 4
a= 1 3 4 ,
b= 0 ,
and
A=
0 1
7
>> a = [1,3,4]
a =
1
3
4
>> b = [4;0;7]
b =
4
0
7
>> A = [2,4;0,7]
A =
2
4
0
7
Note that commas separate entries in a row and semicolons separate rows.
The following MATLAB segment details the commands which can be used to take the transpose of matrix
a; print the (1, 2) entry of matrix A; the inverse of A; and the product of A with itself.
>> a
ans =
1
3
4
>> A(1,2)
ans =
4
>> inv(A)
ans =
0.5000
0
>> A*A
ans =
4
36
0
49
-0.2857
0.1429
Tip 10: To plot multiple graphs on the same axes, use the command hold on before plotting the first
one, and then hold off when finished.
Tip 11: A set of grid lines can be toggled on and off using the command grid, or through the Property
Editor.
Tip 12: The name given to the independent variable in ezplot does not matter.
ezplot(sin(x)) and ezplot(sin(t)) will produce the same output.
For example,
0.4 Systems of Linear Equations
Some functions which are normally applied to single numbers can be applied to all the elements of a
matrix at once. For example,
>> log(a)
ans =
0
1.0986
1.3863
Here, applying the log function to a vector gives the natural logarithm of each element of the vector.
Some functions require a full stop be inserted into the expression to operate in this way; for example,
you can square each element of a matrix as follows:
>> a.^2
ans =
1
16
In general, functions which have special definitions when applied to matrices (for example, multiplication
or exponentiation) will require the full stop, while functions which can only be applied to single numbers
(like logarithms) will not. Consult MATLABs documentation to determine whether a function can
operate on matrices, and whether it requires the full stop operator. See the Appendix for a more
thorough list of MATLAB operations which can be performed on matrices and vectors, as well as the
commands for some special matrices.
0.4
Systems of Linear Equations
You have seen how to create and manipulate matrices in MATLAB. Systems of linear equations can be
written as matrix equations, so we can use MATLAB to easily solve such problems.
Consider the system of linear equations
2x +
x +
You know from lectures that finding

Ax = b, where
2 1
A = 1 1
0 2
y z =
1
y 2z = 1
2y z =
3
the solution to this is equivalent to solving the matrix equation
1
2 ,
1
1
b = 1
3

x
and x = y
z
Remember that the commands

>> A = [2 1 -1; 1 1 -2; 0 2 -1]
A =
2
1
-1
1
1
-2
0
2
-1
>> b = [1; -1; 3]
Tip 13: When entering a matrix, commas separate entries within a row and semicolons start new rows.
Commas can be replaced with spaces (eg. matrix A in the table could be entered as A=[2 4;0 1]).
Tip 14: In MATLAB, the column and row indices of vectors and matrices start at 1, not 0. So the value
in the upper left corner of a matrix A is A(1, 1), not A(0, 0).
0.4 Systems of Linear Equations
b =
1
-1
3
will enter the values of A and b into MATLAB.
0.4.1
Solution by Inversion
One method for solving a matrix equation Ax = b is by noting that x = A1 Ax = A1 b, as long as A1

exists. So to solve the above system, use the command
>> x = inv(A)*b
x =
0.2000
2.4000
1.8000
after entering the values of A and b.
0.4.2
Solution by Gaussian Elimination
Another method for solving systems of linear equations is Gaussian elimination. The function rref
performs Gaussian elimination to find the reduced row echelon form of a matrix. After finding the
reduced row echelon form of an augmented matrix (A|b), the solution to Ax = b can just be read off.
To augment two matrices in MATLAB (after first making sure that the dimensions match up correctly),
simply put them in square brackets separated by a space or comma. For example, to augment A and b
above and then solve Ax = b, use the commands
>> AugA = [A b]
AugA =
2
1
-1
1
1
-2
0
2
-1
>> rref(AugA)
ans =
1
0
0
0
1
0
0
0
1
1
-1
3
0.2000
2.4000
1.8000
Another way to solve a system using Gaussian elimination is to use MATLABs backslash operator.
To solve Ax = b use the command
>> x = A \ b
x =
0.2000
2.4000
1.8000
Tip 15: The command AugA = [A b] concatenates (joins) matrices A and b horizontally, while AugA
= [A; b] concatenates A and b vertically. In both cases, the lengths of the sides being joined together
must be the same.
0.6 Plotting points
0.5
Eigenvalues & Eigenvectors
MATLAB has a built in function called eig to calculate the eigenvalues of a matrix and give corresponding
eigenvectors. For example, to calculate the eigenvalues and eigenvectors of the matrix

1 3
A=
,
4 2
use the following commands.
>> A = [1 3; 4 2];
>> [V,D] = eig(A)
V =
-0.7071
-0.6000
0.7071
-0.8000
D =
-2
0
0
5
Here, the values on the diagonal of D are the eigenvalues of A, and the columns of V are the corresponding
eigenvectors. So A has an eigenvalue = 2 with associated eigenvector x = [0.7071, 0.7071]T and
another eigenvalue = 5 with associated eigenvector x = [0.6, 0.8]T .
0.6
Plotting points
You have seen how to use the ezplot command to plot functions of one variable. There are many
situations where you may instead wish to plot a set of discrete points on a cartesian plane. MATLABs
plot command can be used to do this.
To plot a set of points {(x1 , y1 ), (x2 , y2 ), . . . , (xn , yn )}, use the command plot(x,y), where x = [x1 , x2 , . . . , xn ]
and y = [y1 , y2 , . . . , yn ]. For example, to plot the points (0, 0), (1, 1), (2, 3), (4, 6), use the commands
>> x=[0,1,2,3];
>> y=[0,1,3,6];
>> plot(x,y)
Tip 16: Be careful when using the command [V,D] = eig(A). No matter what you call V and D, the
one which comes first will be the matrix of eigenvectors, and the other will be the matrix of eigenvalues.
Tip 17: You can use hold on to simultaneously plot continuous functions with ezplot and discrete
points with plot.
0.7 Introduction to M-files
10
There is an optional third argument for plot which specifies the colour, type of point and type of line
which will be used to plot the points. For example,
>> plot(x,y,ro)
will plot the points as red circles with no line joining them, and
>> plot(x,y,mx-)
will plot the points as magenta crosses joined by a solid line. Use the command help plot for a full
explanation.
As with ezplot, the properties of a figure (eg. title, axis labels, colour) can be edited using the menus
in the figure window.
0.7
Introduction to M-files
M-files are text files which contain sequences of operations which MATLAB will execute when the files
are run. In comparison to most programming languages, m-files are very easy to write, and are usually
built around only a few core commands.
0.7.1
Creating & Opening M-files
To create a new m-file; either

Click on the New M-File button in the menu bar; or
Select File New M-File.
11
You should save m-files you wish to keep in your H:/ directory.
To open a saved m-file:
Use the drop-down menu at the top of the screen to navigate to the correct folder, and then
double-click the file in the Current Directory panel;
Click the Open file button in the menu bar; or
Select File Open.
0.7.2
Writing & Running M-files
All of the commands which we have thus far covered will work as part of an m-file. For example, create
a new m-file, write the commands
x = sqrt(5);
y = exp(x)
and save the file as example.m. In the main MATLAB window, use the drop-down menu to navigate to
the folder where the file is saved, and then enter the command
>> example
y =
9.3565
in the command window. This illustrates the basic idea of using m-files: you write a sequence of commands, save them all together into an m-file, and then execute them all in order by calling the name of
the m-file as a command. Of course, the above example is very simple, and would take less time to just
evaluate in the command window.
M-files are far more useful when they are intended to be used over and over again, because the commands
only need to be typed out once. For example, say we have a collection of numbers (which we can store as
a vector), and we wish to perform some statistical analysis by calculating the size of the set, the mean,
median, mode, range and standard deviation, and we also wish to sort the numbers into ascending order.
MATLAB has built-in commands for all these operations. If we only have a single data set then we
may as well just type all the commands into the command window; however, if we have many sets then
we would be typing the same commands over again for each set. We can save time by putting all the
commands into an m-file. Put the following into an m-file and save it as statanalysis.m.
% For a vector x of real numbers, outputs the size of the set x,
% the mean, median, mode, range and standard deviation of the data,
% and sorts the data into numerical order.
setsize = length(x)
meanvalue = mean(x)
medianvalue = median(x)
modevalue = mode(x)
range = max(x) - min(x)
standarddev = std(x)
sortedx = sort(x)
Now we can perform all the operations at once by entering the vector x and then running statanalysis.
For example,
>> x = [1,4,10,-2,19,0,6,3,4,7];
>> statanalysis
setsize =
10
meanvalue =
5.2000
medianvalue =
4
modevalue =
4
range =
21
standarddev =
5.9777
sortedx =
-2
0
12
10
19
If we now have another data set we wish to analyze, we only need to redefine x to be that set, and run
statanlysis again.
0.7.3
If Statements
While m-files can be used to save time by collecting together operations which would otherwise take
time to execute in the command window, there are a number of special commands which make m-files
considerably more powerful. The first one we will look at is the if statement. It has the basic form
if condition
commands
end
where condition is a true/false statement and commands is a set of operations for MATLAB to perform.
These commands will be executed if condition is true; if they are false, MATLAB will do nothing. For
example, a valid if statement might be
if x>=0
y=sqrt(x)
end
Here, if x is greater than or equal to 0, then y will be the square root of x. If x is negative then MATLAB
does nothing. If we wanted MATLAB to do something when x is negative, we could add a second if
statement. However, an else command can be inserted into the if statement to accommodate this:
if condition
1st commands
else
2nd commands
end
For example, lets write a program to calculate the absolute value of a number. Create a new m-file and
insert the following code.
if x < 0
y=-x
else
y=x
end
Save the file as absval.m. Set the current directory to the folder containing the file, and enter the
commands
13
>> x = -5;
>> absval
y =
5
If your conditions have more than two possibilities (eg. a piecewise function which evaluates differently
over different intervals), then you can also make use of elseif statements.
if 1st condition
1st commands
elseif 2nd condition
2nd commands
else
3rd commands
end
You can use as many elseif statements as you like. As soon as a condition which is satisfied is found,
the corresponding command will be executed, and no further conditions will be checked. For example,
consider the following m-file (save it as checksign.m)
if x < 0
negative
elseif x > 0
positive
else
zero
end
This will check whether x is negative, positive or zero, and will output a string (the phrase inside the
) saying as much. For example,
>> x = -10;
>> checksign
ans =
negative
The condition in an if statement must be a boolean (true/false) expression. MATLAB has a number of
these, the simplest of which are detailed in the table below.
==
v=
<
<=
>
>=
equal to
not equal to
less than
less than or equal to
greater than
greater than or equal to
Tip 18: In MATLAB a word or phrase inside inverted commas ( ) is known as a string. Strings
can be used to display messages or label figures and axes, among other things. Variables can take string
values, though of course they cannot then be used in mathematical operations.
0.8 Writing Functions
0.7.4
14
Good Habits
It is often the case that MATLAB code makes perfect sense to the author at the time of writing, but
is incomprehensible to anyone else or even the author themselves later on. Here are a few guidelines to
ensure that your code can be understood by anyone reading it.
Use intuitive and descriptive names for m-files, functions (see the next section) and variables.
Indent the commands inside specific statements (see the if statements above) and loops (discussed
in a later section).
Add thorough comments to your code. Anything written after a % character on a line is a comment
and will be ignored by MATLAB. Describe what an m-file does, the meaning of all its variables,
and what is achieved by any statements or loops.
0.8
Writing Functions
The example described in the previous section (absval.m) is fine for calculating the absolute value of a
number, but every time we want to change numbers we have to redefine x, which may be undesirable.
We can avoid this by converting the program into a function, so that we can input any value we like
without having to change any variables. Open absval.m and change the first line so that the file reads
as follows, and save it.
function y = absval(x)
if x< 0
y=-x;
else
y=x;
end
The first line of a function is very important. It gives the name of the function (here, absval; this must
match the name of the m-file itself), the name of any input variables (here, x) and the name of the
output variable (here, y). This function can now be executed like any other MATLAB function (eg.
sin or log). To calculate the absolute value of 8, use the command
>> absval(-8)
ans =
8
The value 8 will be assigned to the input variable x within the function, and the value of the output
variable y will be returned by the function.
Lets look at another example. Say we want to write a function which takes two n n matrices A and
B as input, and calculates (A + 2B)2 . The m-file would look as follows.
function y = matrixsumsquare(A,B)
y = (A+2*B)^2;

To evaluate the function on matrices A =
1 3
4 0

and B =

2 5
, use the commands
1 1
Tip 19: When saving a m-file function, make sure you the m-file has exactly the same name as the name
of the function (ie. the name given in the first line of the m-file).
0.9 Loops
15
>> A = [1 3; 4 0];
>> B = [2 5; 1 1];
>> matrixsumsquare(A,B)
ans =
103
91
42
82
Note that as with any variables, A and B must be defined before matrixsumsquare can operate on them.
The only variables MATLAB can access while evaluating a function are those given as input to the
function, or those which are defined within the function. For example, consider the function
function output = example(y)
output = x + y;
If you now try to evaluate this function by running the commands
>> x = 5;
>> example(10)
You will get an error, because the variable x is neither given as input for the function example, nor
defined within example.m.
The reverse is also true variables which are solely defined within functions will not be accessible outside
of the function, with the exception of the output variable (which will be output as ans).
Note that these caveats only apply to functions; m-files which are not functions can access any variables
defined in the command window, and vice versa.
0.9
0.9.1
Loops
While Loops
In a previous section you were introduced to the if statement, where MATLAB executes a command
only when a condition holds. A while loop has the same structure as an if statement:
while condition
commands
end
The difference is that while condition holds, commands is executed, and then the loop is restarted and
condition is checked again. The first time that condition does not hold MATLAB exits the loop and
moves on to the next line after end.
Consider the following example, which uses a while loop to calculate the factorial of a natural number
n, written n!. Recall that n! = n(n 1)(n 2) 2 1. MATLAB already has a built-in factorial function
(called factorial), but we will write our own. Note that we should not call the function factorial,
since a function with that name is already defined. We instead use the name fact.
function output = fact(x)
% temp is a temporary working variable
temp=1;
while x>1
temp=temp*x;
x=x-1;
end
output=temp;
0.9 Loops
16
Here, temp acts as a partial product: it starts out at 1, we multiply it by x, and then call that value
temp. We then decrease the value of x by 1, and then multiply temp by x again. This is repeated until
x takes the value 1, and we no longer need to multiply. The functions output is then the final value of
temp.
To use our function, we simply call it with a positive integer as the argument:
>> fact(6)
ans =
720
Note that as long as condition holds and commands does not alter any of the values used in condition ,
then the loop will continue running forever and MATLAB will crash. When you write a while loop,
ensure that there will definitely be some point at which condition is false. If you do happen to execute
code containing an infinite loop, use the command Ctrl+c to stop it.
0.9.2
For Loops
So far you have seen if statements and while loops, which execute commands when certain conditions
hold. Sometimes, however, we simply want the commands to be executed a fixed number of times without
depending on any conditions. This situation is handled by a for loop, which has the following form:
for variable = vector of values
commands
end
For each of the values in the vector, variable is assigned that value and the commands are executed.
This is best illustrated with an example. Consider the for loop
for j = [1,2,3,4,5]
t(j) = j^2;
end
Note that as this code is executed, we evaluate t(1), t(2), t(3), t(4), t(5), so the output t is a 1 5 vector.
In the first iteration of the loop, j takes the value 1, and then t(1) (that is, the first entry of the vector t)
takes the value 12 . In the next iteration, j takes the value 2, and t(2) takes the value 22 . This continues
on until j=5, after which the loop is exited.
We can use a for loop to write a factorial function, instead of a while loop as seen in the previous
section. Consider the following function:
function output = fact2(x)
% temp is a temporary working variable
temp=1;
for j=2:x
temp=temp*j;
end
output=temp;
Tip 20: The command Ctrl+c will stop MATLABs current operation.
Tip 21: There is an easier way to create the row vector [1, 2, 3, 4, 5], namely the command 1:5. In
general, the command x:n:y will create the vector [x, x + n, x + 2n, x + 3n, . . . , y]; that is, the numbers
from x to y in increments of n. If the n part is left out (eg. with the command 1:5), then MATLAB
assumes n = 1.
0.10 Useful Commands for MATH1051
17
Here, we defined the vector of values which j takes as 2:x. This is the vector [2, 3, . . . , x]; so the loop
will be evaluated for j = 2, j = 3, . . . , j = x. Again, temp is acting as a partial product. In the first
iteration of the for loop, j takes the value 2, and temp is redefined as temp*2. In the next iteration, j
takes the value 3, and temp is redefined as temp*3. This is repeated until temp has been multiplied by
all the integers 2, 3, . . . , x, and it is this final value of temp which becomes the output.
Note that if either fact or fact2 had been evaluated with a negative number as input, they would both
return the value 1. This is because in both cases no iterations of the loop would be executed: in fact,
the statement x > 1 would be false from the beginning, so the while loop is never entered; and in fact2,
the vector 2:x would be empty, so the for loop is never entered. Hence the value of temp is never altered
from its initial value of 1.
0.10
Useful Commands for MATH1051
In this section we mention a few MATLAB commands which may be useful when completing assignments
and/or other tasks in MATH1051.
0.10.1
The trapz Function
The trapz function uses the trapezoid method to evaluate the integral of a function. Recall that the
trapezoid method involves dividing up the region under a function into vertical strips, finding the values
of the function where the strips meet, and using those values to calculate the area of the region.
The trapz function takes as its argument a vector, which it interprets as the set of values of a function
at the points where the vertical strips meet. By default, MATLAB assumes the strips are one unit wide.
For example, to find the area under the function f (x) = ln(x) between x = 1 and x = 100, use the
commands
>> x=1:100;
>> y=log(x);
>> trapz(y)
ans =
361.4368
To use trapz when the width of the strips is not 1, simply multiply the final answer by the width of
the strips. For example, to compute the same integral as above but with a strip width of 0.1, use the
commands
>> x=1:0.1:100;
>> y=log(x);
>> 0.1*trapz(y)
ans =
361.5162
0.10.2
fminsearch and fminbnd
The MATLAB functions fminsearch and fminbnd are used to find the minimum value of a function in
some region. fminsearch takes as input the function itself and a starting point x0. For example, to find
where the minimum value of f (x) = sin(x) near the value x0 = 0 occurs, use the command
>> fminsearch(sin(x),0)
ans =
-1.5708
0.11 Appendix
18
The answer -1.5708 is of course approximately /2, which is where a minimum of sin(x) occurs. To find
the value of the function itself at the minimum, you can evaluate the function with a separate command
(eg. sin(-1.5708)), or use the command
>> [xvalue,fvalue]=fminsearch(sin(x),0)
xvalue =
-1.5708
fvalue =
-1.0000
The function fminbnd performs a similar task to fminsearch, except instead of starting at an initial guess
x0, we provide an interval [a,b] inside which MATLAB will search for a minimum value. For example,
to find where the minimum value of f (x) = x3 + 10x2 in the interval [4, 4] occurs, use the command
>> fminbnd(x^3+10*x^2,-4,4)
ans =
-1.2543e-005
The answer 1.2543 105 is very close to 0, and it easily be verified analytically that the minimum
value of this function on the interval [4, 4] does occur at 0 (where the function takes the value 0). As
with fminsearch, to find the value of the function at this minimum, we can use the command
>> [xvalue,fvalue]=fminbnd(x^3+10*x^2,-4,4)
xvalue =
-1.2543e-005
fvalue =
1.5732e-009
0.11
0.11.1
Appendix
Common Numerical Operations & Constants
The following table lists the MATLAB commands for a number of commonly used functions which operate
on numbers, as well as some constants. Use MATLABs help command for more information.
Operation
x+y
xy
xy
x/y
xn
x
ln(x)
log10 (x)
x 10n
0.11.2
MATLAB
Code
x+y
x-y
x*y
x/y
x^n
sqrt(x)
log(x)
log10(x)
xen
Operation
ex
|x|
sin(x)
cos(x)
tan(x)
arcsin(x)
arccos(x)
arctan(x)
n!
MATLAB
Code
exp(x)
abs(x)
sin(x)
cos(x)
tan(x)
asin(x)
acos(x)
atan(x)
factorial(n)
Constant
MATLAB
Code
pi
i
exp(1)
Inf
-Inf
Vector & Matrix Operations
The following table lists the MATLAB commands for a number of commonly used operations which can
be performed on vectors and/or matrices.
0.11 Appendix
19
Operation

3
4
Entering a row vector: a = 1
4
Entering a column vector: b = 0

7
2 4
Entering a matrix: A =
0 1
Second component of vector b:
Entry (1, 2) of a matrix A:
Transpose: c = aT
Adding vectors or matrices: b + c
Dot product: b c
Cross product: b c
Norm: kak
Multiplying matrices: AB
Matrix determinant: det(A)
Inverse: A1
Vector dimension:
Matrix dimensions:
Minimum value in vector b:
Maximum value in vector b:
0.11.3
MATLAB Code
a=[1,3,4]
b=[4;0;7]
A=[2,4;0,1]
b(2)
A(1,2)
c=a
b+c
dot(b,c)
cross(b,c)
norm(a)
A*B
det(A)
inv(A)
length(a)
size(A)
min(b)
max(b)
Special Matrices
MATLAB has a number of special matrices built-in, and these can be produced with certain commands,
given in the table below.
Matrix
n n identity matrix
m n matrix of ones
m n matrix of zeros
MATLAB Code
eye(n)
ones(m,n)
zeros(m,n)
1.2 Complex Numbers
20
Complex Numbers
1.1
Notation
Let R denote the set of real numbers. The symbol means is an element of. Thus we write
xR
to mean x is an element of R. That is, x is a real number. Eg 2 R, R, 3 R etc.
1.2
Complex Numbers
Complex numbers were introduced in the 16th century to obtain roots of polynomial equations. A complex
number is of the form
z = x + iy
where x, y R and i is (formally) a symbol satisfying i2 = 1. The quantity x is called the real part of
z and y is called the imaginary part of z.
The set of all complex numbers is denoted C. Eg 3 2i C.
1.2.1
Example
The real part of 3 2i is 3 and the imaginary part is 2 (not 2i).

Complex numbers can be added and multiplied by replacing i2 everywhere with 1. For example (2i)2 =
4i2 = 4.
1.2.2
Example
Simplify (3 2i)(1 + i).
(3 2i)(1 + i) = 3 + 3i 2i 2i2
= 3+i+2=5+i
(3 2i)(1 + i)
1.2.3
Example
Suppose a, b R. Simplify (a + bi)(a bi).
3 + 3i 2i 2i
3+i+2=5+i
1.2 Complex Numbers
21
(a + bi)(a bi) = a2 abi + abi b2 i2

= a2 + b2
(a + bi)(a bi)
2 2
a abi + abi b i
a +b
If z = a + bi is a complex number, the number a bi is called the complex conjugate of z, denoted z. Eg

the complex conjugate of 3 + 2i is 3 + 2i = 3 2i.
The previous example shows that z z is always a real number.
1.2.4
Example
Simplify
3 2i
.
1i
We multiply top and bottom by the complex conjugate of the denominator. This does not change the
value of the fraction, but the new denominator is a real number.
3 2i
1i
3 2i (1 + i)
1i
(1 + i)
(3 2i)(1 + i)
12 + 12
1
(3 2i)(1 + i)
2
1
(5 + i)
2
5 1
+ i
2 2
=
=
=
=
=
We multiply top and bottom by the complex conjugate of the denominator. This does not change the value of the fraction, but the new denominator is a
real number.
3 2i
1i
=
=
=
=
=
3 2i
1i
(1 + i)
(1 + i)
(3 2i)(1 + i)
12 + 1 2
1
2
1
2
5
2
(3 2i)(1 + i)
(5 + i)
+
1
2
It is a fact that if we consider complex roots of polynomials (and count them with their correct multiplicity), then a polynomial of degree n always has n roots. For example, every quadratic has two
roots.
1.3 Polar form
1.2.5
22
Example
Find the roots of x2 + 2x + 2 = 0.
Use the quadratic formula:
1
1
(2 4 8) = (2 2i)
2
2
= 1 i.
x =
Alternatively, complete the square. Take half the coefficient of the linear term 2x, namely 1. So consider
(x + 1)2 :
x2 + 2x + 2 = (x + 1)2 + 1 = 0
(x + 1)2 = 1 = i2
x + 1 = i
x = 1 i are the roots.
Use the quadratic formula:
=
=
(2
2
1 i.
4 8) =
1
2
(2 2i)
Alternatively, complete the square. Take half the coefficient of the linear term 2x, namely 1. So consider (x + 1)2 :
x + 2x + 2
(x + 1) + 1 = 0
x + 1
1 i are the roots.
(x + 1)
1.3
1 = i
Polar form
Real numbers are often represented on the real line. A complex number z = x + iy may be represented
by a point in the complex plane, where the horizontal axis is the real axis and the vertical axis is the
imaginary axis.
We can also specify z by giving the length r and the angle in figure 1. The quantity r is called the
modulus of z, denoted |z|. It measures the distance of z from the origin. The angle is called the
argument of z. We have:
1.3 Polar form
23
Imaginary
z=x+iy
r
Real
Figure 1: A complex number can be represented in rectangular or polar form.
x =
r cos
y =
r sin
z = x + iy =
Also
p
x2 + y 2
y
tan =
if x 6= 0.
x
r = |z| =
r cos
r sin
z = x + iy
r(cos + i sin ).
r = |z|
Also
tan
1.3.1
Example
Write z = 1 + i in polar form.
r(cos + i sin ).
q
y
x
x2 + y 2
if x 6= 0.
1.4 Eulers formula
24
First find the modulus:

|z| =
1+1=
2.
y
We have for the argument tan = , so we can take
x
y
x
= arctan 1
=
4

z =
2 cos + i sin
.
4
4
= arctan
First find the modulus:
|z| =
We have for the argument tan =
y
x
1+1=
2.
, so we can take
=
=
=
1.4
arctan
x
arctan 1

2
cos
+ i sin

.
Eulers formula
Eulers formula states for any real number :

cos + i sin = ei .
(To make sense of this, one has to define the exponential function for complex arguments. This may be
done using a series.)
Thus every complex number z = x + iy can be represented in polar form
z = rei .
1.4 Eulers formula
Notes
25
2.1 Review
26
Vectors
A vector quantity has both a magnitude and a direction. Force and velocity are examples of vector
quantities.
A scalar quantity has only a magnitude (it has no direction). Time, area and temperature are
examples of scalar quantities.
2.1
Review
A vector is represented geometrically in the (x, y) plane (or in (x, y, z) space) by a directed line segment
(arrow). The direction of the arrow is the direction of the vector, and the length of the arrow is proportional to the magnitude of the vector. Only the length and direction of the arrow are significant: it can
be placed anywhere convenient in the (x, y) plane (or (x, y, z) space).
y
Figure 2: The vector of unit length at 45 to the x-axis has many representations.
(v 1 ,v 2)
y=v 2
Q=(xQ ,yQ )
v=<v 1 ,v 2>
PQ
yQ y P = v 2
xQ x P = v 1
x=v 1
(a)
P=(x P ,y P )
(b)
Figure 3: Geometric representation of a vector.
If P , Q are points, P Q denotes the vector from P to Q.
A vector v =P Q in the (x, y) plane may be represented by a pair of numbers

v1
xQ xP
v=
=
v2
yQ yP
2.4 Vector Addition
27
which is the same for all representations P Q of v. We call v1 , v2 the components of the vector v.

0
We call the vector
the zero vector. It is denoted by 0.
0
Vectors will be indicated by bold lowercase letters v, w etc. Writing by hand you may use v or ~v or ve.
2.2
Position Vectors
Let P = (xp , yp ) be a point in the (x, y) plane. The vector OP , where O is the origin, is called the
position vector of P . Obviously

xp
OP =
.
yp
2.3
Definition: Norm
For vector v =P Q,
norm (or length or magnitude ) of v, written kvk, is the distance between P and
the
v1
Q. Thus for v =
we have
v2
kvk =
q
v12 + v22
kvk =
2.4
2 + v2
v1
2
Vector Addition
We add vectors by the triangle rule.
Consider the triangle P QR with v =P Q, w =QR. Then v + w =P R; see Figure 4. In terms of

components, if

v1
w1
v=
, w=
v2
w2
then

v+w =
v 1 + w1
v 2 + w2

.
It follows from the component description that vector addition satisfies the following properties:
v+w =w+v
u + (v + w) = (u + v) + w
v+0 =0+v =v
(commutative law)
(associative law)
2.6 Unit Vectors
28
Figure 4: P Q + QR=P R
v
v2
v
v2
v1
v1
Figure 5: We can multiply the vector v by a number (scalar).
2.5
Scalar multiplication
If is a real number (called a scalar ), we define v to be the vector of norm

kvk = || kvk
in the same direction as v if > 0, and opposite direction if < 0.

v1
v1
Using similar triangles it follows that if v =
then v =
.
v2
v2

v1
If we multiply any vector v =
by zero we obtain the zero vector:
v2

0
0v =
=0
0
2.6
Unit Vectors
A unit vector is a vector of norm 1. If v 6= 0 is a vector, then
2.7 Vectors in 3 dimensions
29
1
v
kvk
1
kvk
is a unit vector in the direction of v.

In particular

i=
1
0

,
j=
0
1
determine unit vectors along the x and y axes respectively.

y
Figure 6: Unit vectors i and j in the x and y directions respectively.

For any vector v =
v1
v2

we have

v=
v1
0

+
0
v2

= v1 i + v2 j,
Hence we can decompose v into a vector v1 i along the x-axis and v2 j along the y-axis. The numbers v1
and v2 are called the components of v with respect to i and j.
2.7
Vectors in 3 dimensions
Similarly in (x, y, z)-space a vector v =P Q is represented in component form by
v1
xQ xP
v = v2 = yQ yP
v3
zQ zP
which is the same for all representations P Q of v.
v1
For v =OP = v2 the norm of the vector v is

v3
2.7 Vectors in 3 dimensions
30
z
P
v3
y
v1
kvk =
v2
q
p
ON 2 + N P 2 = v12 + v22 + v32 .
kvk =
ON 2 + N P 2 =
2 + v2 + v2 .
v1
3
2
As before, we add vectors component by component. We also
v1
v = v2 ,
w=
v3
define multiplication by a scalar . So if
w1
w2
w3
are vectors then
v 1 + w1
v + w = v 2 + w2
v 3 + w3
v1 + w1
v + w = v2 + w2
v3 + w3
v1
and v = v2 , for R.
v3
and
v1
v = v2 , for R.
v3
The unit vectors along the x, y, z axes are respectively

1
0
0
1 ,
i=
,
j=
0
0
v1
Any vector v = v2 may be expressed as
v3
v = v1 i + v2 j + v3 k.
0
k = 0 .
1
2.8 Row and column vectors
31
We call v1 , v2 , v3 the components of v in the i, j, k directions respectively.

We usually denote 2 and 3 dimensional space by R2 and R3 , respectively. Thus

x
x
2
3
y
R =
| x, y R
and R =
| x, y, z R .
y
2.8
Row and column vectors

We may write vectors using columns eg.

vectors in this course.
a
b

, or as row vectors

a b . We usually use column
2.8.1
32
Example
An albatross is flying NE at 20 km/h into a 10 km/h wind in direction E60 S. Find the speed and
direction of the bird relative to the ground.
60 o
10
20
v
45 o
Figure 7: v gives the velocity vector of the albatross.
33
We have:
z = 20 cos 45 i + 20 sin 45 j
= 10 2i + 10 2j;
w = 10 cos 60 i 10 sin 60 j
= 5i 5 3j.
So v = z + w = (10 2 + 5)i + (10 2 5 3)j

Therefore the magnitude of v is
q
kvk =
(10 2 + 5)2 + (10 2 5 3)2
q
500 + 100 2(1 3)

=
19.91.
To calculate the angle we use
tan =
10 2 5 3
10 2 + 5
16
So v = 19.9kmh1 at E16 N.
We have:
20 cos 45 i + 20 sin 45 j
10 2i + 10 2j;
10 cos 60 i 10 sin 60 j
5i 5
3j.
So v = z + w = (10 2 + 5)i + (10 2 5 3)j

Therefore the magnitude of v is
kvk
(10 2 + 5)2 + (10 2 5 3)2
500 + 100 2(1 3)
19.91.
To calculate the angle we use
So v = 19.9kmh1 at E16 N.
tan
10 2 5 3
10 2 + 5
16
2.9 Dot Product
2.9
34
Dot Product
For non zero vectors v =OP , w =OQ the angle between v and w is the angle with 0 radians
between OP and OQ at the origin O; see Figure 8.

y
v
w
Figure 8: is the angle between v and w.

The dot (or scalar or inner ) product of vectors v and w, denoted by v w, is the number given by
vw =
0,
if v or w = 0
kvk kwk cos , otherwise
where is the angle between v and w.

If v, w 6= 0 and v w = 0 then v and w are said to be orthogonal or perpendicular.
If v = v1 i + v2 j + v3 k and w = w1 i + w2 j + w3 k are two vectors, then v w is given by:
v w = v 1 w1 + v 2 w2 + v 3 w3 ,
In particular, for v R3 ,
2.9.1
kvk2 = v v = v12 + v22 + v32 .
Example
Find the angle between the vectors:
1
v = 5 ,
4
3
w= 3
3
2.9 Dot Product
35
cos =
vw
.
kvk kwk
v w = 1 3 5 3 + 4 3 = 0.
So cos = 0, so the vectors are perpendicular, i.e. = .

2
cos =
vw
kvk kwk
v w = 1 3 5 3 + 4 3 = 0.
So cos = 0, so the vectors are perpendicular, i.e. =
2.9.2
Example
If P = (2, 4, 1), Q = (1, 1, 1), R = (2, 2, 3), find the angle = P QR.
Find vectors joining Q to P , and Q to R:
QP

1
1
2
= 4 1 = 3 ,
2
1
1
2
1
3
QR = 2 1 = 1 .
3
1
2

QP = QR
1
+
9
+
4
=
14.
=

1
3
QP QR = 3 1 = 3 + 3 4 = 4
2
2
QP QR
4

cos =
= 14 ' 107 .
QP QR

QP
QR
2
1
1
3 ,
4 1 =
1
1
2
2
1
3
1 .
2 1 =
1
2
3
QP = QR = 1 + 9 + 4 = 14.

QP QR =
Find vectors joining Q to P , and Q to R:
QP QR
4
cos = =
' 107 .

14
QP QR

1
3
3
1 = 3 + 3 4 = 4
2
2
2.10 The Projection Formula
2.10
36
The Projection Formula
Fix a vector v. Given another vector w we can write w as

w = w1 + w2
where:
w1 is in the direction of v
w2 is perpendicular to v.
See Figure 9. We want to find w1 and w2 in terms of v and w.
v
w1
w2
w
x
Figure 9: w can be decomposed into a component w1 in the direction of v and a component w2 perpendicular to v.
Since w1 is in the direction of v, let
for some R.
w1 = v
Then w2 = w v.
We need to choose to make v and w2 orthogonal.
0 = w2 v = (w v) v = w v v v = w v kvk2 .
So we need
=
wv
.
kvk2
We have derived the projection formula:

w1 =
(w v)
v,
kvk2
w2 = w
(w v)
v.
kvk2
(2.1)
2.10 The Projection Formula
2.10.1
37
Example

Find the projection of w =
6
2

onto v =
6
6

.
2.11 Vectors in Rn
2.11
38
Vectors in Rn
We are familiar with vectors in two and three dimensional space, R2 and R3 . These generalize to n
dimensional space, denoted Rn . A vector v in Rn is specified by n components:
v1
v2
vi R.
v= .
..
vn
We defineaddition
of vectors and multiplication by a scalar component-wise. Thus if R is a scalar
w1
w2
and w = . is another vector then

.
.
wn
v1 + w1
v1
v2 + w2
v2
v+w =
v = . R,
= w + v,
..
.
.
.
v n + wn
vn
We define the dot (or scalar ) product between v and w as
v w = v 1 w1 + v 2 w2 + + v n wn
n
X
=
v k wk .
k=1
vw
=
=
v1 w1 + v2 w2 + + vn wn
n
X
vk wk .
k=1
The length or norm of the vector v is
vv
q
=
v12 + v22 + + vn2
kvk =
kvk
=
=
vv
q
2 + v2 + + v2
v1
n
2
Every vector v in Rn is expressible in terms of n special vectors e1 , . . . , en called coordinate vectors,
2.12 Properties of Dot Product
39
where
0
..
.

0

ei =
1 ith position.
0

..
.
0
Proof:
An arbitrary vector v Rn can be written as
v1
v1
0
v2
v = . = 0
..
..
.
vn
0
0
v2
0
..
.
0
0
..
.
0
0
vn
= v1 e1 + v2 e2 + + vn en .
An arbitrary vector v Rn can be written as
2.11.1
v1
0
0
.
.
.
0
v2
0
+
.
0
v1
v2
.
.
.
vn
v1 e1 + v2 e2 + + vn en .
+ +
0
0
.
.
.
0
vn
Notation
(1) E.g. in R3 we have

i = e1 ,
j = e2 ,
k = e3 .
(2) We have also the zero vector
0=
0
0
..
.
satisfying
0
0 + v = v + 0 = v,
0 v = 0,
for every vector v.
(3) If v 6= 0, w 6= 0, we say that vectors v, w are perpendicular (or orthogonal ) if

v w = 0.
2.12
Properties of Dot Product
If u, v and w Rn and R then

(1) u v = v u.
(2) v (w + u) = v w + v u.
(3) (v) w = (v w) = v (w).
(4) v v 0 and v v = 0 if and only if v = 0.
40
Notes
41
3.2 The effect of matrix multiplication on vectors
42
Matrices and Linear Transformations
3.1
2 2 matrices
Recall that a 2 2 array of numbers

A=
a b
c d

x
is called a 2 2 matrix (plural matrices). Suppose v =
is a 2 element column vector. Then
y
the product Av is defined to be the 2 element column vector whose entries are given by taking the dot
product of each row of A with the vector v:

a b
x
ax + by
Av =
=
.
c d
y
cx + dy
3.1.1
Example

1 2
5
Let A =
, and let v =
. Calculate Av.
3 4
6
5
6
15+26
=
35+46

17
.
=
39

Av =

1 2
3 4
Av
1
3
15+26
35+46
17
39
3.2

2
4
5
6

.
The effect of matrix multiplication on vectors
We can visualize geometrically what happens when we multiply a 2 2 matrix by a vector.

Multiplying 2 2 matrix A by i and j gives

a b
1
a
Ai =
=
c d
0
c

Aj =
a b
c d

0
1

=
b
d
and

.
Ai is the 1st column of A, and Aj is the 2nd column of A.

This will be used again many times in the course.
3.3 Composing Transformations & Matrix Multiplication
3.2.1
43
Example

Consider the matrix A =

A maps i 7
2
1
2 1
1 2

and j 7
1
2

. Thus Ai =
2
1

and Aj =
1
2

.

.
In this way 2 2 matrices can be used to manipulate position vectors in the plane. This has application
in computer graphics. We often want to transform a set of points forming a figure on the screen. Eg
rescaling the figure, or rotating it about a central point etc. Such transformations can be accomplished
by multiplying by the appropriate matrix.
It can be shown that multiplying by a 2 2 matrix in this way always maps collinear points to collinear
points. For this reason a matrix is said to represent a linear transformation.
3.3
Composing Transformations & Matrix Multiplication
Suppose we perform two linear transformations,

we
one after the other. Suppose the first transformation

a b
s t
. A
, and the second is given by the matrix A =
perform is given by the matrix B =
c d
u v

x
vector v =
is first sent to
y

sx + ty
x
s t
.
Bv =
=
ux + vy
y
u v
The resulting vector Bv is then acted on by A, giving

a b
sx + ty
sx + ty
A
=
c d
ux + vy
ux + vy

a(sx + ty) + b(ux + vy)
=
c(sx + ty) + d(ux + vy)

(as + bu)x + (at + bv)y
as + bu
=
=
(cs + du)x + (ct + dv)y
cs + du
at + bv
ct + dv

x
y

.
We define the product AB to be the 2 2 matrix

a b
s t
as + bu at + bv
AB =
=
.
c d
u v
cs + du ct + dv
The effect of performing the transformation represented by B and then the transformation represented
by A is to perform the transformation represented by AB. In symbols
A(Bv) = (AB)v.
Note that the matrix AB represents the transformation of first applying B, and then applying A. Compare
composition of functions F G.
Linear transformations in three dimensions can be accomplished using larger matrices.
3.6 Addition
3.4
44
Definition: Matrix
A rectangular array of numbers
A=
a1n
a2n
..
.
am1 am2
amn
a11
a21
..
.
a12
a22
is called an m n matrix. It is made up of m rows and n columns. Entries of the jth column may be
assembled into a vector
a1j
a2j
..
.
amj
called the jth column vector of A. Similarly the ith row vector of A is
(ai1 ai2 ain ).
Note that an m 1 matrix is a column vector and a 1 n matrix is a row vector. The number aij in the
ith row and j column of A is called the (i, j)th entry of A. For brevity we write A = (aij ).
3.4.1
Example
The 2 3 matrix A with entries aij = i j is

A=

A=
3.4.2
0 1 2
1
0 1
0
1
1
0
2
1

.

.
Example
The m n matrix A whose entries are all zero is called the zero matrix, denoted 0; eg. the zero 3 2
matrix is
0 0
0 = 0 0 .
0 0
3.5
Equality
Matrices A, B are said to be equal, written A = B, if they are the same size and aij = bij , for all i, j.
3.6
Addition
The sum of two m n matrices A and B is defined to be the m n matrix A + B with entries
(A + B)ij = aij + bij
We add matrices element wise, as for vectors.
Addition is only defined between matrices of the same size.
3.8 Matrix Multiplication
3.6.1
45
Properties
For matrices A, B, C of the same size we have

A+B = B+A
(A + B) + C = A + (B + C).
3.7
Scalar Multiplication
Let A be an m n matrix and R. We define A to be the m n matrix with entries

(A)ij = aij
for all i, j.
Write A for (1) A. Define subtraction between matrices of the same size by
A B = A + (B).
3.7.1
Example
Find A + B and A B if

A=

A+B =
1 5 3
3 2 2

A+B =
3.7.2
3.8
1
3

,
B=
5
2

,
3
2
AB =

,
AB =
5 1 0
3
1 0
9
3
9 7 3
3 0 2
7
0
3
2

.

.
Example

If A =
4 6 3
0 1 2
6
0
2 1

then A + A =
12
0
4 2

= 2A.
Matrix Multiplication
We generalize the multiplication of 2 2 matrices. Let A = (aij ) be an m n and B = (bjk ) an n r

matrix. Then AB is the m r matrix with ik entry
(ab)ik = ai1 b1k + ai2 b2k + + ain bnk :
a11
..
.
ai1
..
.
am1
a12
..
.
ai2
..
.
am2
A
(m n)
a1n
..
ain
..
.
amn
b11 b1k
b21 b2k
..
.
..
.
bn1 bnk
B
(n r)
b1r
b2r
..
.
..
.
bnr
= AB
(m r)
where (ab)ik = dot product between ith row of A and kth column of B.
AB is only defined if the number of columns of A = the number of rows of B.
3.8 Matrix Multiplication
3.8.1
46
Example
0 1
2
Let A = (1 0 1) and B = 1
13
2
5
32
Then AB is defined, but B A is not defined.
32 13
Calculate AB.
0 1
2
AB = (1 0 1) 1
2
5
= (1 0 + 0 1 + 1 2
= (2
4)
AB
3.8.2
1 1 + 0 2 + 1 5)
0
1
2
1
2
5
(1 0 1)
(1 0 + 0 1 + 1 2
(2
1 1 + 0 2 + 1 5)
4)
Example
2
A = (1 3 9) , B = 1 .
7
Calculate AB and BA.
2
(1 3 9) 1 = 2 + 3 + 63 = 68
13
7
31
2
6
18

1 3 9
= 1 3 9 .
13
7 21 63
2
1
7
31
2
(1 3 9) 1
13
7
2 + 3 + 63 = 68
31
2
1
7
31
3
13
2
1
7
6
3
21
18
9 .
63
3.9 Transposition
3.8.3
47
Properties
For matrices of appropriate size

(1) (AB)C = A(BC),
Associativity
(2) (A + B)C = AC + BC,
A(B + C) = AB + AC
Distributive Laws
Another unusual property of matrices is that AB = 0 does not imply A = 0 or B = 0. It is possible for
the product of two non-zero matrices to be zero:

1 1
1
1
0 0
=
2 2
1 1
0 0
Also, if AC = BC, or CA = CB then it is not true in general that A = B. (If AC = BC, then
AC BC = 0 and (A B)C = 0, but this does not imply A B = 0 or C = 0.)
3.9
Transposition
The transpose of an m n matrix A = (aij ) is the n m matrix AT with entries

aji = aTij
ie.
a1n
a2n
..
.
am1 am2
amn
a11
a21
..
.
a12
a22
for all i, j
am1
am2
..
.
a1n a2n
amn
a11
a12
..
.
a21
a22
So the row vectors of A become column vectors of AT and vice versa.

3.9.1
Examples of transpose

If A =
5 8 1
4
0 0

1 1
, C = (7 5 2) then find AT , B T , C T .
,B=
2
0
5 4
5 8 1
AT =
= 8 0
4
0 0
1 0

T

1 1
1 2
T
B =
=
2
0
1 0
7
.
C T =(7 5 2)TT = 5 54
5
8
1
T
0
A =
= 8
4
0
0
2
1
0
B
T

=
1
2
1
0
T
1
1
7
5 .
2
= (7 5 2)
2
0
3.9 Transposition
3.9.2
48
Properties
For matrices of appropriate size

(1) (A)T = AT
(2) (A + B)T = AT + B T
(3) (AT )T = A
(4) (AB)T = B T AT
(not AT B T !).
3.9.3
Dot product expressed as matrix multiplication
v1
v2
A column vector
v = . Rn
.
.
vn
may be interpreted as an n 1 matrix. Then the dot product (for two column vectors v and w) may be
expressed using matrix multiplication:
w1
w2
v w = v1 w1 + v2 w2 + + vn wn = (v1 v2 vn ) .
..
1n
wn
n1
= v T w.
Also
v v = v12 + v22 + + vn2 = kvk2
= v T v.
3.9 Transposition
Notes
49
4.2 Gaussian elimination
4
4.1
50
Gaussian elimination
Simultaneous equations
One of the main applications of matrices is solving simultaneous equations of the form
a11 x1 + a12 x2 + + a1n xn
a21 x1 + a22 x2 + + a2n xn
..
.
= b1
= b2
..
.
am1 x1 + am2 x2 + + amn xn = bm .

Here x1 , x2 , , xn are the variables (unknowns) and the coefficients aij and bi are given real numbers.
This set of equations can be written using matrix multiplication as
Ax = b
where x =
x1
x2
..
.
(4.1)
is a column vector of unknowns, b =
xn
vector) and A = (aij ) is the m n matrix of coefficients.
b1
b2
..
.
is a given m 1 matrix (column
bm
An efficient way of solving a system is by Gaussian elimination.
4.2
Gaussian elimination
We illustrate the process of Gaussian elimination. Given a set of equations, the idea is to perform
operations, called elementary row operations, in order to simplify the equations to the point where the
solution may be obtained easily.
Solve the following system of equations:
2x1 + 5x2 + 7x3 = 25
x1 + 2x2 + 2x3 = 9
2x1
+ 11x3 = 16
51
Swap the order of the first two equations:

x1 + 2x2 + 2x3 = 9
2x1 + 5x2 + 7x3 = 25
2x1
+ 11x3 = 16
Subtract 2 times first equation from second:
x1 + 2x2 + 2x3 = 9
x2 + 3x3 = 7
2x1
+ 11x3 = 16
Add 2 times first equation to third:
x1 + 2x2 + 2x3 = 9
x2 + 3x3 = 7
4x2 + 15x3 = 34
Subtract 4 times second equation from third:
x1 + 2x2 + 2x3 = 9
x2 + 3x3 = 7
3x3 = 6
Multiply third equation by 1/3:
x1 + 2x2 + 2x3 = 9
x2 + 3x3 = 7
x3 = 2
Now we can solve by back substitution:
Third equation gives x3 = 2.
Substitute x3 into second: x2 + 6 = 7 so x2 = 1.

Substitute x2 , x3 into first: x1 + 2 + 4 = 9 so x1 = 3.
Swap the order of the first two equations:
x1
2x1
2x1
+
+
2x2
5x2
+
+
+
2x3
7x3
11x3
=
=
=
9
25
16
x1
2x2
x2
+
+
+
2x3
3x3
11x3
=
=
=
9
7
16
Subtract 2 times first equation from second:
2x1
Add 2 times first equation to third:
x1
2x2
x2
4x2
+
+
+
2x3
3x3
15x3
=
=
=
9
7
34
Subtract 4 times second equation from third:

x1
2x2
x2
+
+
2x3
3x3
3x3
=
=
=
9
7
6
x1
2x2
x2
+
+
2x3
3x3
x3
=
=
=
9
7
2
Multiply third equation by 1/3:
Now we can solve by back substitution:
Third equation gives x3 = 2.

Substitute x3 into second: x2 + 6 = 7 so x2 = 1.
Substitute x2 , x3 into first: x1 + 2 + 4 = 9 so x1 = 3.
52
This process is called Gaussian elimination. We call this simplified set of equations
x1 + 2x2 + 2x3 = 9
x2 + 3x3 = 7
x3 = 2
the reduced form of the equations. Once in reduced form the set of equations is easily solved by back
substitution.
In this example we repeatedly performed the following operations:
(1) Interchange the order of two equations: swap the ith and j equations.
(2) Multiply an equation through by a non-zero number .
(3) Add times the jth equation to the ith equation.
Performing these operations does not change the solution of the equations. But eventually we obtained
such a simple system that we could easily find the solution.
4.2.1
Gaussian elimination using matrices
The above procedure may be described most conveniently in terms of matrices. The information present
in a set of simultaneous equations
A x = b
mn n1
m1
may be summarized by constructing a new matrix

(A | b)
A with one extra column, b. This is called the augmented matrix. Instead of manipulating equations we
manipulate the corresponding augmented matrix. This way we keep writing to a minimum.
The augmented matrix for the set of equations in the previous example 4.2 may be written
25
9
16
2 5 7
1 2 2
2 0 11
2
1
2
5
2
0
7
2
11
25
9
16
Always keep in mind that each row of the augmented matrix represents an equation. The last column is
the RHS. The other columns are the coefficients of the variables. Eg the last row above corresponds to
2x1 + 0x2 + 11x3 = 16.
Another advantage of the augmented matrix notation is that if we want to solve two sets of equations
with the same coefficient matrix but different RHS:
Ax = b,
and Ay = c,
we can do both at once by forming the augmented matrix

(A | b | c).
See later.
4.2.2
53
Elementary row operations (EROs)
We rewrite the three allowable operations on equations discussed above, in terms of their effect on
the augmented matrix. Call the row vectors of the augmented matrix r 1 , r 2 , . . . etc. Since each row
corresponds to an equation, the operations are:
(1) Interchange rows i and j, written r i r j .
(2) Replace r i with r i , R, 6= 0, written r i
r i .
(3) Replace r i with r i + r j , R, where r j is another row of A, written r i
r i + r j .
These three operations are known as elementary row operations (EROs). In general we say that m n
matrices A, B are row equivalent, written A B, if and only if they are related by a series of EROs.
To solve a system of equations:
Set up the augmented matrix (A | b) representing the system.
Apply EROs to (A | b) to reduce it to a simpler matrix.
Eventually the system is so simple that the solution may be read off.
This is one of the most efficient procedures for solving linear equations.
4.2.3
Example
Redo the previous example 4.2, using row reductions. You should check that we are doing exactly the
same calculations as before, just with less writing.
54
The augmented matrix is:
25
9
16
2 5 7
1 2 2
2 0 11
Swap the first two rows: r 1 r 2 .
1 2 2
2 5 7
2 0 11
9
25
16
r 2 2r 1 .
1 2 2
0 1 3
2 0 11
9
7
16
Subtract 2 times row 1 from row 2: r 2
Add 2 times row 1 to row 3: r 3
r 3 + 2r 1 .
Multiply row 3 by 1/3: r 3
1 2 2
0 1 3
0 4 15
9
7
34
r 3 4r 2 .
1 2 2
0 1 3
0 0 3
9
7
6
1
3 r3 .
1 2 2
0 1 3
0 0 1
9
7
2
This is called a Gauss reduced form, or row-echelon form. The solution can now be read off directly by
back substitution as before.
r 3 gives x3 = 2.
Substitute x3 into r 2 : x2 + 6 = 7 so x2 = 1.
Substitute x2 , x3 into r 1 : x1 + 2 + 4 = 9 so x1 = 3.
2
1
2
5
2
0

7
2
11
25
9
16
Swap the first two rows: r 1 r 2 .
2
5
0
2
7
11
9
25
16
1
0
2
2
1
0
2
3
11
9
7
16
r 2 2r 1 .
Add 2 times row 1 to row 3: r 3
1
2
2
r 3 + 2r 1 .
1
0
0
Multiply row 3 by 1/3: r 3
2
1
4
2
3
15
9
7
34
r 3 4r 2 .
1
0
0
2
1
0
2
3
3
9
7
6
1
0
0
2
1
0
2
3
1
9
7
2
1r .
3 3
This is called a Gauss reduced form, or row-echelon form. The solution can now be read off directly by back substitution as before.
r 3 gives x3 = 2.
4.2.4
55
The Gaussian algorithm
To solve
A
x = b ,
mn n1
m1
first set up the augmented matrix

(A | b) .
Then
- Find a row with entry 6= 0 in column 1. (If none go to column 2)
- Multiply row by
to make leading entry 1
- Swap rows to bring the 1 to the top
..
.
- Subtract appropriate multiples of row 1 to clear out the first column
1
0
..
.
First column is now done.
Repeat on second column:
Find a row (not the first) with entry 6= 0 in column 2. (If none exists, go on to third column). Make
the leading entry 1, and clear out all the entries below this 1:
1

0
1
..
.
Then go on to the third column etc.
4.2.5
Gauss reduced form or row echelon form
When solving a system Ax = b, we can stop doing EROs on (A | b) once A is simple enough to find
the solution by back substitution. This means we reduce (A | b) until:
All rows consisting entirely of 0s are at the bottom.
The first non-zero entry (leading entry or pivot) of each row is 1, and appears in a column to the
right of the leading entry of the row above it.
All entries in a column below a leading entry are zeros.
56
The reduced matrix is called the Gauss reduced form (or row echelon form) of A. For example:
1
0
0
0
1
0
0
0
0
1
0
The leading entries 1 (pivots) must be 1 and must have zeros below them. Not every column must have
a pivot. Above, the third column has no pivot.
The leading entries must form a stair case from top left to bottom right.
(Note that a matrix A may have more than one Gauss reduced form. Eg the entries marked above are
not unique. It doesnt matter which Gauss reduced form is used.) Tips:
Try to do row reductions that lead to the easiest algebra.
Try to avoid fractions.
Use row swaps to take advantage of existing 1s in the matrix.
We will look at the Gaussian Algorithm in a MATLAB session.
4.2.6
Rank of a Matrix
The number of pivots 1 in the row echelon form of a matrix A is called its rank, denoted rank A. The
matrix above has rank 3.
4.2.7
Example
Which of the following matrices are in Gauss reduced form (row echelon form)? For those in row echelon
form, determine the rank.
1 2 3 4
A = 0 0 1 4 ,
0 0 0 1
0 1 0 2
B = 0 0 1 2 ,
0 0 0 0
0 1 0 2
C= 1 0 0 0
0 0 1 0
A, B are row-reduced. C is not because the leading 1 in the second column is not to the right of the
leading 1 in the 1st column.
rank A = 3. rank B = 2. A, B are row-reduced. C is not because the leading 1 in the second column is not to the right of the leading 1 in
the 1st column.
rank A = 3. rank B = 2.
4.2.8
Inconsistent Equations
Not every system of linear equations has a solution. Consider the following example.
Solve
x1 + 2x2 x3 = 4
2x1 + 7x2 + x3 = 14
3x1 + 8x2 x3 = 17
57
The augmented matrix is
1 2 1
2 7
1
3 8 1
4
4
1 2 1
3 6 (r 2
14 0 3
r 2 2r 1 )
17
0 2
2 5
(r 3
r 3 3r 1 )
1 2 1 4
1
0 1
1 2 (r 2
3 r2 )
0 2
2 5
1 2 1 4
1 2
0 1
0 0
0 1
r 3 2r 2 )
(r 3
This is in row echelon form. The bottom line corresponds to:

0 x1 + 0 x2 + 0 x3 = 1,
which is impossible. Therefore there are no solutions to the system.
1
2
3
2
7
8
1
1
1
4
14
17
1
0
0
2
3
2
1
3
2
4
6
5
1
0
0
2
1
2
1
1
2
4
2
5
2
1
0
1
1
0
4
2
1
1
0
0
The augmented matrix is
(r 2
(r 3
r 2 2r 1 )
r 3 3r 1 )
(r 2
1r )
3 2
(r 3
r 3 2r 2 )
This is in row echelon form. The bottom line corresponds to:

0 x1 + 0 x2 + 0 x3 = 1,
which is impossible. Therefore there are no solutions to the system.
4.2.9
How to recognize an inconsistent system
The Gauss reduced form of the augmented matrix will have one or more rows of the form (0 0 . . . 0 | a)
where a 6= 0. This leads to an impossible equation 0x1 + 0x2 + + 0xn = a 6= 0.
Geometrically, a linear equation in R2 represents a line. An equation in R3 represents a plane etc. Solving
two equations in R2 thus corresponds to finding the intersection of two lines in R2 . Usually two lines in
R2 have a unique point of intersection. But not always: parallel lines do not.
In fact parallel lines do not intersect at all, so there are no solutionsunless the lines are right on top of
each other, in which case they intersect in infinitely many points.
Solving three equations in R3 corresponds to finding the intersection of three planes. Usually three
planes have a unique point of intersection, but not always. Similarly for equations in more variables.
This underlies the next two examples.
58
Three planes meeting in a point
Three planes meeting in a line
4.2.10
More than one solution
A system of equations may have more than one solution.

Solve the system of equations
x1 + 3x2 2x3 = 5
x1 + x2 2x3 = 9
2x1 + 4x2 2x3 = 12
59
1
1
2
3 2
1 2
4 2
1
1
r2 ) 0
4
0
(r 2
5
1
9 0
12
0
3
1
2
2
1
2
3
4
2
5
4
2
2
4
2
5
1
1 0
2
0
2
1
0
3
1
0
(r 2
(r 3
r2 + r1 )
r 3 2r 1 )
5
1 (r 3
0
r 3 +2r 2 )
(Gauss reduced form).

The last row is equivalent to
0 x1 + 0 x2 + 0 x3 = 0,
which gives no information.
Therefore we effectively have just 2 equations in 3 unknowns

x1 + 3x2 2x3
x2 x3
= 1
No leading 1 in the third column (non-pivotal). So set x3 = arbitrary, and back substitute into r 2 and r 1 to give
x2
= 1 + x3 = 1 + ,
x1
5 3x2 + 2x3 = 8 3 + 2 = 8
1
1
2
(r 2
3 2
1 2
4 2
1
1
r2 ) 0
4
0
1
5
9 0
12
0
3
1
2
2
1
2
3
4
2
5
4
2
2
4
2
5
1
1 0
0
2
3
1
0
2
1
0
(r 2
(r 3
r2 + r1 )
r 3 2r 1 )
5
1 (r 3
0
r 3 +2r 2 )

The last row is equivalent to
0 x1 + 0 x2 + 0 x3 = 0,
which gives no information.
Therefore we effectively have just 2 equations in 3 unknowns

x1 + 3x2 2x3
x2 x3
= 1
No leading 1 in the third column (non-pivotal). So set x3 = arbitrary, and back substitute into r 2 and r 1 to give
x2
= 1 + x3 = 1 + ,
x1
5 3x2 + 2x3 = 8 3 + 2 = 8
4.3 The general solution
60
This means, for any choice of R, a solution is:
x1
8
8
1
x = x2 = 1 + = 1 + 1 .
x3
0
1
8
The last equation is the equation of a line in R3 : x is given by a position vector p = 1 plus
0
1
times a vector v = 1 indicating direction.
1
Thus geometrically the three planes intersect along a common line rather than at a single point. Every
solution x corresponds to a point on this line obtained by a particular choice of . This means, for any choice
of R, a solution is:
8
8
1
x1
1 .
x = x2 = 1 + = 1 +
0
1
x3
8
1
1 plus times a vector v =
1 indicating direction.
0
1
Thus geometrically the three planes intersect along a common line rather than at a single point. Every solution x corresponds to a point on this line
The last equation is the equation of a line in R3 : x is given by a position vector p =
obtained by a particular choice of .
4.2.11
How to recognize non-unique solutions of Ax = b
The system is consistent, and the Gauss reduced matrix will have one or more columns without a pivot
(leading entry). That is, the rank of A is less than the number of columns of A.
For example, if the second column does not have a leading 1 then there will never be an equation we can
use to isolate x2 , so x2 will be a free variable: we will be able to assign it any value ( say), and still
solve the system. In a big system we may have several free variables.
Any system with more than one solution will have a free variable. Since this variable can take any value,
the system will then have infinitely many solutions.
4.3
4.3.1
The general solution

General form of solutions
In general a system of equations

A
x = b
mn n1
m1
will have one of the following

A unique solution
An infinite number of solutions
No solution.
Which behaviour occurs can be seen from the Gauss reduced form of the matrix.
We shall show that if there are an infinite number of solutions, they will all be of the form p + y where
p is a particular solution and y is the part arising from the free variables. We shall show that y is a
solution of Ax = 0.
It is useful to introduce some notation.
4.3 The general solution
4.3.2
61
The Nullspace of a matrix
Let A be a matrix. The set of vectors x with Ax = 0 is called the nullspace of A, denoted NS(A).
The nullspace always contains the zero vector, since A0 = 0. We shall show later that the nullspace
either contains only the zero vector, or else contains infinitely many vectors.
The linear system Ax = 0 is called a homogeneous system. (Homogeneous means all the same.) Thus
the nullspace of A is the set of solutions of the homogeneous system.
A linear system Ax = b with b 6= 0 is called inhomogeneous.
4.3.3
Example
1 3 2
5
Let A = 1 1 2 and let b = 9 .
2 4 2
12
(a)
Calculate NS(A).
Solve the system Ax = b.
(b)
For the nullspace we have
1
1
2
(r 2
to solve Ax = 0.
3 2 0
1 3 2 0
1 2 0 0 4 4 0
r2 + r1 )
(r 2
4 2 0
0 2 2 0
(r 3
r 3 2r 1 )
1 3 2 0
1 3 2 0
r2
r 3 +2r 2 )
) 0 1 1 0 0 1 1 0 (r 3
4
0 2 2 0
0 0 0 0

The bottom row gives no information. So x3 can be anything. Say x3 = . The middle row says
x2 x3 = 0 so x2 = . The top row gives x1 + 3x2 2x3 = 0, so x1 + 3 2 = 0 so x1 = .

NS(A) = 1 | R .
1
For the nullspace we have to solve Ax = 0.
1
1
2
(r 2
3
1
4
1
) 0
4
0
r2
2
2
2
3
1
2
0
1
0 0
0
0
2
1
2
0
0
0
3
4
2
0
0
0
2
4
2
1
0
0
3
1
0
2
1
0
(r 2
(r 3
r2 + r1 )
r 3 2r 1 )
0
0
0
(r 3
r 3 +2r 2 )

The bottom row gives no information. So x3 can be anything. Say x3 = . The middle row says x2 x3 = 0 so x2 = . The top row gives x1 +3x2 2x3 = 0,
so x1 + 3 2 = 0 so x1 = .

1 |R .
NS(A) =
4.4 Gauss-Jordan elimination
62
Inhomogeneous case Ax = b. We did this in the previous example.

We found
x1
8
8
1
x = x2 = 1 + = 1 + 1 = p + y,
x3
0
1
1 is a particular solution of Ax = b and y is any element in the nullspace NS(A).

where p =
0
Inhomogeneous case Ax = b. We did this in the previous example.
We found
1
8
8
x1
1 = p + y,
x = x2 = 1 + = 1 +
1
0
x3
8
where p = 1 is a particular solution of Ax = b and y is any element in the nullspace NS(A).
0
4.3.4
Relation between homogeneous & inhomogeneous systems
The solution of the homogeneous Ax = 0 and inhomogeneous Ax = b systems are related.

Let A be a matrix, and consider the system
() Ax = b.
Let p be a fixed solution of (). Then the complete set of solutions of () is
{x = p + y | y NS(A)}.
Proof: If p is a solution and y NS(A) then A(p + y) = Ap + Ay = b + 0 = b. So every x of the form
p + y is a solution.
Conversely, if x is a solution of () then Ax = b = Ap. Thus A(x p) = Ax Ap = b b = 0. So let
y = x p NS(A). Then x = p + y has the stated form.
4.3.5
Summary
There are 3 possible behaviours for a system:

1. The system Ax = b may have no solutions. (Inconsistent case.)
2. If a solution to Ax = b exists and NS(A) = {0} then Ax = b has a unique solution.
3. If Ax = b has a solution x = p and NS(A) 6= {0} then there are infinitely many solutions: all x of
the form x = p + y with y NS(A).
Note that a homogeneous system is always consistent (case (1) cannot apply) because the 0 vector is
always a solution.
4.4
Gauss-Jordan elimination
Gauss-Jordan (or simply Jordan) reduction involves performing more EROs on the Gauss reduced form
of a matrix until it has the additional property that each leading 1 is the only non-zero entry in its
column.
Solve
2x1 + 5x2 + 7x3 = 25
x1 + 2x2 + 2x3 = 9
2x1
+ 11x3 = 16
4.4.1
63
Example of Gauss-Jordan Reduction
This is example 4.2.3 again. The augmented matrix is:
2 5 7 25
1 2 2
1 2 2
9 0 1 3
2 0 11 16
0 0 1
This is the Gauss reduced form. But we can
1 2 0
0 1 0
0 0 1
r1
r2
continue doing row
5
1 0
0 1
1
2
0 0
r 1 2r 3
r 2 3r 3
r1
9
7
2
operations.
0 3
0 1
1 2
r 1 2r 2
This is the Jordan reduced form. We can read off the solution directly: x1 = 3, x2 = 1, x3 = 2.
4.4.2
The difference between Gauss and Gauss-Jordan reduction
Gauss reduction leads to a matrix with 0s below the leading 1s, such as
1
0
0
0
1
0
0
0
0
1
0
while Gauss-Jordan reduction results in a matrix with 0s above and below the leading 1s such as
1
0
0
0
0
1
0
0
0
1
0
0
0
0
.
1
0
Jordan form has the advantage that the solution of the original system can be read off directly without
any back substitution. However Gaussian reduction is often preferred for solving systems.
Gauss-Jordan reduction is better if we want to solve many systems Ax = b, Ax = c etc with the same
coefficient matrix A, since it eliminates the need to do multiple back substitution steps. This is useful in
the next chapter.
Notes
64
5.1 The Identity Matrix
65
Inverses
For the rest of the course, the matrices will be square.
5.1
The Identity Matrix
An n n matrix (m = n) is called a square matrix of order n. The diagonal containing the entries
a11 , a22 , , ann
is called the main diagonal (or principal diagonal ) of A. If the entries above this diagonal are all zero then
A is called lower triangular. If all the entries below the diagonal are zero, A is called upper triangular
a11 0
0
0
a11
a1n
..
..
. a22 0
0 a22
0
.
.
.
.
.
.
.
.
.
.
.
.
.
0
..
..
..
..
..
..
..
.
.
.
.
.
0
.
.
an1
ann
0
0 0 ann
lower triangular
upper triangular
If elements above and below the principal diagonal are zero, so
aij = 0 ,
i 6= j
then A is called a diagonal matrix. Note that if A is diagonal then A = AT .

5.1.1
Example
1
0 0
A = 0 3 0 is a diagonal 3 3 matrix.
0
0 2
5.1.2
Definition: Identity matrix
The n n identity matrix I = In , is the diagonal matrix whose entries are all 1:
1 0 0 0
0 1 0 0
I = 0 0 1 0 .
..
.
.
. . ..
.
0 0 1
5.1.3
Comments on the identity matrix
Recall from linear transformations of R2 : Ai is the first column of A and Aj is the second column.
For example, for the 2 2 case we have

a11 a12
1 0
a11 a12
AI =
=
= A = IA.
a21 a22
0 1
a21 a22
5.2 Definition: Inverse
66
In general the jth column vector of I is the coordinate vector

0
..
.

0

ej =
1 j.
0

..
.
0
Thus
0
.
..

0
j
1

0

..
.
0
Aej
a1j
a2j
..
.
a1n
a2n
..
.
an1
anj
ann
a11
a21
..
.
a1j
a2j
..
.
anj
jth column
= vector of
A.
So AI = A for all A.
Replacing A by AT , AT I = AT also. Transpose both sides: A = (AT )T = (AT I)T = I T (AT )T = IA, so
IA = A for all A.
We have proved that for all square matrices A
IA = AI = A.
5.2
Definition: Inverse
Let I denote the identity matrix.

A square matrix A is invertible (or non-singular ) if there exists a matrix B such that
AB = BA = I.
Then B is called the inverse of A and is denoted A1 . A matrix that is not invertible is also said to be
singular.
5.2.1
Inverse for the 2 2 case

Let A be a 2 2 matrix A =

AB =
a b
c d
a b
c d

d b
. Set B =
. Then
c a

d b
c a

=
ad bc
0
0
ad bc

= BA.
67
Let
= ad bc.
If 6= 0, then
1
1
B=
A1 =
d b
c
a
We call the determinant of A. We will consider determinants in more detail later (chapter 6). The
problem of obtaining inverses of larger square matrices will also be considered later (section 5.2.5).
5.2.2
Example

Find the inverse matrix of A =

. Check your answer.
1 2
3 5
= 1 5 2 3 = 1.
A
1
=
1
5 2
3
1

=
5
2
3 1
Check whether AA1 = A1 A = I:

Multiplying the two matrices A and A1 gives

1 0
5
2
1 2
,
=
0 1
3 1
3 5

1 0
1 2
5
2
=
0 1
3 5
3 1
So A is invertible with inverse
5
2
3 1

.
= 1 5 2 3 = 1.
A
1
1
2
1
5
3

=
5
3
2
1

.
Check whether AA1 = A1 A = I:

Multiplying the two matrices A and A1 gives

1
3
5
3

2
5
2
1
So A is invertible with inverse
5.2.3
Example

Show that A =
1 2
3 6
5
3

is not invertible.
2
1
2
5
1
3
5
3
1
0
0
1
1
0
0
1
=
=
2
1

.
68

Suppose A has inverse

a b
.
c d

1 2
a b
1 0
Then
=
3 6
c d
0 1
a + 2c = 1
b + 2d = 0
3a + 6c = 0
3b + 6d = 1
(1)
(2)
.
(3)
(4)
Hence 3 (1) (3) 0 = 3. This is impossible, so A cant have an inverse.

Note that = 6 2 3 = 0 in this case.
Later we shall
see that a square matrix is invertible exactly when its determinant is not 0.

pose A has inverse
a
c

1
Then
3

b
1
=
d
0

2
a
6
c
b
d
a + 2c = 1
b + 2d = 0
3a
+ 6c = 0
3b + 6d = 1

0
(1)
(2)
.
(3)
(4)
Hence 3 (1) (3) 0 = 3. This is impossible, so A cant have an inverse.
Note that = 6 2 3 = 0 in this case.
Later we shall see that a square matrix is invertible exactly when its determinant is not 0.
5.2.4
Example: Inverse of a 3 3 Matrix
2 3 1
16 17
7
5 . Show that B = A1 .
Let A = 1 2 3 and B = 11 12
2
2 5
2
2 1
Just multiply the two matrices together.
2 3 1
16 17
7
1 0 0
5 = 0 1 0 .
AB = 1 2 3 11 12
2
2 5
2
2 1
0 0 1
And, similarly
16 17
7
2 3 1
1 0 0
5 1 2 3 = 0 1 0 .
BA = 11 12
2
2 1
2
2 5
0 0 1
Just multiply the two matrices together.
AB =
2
1
2
3
2
2
16
11
2
17
12
2
1
16
3 11
5
2
17
12
2
7
1
5 = 0
1
0
0
1
0
0
0 .
1
3
2
2
1
1
3 = 0
5
0
0
1
0
0
0 .
1
And, similarly
BA =
5.2.5
7
2
5
1
1
2
Inverses using GaussJordan reduction
We can now give a systematic way of finding inverses of matrices, using GaussJordan reduction.
Sup-
69
Let A be an invertible n n matrix. We want to find the inverse X of A. We try to solve

AX = I.
5.2.6
Example
2 3
1
x1 y1 z1
Above A = 1 2 3 . We want a matrix X = x2 y2 z2 with AX = I.
2
2
5
x3 y3 z3

x1
y1
z1
1
0
0
Let x = x2 , y = y2 , z = z2 . We need Ax = 0 , Ay = 1 , Az = 0 .
x3
y3
z3
0
0
1
So Ax = i,
Ay = j,
Az = k. We can solve these 3 systems by augmenting A with all 3 RHS
vectors: (A | i j k). That is, we form (A | I). We use Gauss-Jordan reduction to solve the 3 systems at
once and find x, y, z.
2 3 1
Let A = 1 2 3 . Calculate A1 . (This is example 5.2.4).
2
2 5
70
Perform Gauss-Jordan reduction on the augmented matrix (A | I):
2 3 1 1 0 0
1 2 3 0 1 0
1 2 3 0 1 0 2 3 1 1 0 0
2
2 5 0 0 1
2
2 5 0 0 1
r 1 r 2
1 0
7 2 3 0
5 1 2 0
0 1
0 0 1 2 2 1
r1
r 1 + 2r 2
r3
r 3 + 2r 2
16 17
7
1 0 0
11 12
5
0 1 0
0 0 1 2
2 1
r1
r 1 7r 3
r2
r 2 5r 3
1 2 3 0
1 0
0
1
5 1 2 0
0 2 11 0
2 1
r 2 2r 1
r2
r3
r 3 + 2r 1
2 3
0
1 0 7
0 1 5
1 2
0
0 0 1 2
2 1
r 3
r3
So A is invertible, and A1
16 17
7
5 .
= 11 12
2
2 1
Compare this to 5.2.4.

Perform Gauss-Jordan reduction on the augmented matrix (A | I):
1
0
0
1
0
0
So A is invertible, and A1 =
16
11
2
3
2
2
2
1
2
17
12
2
2
1
2
0
1
0
1
3
5
1
0
0
0
1
0
3
0
1
1
2
5
11
0
2
r2
r 2 2r 1
r3
r 3 + 2r 1
7
5
1
r3
2
3
1
2
2
2
r 3
0
0
1
0
0
1
0
0
1
2
3
0
3
1
1
0
2
5
r 1 r 2
1
2
2
1
0
0
1
0
0
1
0
0
0
7
2
3
1
2
1
5
0
1
2
2
r1
r 1 + 2r 2
r3
r 3 + 2r 2
0
1
0
0
0
1
r1
r2
16
17
11
12
2
2
r 1 7r 3
r 2 5r 3
0
0
1
0
0
1
7
5
1
7
5 .
1
Compare this to 5.2.4.
In general, let xj be the jth column of the matrix X, and recall that ej is the jth column of I. Breaking
the previous matrix equation AX = I into one equation for each column gives rise to n equations
Ax1 = e1 , Ax2 = e2 , . . . , Axn = en .
(5.1)
We can regard each of these as a system to be solved. As we noted when introducing the augmented
matrix, we can solve more than one system of equations at once by forming a larger augmented matrix.
Eg
Ax = b, Ax = c 99K (A | b c)
Thus we may solve (5.1) by applying GaussJordan elimination to the augmented matrix
(A | e1 e2 , . . . , en ) = (A | I).
Reduce until we reach
(I | X).
The solution X can now be read off, so X = A1 is the inverse of A. If this reduction cannot be
completed, A has no inverse.
5.3 Inverses via GaussJordan. Summary
5.3
71
Inverses via GaussJordan. Summary
If A is n n, to find A1 :
Set up augmented n 2n matrix (A | I).
If this can be reduced to (I | X) then X = A1 .
If it cannot be reduced, then A is not invertible.
5.3.1
Invertibility and the nullspace
Above to find A1 we found X with AX = I. However for X to be the inverse of A, the definition of
the inverse matrix requires XA = I also. We shall show the surprising fact that if AX = I then XA = I
automatically, for square matrices.
We need a little theory. Let A be an n n matrix. We shall find several properties equivalent to A being
invertible.
Consider the following statements:
1. A is invertible.
2. NS(A) = {0}.
3. For every n 1 vector b the system Ax = b is solvable, with a unique solution x.
4. There exists an n n matrix B with AB = I.
We first show that (1) = (2) = (3) = (4):
Proof: (1) = (2) Suppose x NS(A). Thus Ax = 0. Multiply both sides by A1 to see that x = 0,
so NS(A) = {0}.
(2) = (3) If NS(A) = {0} then when we solve
variables. Row reducing must give
1
0
(A | 0) 0
..
.
Ax = 0 the only solution is x = 0, so there are no free

1
0 1
.. ..
..
. .
.
0 0 0
0
0
0
,
.. ..
. .
1 0
with a leading entry in every column.

Now consider solving Ax = b. The row reductions on A are
changes. So we get
1
0 1
(A | b) 0 0 1
.. .. ..
..
..
. . .
.
.
0 0 0
the same. Only the augmented column
..
.
This will have also have a unique solution x, by back substitution.

(3) = (4) We can solve Ab1 = e1 , Ab2 = e2 , . . . where e1 , e2 , . . . are the standard coordinate vectors.
Let B be the n n matrix with columns b1 , b2 , . . . The first column of AB will be Ab1 = e1 . The second
column will be Ab2 = e2 etc. Thus AB = I.
72
It is not surprising that (1) = (4): (4) is the statement AB = I, whereas (1) is the statement AB = I
and BA = I. Amazingly, once AB = I, we shall show that BA is forced to be I (for square matrices);
see below. This means (4) = (1) also.
FACT:
If C and D are n n matrices and CD = I then DC = I also.
Consider the nullspace of D. Let x be any element in NS(D), so Dx = 0. Multiply both sides by C
on the left to get (CD)x = C(Dx) = C0 = 0. But CD = I (given), so Ix = 0 so x = 0. That is
NS(D) = {0}.
Thus property (2) holds for D. We have seen that (2) = (4), so (4) holds for D. That is, there exists
an n n matrix B with DB = I.
So C = CI = C(DB) = (CD)B = IB = B, that is, B = C. So DC = DB = I. Consider the nullspace of D. Let x
be any element in NS(D), so Dx = 0. Multiply both sides by C on the left to get (CD)x = C(Dx) = C0 = 0. But CD = I (given), so Ix = 0 so x = 0.
That is NS(D) = {0}.
Thus property (2) holds for D. We have seen that (2) = (4), so (4) holds for D. That is, there exists an n n matrix B with DB = I.
So C = CI = C(DB) = (CD)B = IB = B, that is, B = C. So DC = DB = I.
We have proved:
5.3.2
Properties equivalent to invertibility
Let A be an n n matrix. The following statements are all equivalent:

1. A is invertible.
73
2. NS(A) = {0}.
3. For every n 1 vector b the system Ax = b is solvable, with a unique solution x.
4. There exists an n n matrix B with AB = I.
If any one of these statements holds, the other three also hold. If any one fails the other three also fail.
5.3.3
Special case: n equations in n unknowns
Consider the system Ax = b. In the special case that A is n n and invertible, we may solve this system
by multiplication on the left by A1 to give
1
1
1
A
| {z A} x = A b unique solution x = A b.
I
This is another demonstration that (1) = (3) above:

If A is invertible, the system Ax = b has a unique solution for each b.
This is often inaccurately remembered as n equations in n unknowns have a unique solution. This
statement is only true if A is invertible.
The method of multiplying by A1 only works if A is square and A is invertible. It is also more work to
find A1 than to solve the system using Gaussian elimination. In general, use Gaussian elimination to
solve systems.
5.3.4
Properties of inverses
(1) A has at most one inverse

(2) If A, B are invertible so is AB, and
(AB)1 = B 1 A1
(not A1 B 1 !)
(3) A is invertible if and only if AT is invertible

(4) If A is invertible then (AT )1 = (A1 )T .
5.3.5
Proof
(1) Suppose B, C are both inverses of A, so

AB = BA = I
AC = CA = I.
Then B = BI = B(AC) = (BA)C = IC = C.

Therefore the inverse, if it exists, is unique.
(2) Suppose A, B have inverses A1 , B 1 respectively. We have to show that AB has inverse B 1 A1 .
So we just multiply AB by B 1 A1 , and check that we get the identity matrix I:
(AB)B 1 A1 = A(BB 1 )A1
= AIA1 = AA1 = I
74
and similarly
(B 1 A1 )(AB) = I.
Therefore AB is invertible with inverse
(AB)1 = B 1 A1 .
(3) and (4) are an exercise:
We prove (4). Assume that A is invertible. We have to show that the inverse of AT is (A1 )T . To show
that one matrix is the inverse of the other, we just multiply them together and check we get the identity.
We will need the fact that (AB)T = B T AT . Setting B = A1 in this, we have
(A1 )T AT = (AA1 )T = I T = I.
Similarly
AT (A1 )T = (A1 A)T = I T = I.
Therefore if A is invertible then AT is invertible with inverse
(AT )1 = (A1 )T
This also proves half of (3), namely A invertible implies A1 invertible. To prove the converse, just
replace A with AT in the above calculation. This concludes the proof of both (3) and (4). We prove (4).
Assume that A is invertible. We have to show that the inverse of AT is (A1 )T . To show that one matrix is the inverse of the other, we just multiply
them together and check we get the identity. We will need the fact that (AB)T = B T AT . Setting B = A1 in this, we have
(A
1 T
) A
= (AA
1 T
=I
= I.
=I
= I.
Similarly
T
1 T
A (A
Therefore if A is invertible then AT is invertible with inverse
= (A
T 1
(A )
A)
= (A
1 T
This also proves half of (3), namely A invertible implies A1 invertible. To prove the converse, just replace A with AT in the above calculation. This
concludes the proof of both (3) and (4).
Notes
75
6.1 Definition: Determinant
76
Determinants
6.1
Definition: Determinant
Associated to each square matrix A is a number called the determinant of A and denoted |A| or det(A).
It is defined as follows.
If A is a 1 1 matrix, say A = (a) then |A| is defined to be a.
If A = (aij ) is 2 2 then we define

a
a
|A| = 11 12
a21 a22

= a11 a22 a12 a21 .

We already encountered this in 5.2.1.

In general if A = (aij ) is n n, first set

a11
a12

a21
a22

..
..
.
.

Cij = (1)i+j ai1 1 ai1
ai+1 1 ai+1

.
..
..
.

an1
an2
.... a1
.... a2
..
.
2
2
j1
j1
.... ai1
.... ai+1
..
.
.... an
j1
j1
j1
a1
a2
..
.
j+1
j+1
ai1
ai+1
..
.
an
j+1
j+1
j+1
.... a1n
.... a2n
..
.
.... ai1
.... ai+1
..
.
.... ann
n
n
called the cofactor of aij . The (n 1) (n 1) determinant is obtained by omitting the ith row
and jth column from A (indicated by the horizontal and vertical lines in the matrix).
We then define
|A| = a11 C11 + a12 C12 + + a1n C1n
which gives a recursive definition of the determinant.
Observe that
(1)i+j
gives the pattern
+ +
+ +
+ +
Thus for a 3 3 matrix A we have

a11 a12 a13

|A| = a21 a22 a23
a31 a32 a33
where
C11

a
a
= 22 23
a32 a33
C12

= a11 C11 + a12 C12 + a13 C13

a
a
= 21 23
a31 a33
C13

a
a
= 21 22
a31 a32
6.2 Properties of Determinants
6.1.1
77
Example
Find the cofactors C13 and C23 of the matrix
3 5 7
1 0 2 .
0 3 0
1+3
C13 = (1)
2+3
C23 = (1)
M13

1 0
= 3,

=1
0 3
M23

3 5
= 9.
= 1
0 3
1+3
C13 = (1)
2+3
C23 = (1)
6.1.2

1
M13 = 1
0

0
= 3,
3

3
M23 = 1
0

5
= 9.
3
Example

3 5 7

Calculate the determinant 1 0 2
0 3 0

3 5 7

1 0 2

0 3 0

= (1) (3) 0 2

3 0

+ (1) (5) 1 2

0 0

+ (1) (7) 1 0

0 3
= 3(0 6) 5(0 0) + 7(3 0)

= 3.

3

1

0
6.2
5
0
3
7
2
0

0
(1) (3)
3

1
2
+ (1) (5)

0
0
3(0 6) 5(0 0) + 7(3 0)
3.

1
2
+ (1) (7)

0
0
Properties of Determinants
Property (1):

|A| = AT
Consider for example the 2 2 case. Then

a11 a12
A=
a21 a22
A =
a11 a21
a12 a22

0
3
78

Thus |A| = a11 a22 a12 a21 = AT . It can be shown that this holds for any square matrix, not just in
the 2 2 case.
This means that any results about the rows in a general determinant is also true about the columns (since
the rows of AT are the columns of A). In particular, any statement about the effect of row operations
on determinants is also true for column operations.
Warning: We used row operations to simplify systems of equations, because they do not change the
solution. Column operations may change the solution of linear systems, so we should not use column
operations on such systems.
Property (2): The determinant may be found by taking cofactors along any row (not just the first)
or down any column. Eg. for a 3 3 matrix A
|A| = a11 C11 + a12 C12 + a13 C13
= a21 C21 + a22 C22 + a23 C23
= a13 C13 + a23 C23 + a33 C33
(definition, expansion along 1st row)

(expansion along 2nd row)
(expansion down 3rd column) etc.
This is useful if one row or column contains a larger number of zeros.

6.2.1
Examples

3 5 7

(i) Calculate 1 0 2 by (a) expanding down the third column, (b) expanding along the third row.
0 3 0
Which is more efficient?

3 5 7

1 0 2

0 3 0

= 7 1 0 2 3 5 = 21 18 = 3

0 3
0 3

3 7
= 3(6 7) = 3
= 3
1 2

3

1

0
5
0
3
7
2
0

1
3
0
5
= 7
2
= 21 18 = 3
0
3
0
3

3
7
= 3
= 3(6 7) = 3
1
2
(ii) Find

3 6 9

0 0 0

2 4 5
exp. down
3rd. col.
exp. along
r3
exp. down
3rd. col.
exp. along
r3
79

3 6 9

0 0 0

2 4 5

= 0 C21 + 0 C22 + 0 C23

= 0

3

0

2
6
0
4
9
0
5
(expanding along r 2 )
0 C21 + 0 C22 + 0 C23
In general the determinant of any matrix with a row or column of zeros is equal to zero.
6.2.2
Effect of Elementary Row Operations on Determinants
The calculation of large determinants from the definition is very tedious. It is much easier to calculate the
determinant after row reducing a matrix, because the reduced matrix will contain many zeros. However,
performing elementary row operations changes the value of a determinant. Luckily the changes are simple:
Property (3): A common factor in all entries of a row (or column) can be taken out as a multiplier
of the determinant.
6.2.3
Example

3 6 9

0 1 2

3 4 5

= 3C11 + 6C12 + 9C13

= 3[C11 + 2C12 + 3C13 ]

1 2 3

= 3 0 1 2
3 4 5
Property (4):
changes sign.
6.2.4
(check!)
If 2 different rows (or columns) of A are interchanged (r i r j or ci cj ), |A|
Example

3 5 7

We have seen 1 0 2
0 3 0

= 3. Interchanging columns 2 and 3 gives

3 7 5

1 2 0

0 0 3

3 7

= 3

1 2 (expanding along r 3 )

= 3.
Let A be a matrix, and let B be the matrix obtained by interchanging two rows or columns. Then
det(A) = det(B). If A has two equal rows (or columns) and these are interchanged then A = B so
det(A) = det(B) = det(A) so det(A) = 0. Hence:
80
If A has two equal rows or columns then det(A) = 0.

Property (5): If a multiple of one row (or column) is added to another (r i
the determinant is unchanged.
r i +r j or ci
ci +cj ),
The above 5 properties give a shortcut for evaluating determinants. We can use a sequence of elementary
row operations (and/or column operations) to produce a matrix with a row (or column) with at most
one non-zero entry. We then expand along this row (or column) to obtain a smaller determinant, and
repeat.
Important: row (and column) operations can change the value of the determinant, so we must always
keep track of which operations we perform. (Unlike doing Gaussian elimination.)
6.2.5
Example

Calculate the determinant

2
1
3
0

6
1
3 2
.
4
4
4
6
5
2
2
2

1
0
5
6
1 0
(r 1
r 1 2r 2 )
2
3 2
3 2 1
=
4
4 0 4 5 10
(r 3
r 3 3r 2 )
2
4
6
4
6 0

1
0 5

= (1) 4 5 10 (expanding down column 1)
2
4 6
2
1
3
0
5
2
2
2

1
0
0

= (1) 4 5 30
2
4 4

5 30
=
4 4
2
1
3
0
5
2
2
2
6
3
4
4
(c3
1
2
4
6

= 20(1 6) = 100.

0

= 1

0

0
1
2
4
2

1

= (1) 4

2
0
5
4

5
10
6

1

= (1) 4

2
0
5
4
0
30
4

5
=
4
c3 5c1 )


1
6
= (5) 4
1 1

30
4

1
= (5) 4
1
0
3
5
4
5
2
10
6
(r 1
r 1 2r 2 )
(r 3
r 3 3r 2 )
(expanding down column 1)
(c3
c3 5c1 )

6
= 20(1 6) = 100.
1
Using the definition with cofactors, to calculate an n n determinant, we have to find n cofactors, each a
(n 1) (n 1) determinant. For each of these we need to find (n 1) cofactors of size (n 2) (n 2)
etc. The total number of calculations is about
n (n 1) (n 2) 3 2 = n!
81
For a 25 25, this is about 1025 operations. This would take thousands of years, even on a computer.
Doing row reductions, only a few thousand calculations are required, which could be done in a fraction
of a second on a computer.
6.2.6
Example

1 1 1

Evaluate a b c . This matrix is called Vandermondes matrix, and it occurs in applications such
a2 b2 c2
as signal processing, error-correcting codes, and polynomial interpolation.

1 1 1 1

1
1

a b c = 0 ba
c a (r 2
r 2 ar1 )

a2 b2 c2 0 b2 a2 c2 a2 (r
r 3 a2 r 1 )
3

ba
c a

= 2
(expand down first column)
b a2 c2 a2
Now we extract a common factor from each of columns 1 and 2 giving

1
1

= (b a)(c a)
b+a c+a
= (b a)(c a)[c + a (b + a)]
= (b a)(c a)(c b).

1

a

2
a
1
b
b2

1

= 0

2
0
c
1
c

ba
= 2
b a2
1
ba
2
b a2

c a
c2 a2
1
ca
2
c a2

(r
2

(r 3
(expand down first column)
Now we extract a common factor from each of columns 1 and 2 giving
6.2.7

(b a)(c a)
(b a)(c a)[c + a (b + a)]
(b a)(c a)(c b).
Determinants of Triangular Matrices
a11
a21 a22 0
.
..
Suppose A =
..
.
an1
0
..
.
..
.
0
ann
r 2 ar1 )
1
b+a

1

c+a
r3 a r1 )
6.3 Connection with inverses
82
is lower triangular. Then by repeated expansion along r 1 we obtain

a22

..

..

.
.

= . . . = a11 a22 ann ,
|A| = a11 .

..
..

.

an2 ann
which is the product of the diagonal entries. The same result holds for upper triangular and diagonal
matrices.
In particular
|I| = 1.
6.3
Connection with inverses
An important property of determinants is the following.

6.3.1
Fact (product of determinants)
Let A, B be n n matrices. Then

|AB| = |A| |B| .
6.3.2
Theorem: Invertible matrices

A is invertible |A| =
6 0.
6.3.3
Proof
= If A is invertible, then
= Follows from 6.3.4 (see below).
I = AA1

1 = |I| = AA1 = |A| A1 ,
So |A| =
6 0.
= If A is invertible, then
1
I = AA

1 = |I| = AA1 = |A| A1 ,
So |A| =
6 0.
= Follows from 6.3.4 (see below).
Remark: It follows immediately from the proof that if A is invertible, then

|A1 | =
1
,
|A|
i.e., the inverse of the determinant is the determinant of the inverse.
6.3.4
83
A Theoretical Formula
There is an explicit formula for the inverse of A. If Cij is the cofactor of the i, j entry of A, let C be the
matrix of cofactors C = (Cij ). Then
1 T
A1 =
C .
|A|
Here C T is called the adjoint matrix of A. This formula is useful for some theoretical problems, but is
useless as a tool for calculations because it is so time consuming to calculate all the cofactors.
Never use this formula for calculating inverses!
6.3.5
Example
In the 2 2 case we have

a
a
|A| = 11 12
a21 a22

= a11 a22 a12 a21

and co-factors
C11 = a22
C12 = a21
|A| =
6 0 A1 =
in agreement with a previous result, 5.2.1.
1
|A|
C21 = a12
C11 C21
C12 C22

=
1
|A|
,

C22 = a11
a22 a12
a21
a11
Notes
84
7.1 Definition: Cross product
85
Vector Products in 3-Space
7.1
Definition: Cross product
For vectors v, w in R3 , the cross (or vector) product of v and w is a vector v w such that
(i) kv wk = kvk kwk sin , where is the angle between v and w
(ii) v w is perpendicular to v and w, in right hand direction from v w (right hand rule).
v xw
Figure 10: The vector v w is perpendicular to both v and w in R3

7.1.1
Application: angular momentum
Vector products also arise naturally in physics and engineering eg. a particle of unit charge moving with
velocity v in a magnetic field B experiences a force F = v B. The angular momentum of a particle of
mass m moving with velocity v is given by L = m(r v) where r is the position vector of the particle.
This is important for central force problems.
7.1.2
(1)
Notes on the cross product
For v, w 6= 0
v w = 0 sin = 0 = 0 or 180
(2)
v w = 0 v and w are parallel (or anti-parallel).

v w = w v since they have the same magnitude but opposite direction
(3)
v v = 0, since sin 0 = 0.
(4) If v w 6= 0, then v w is perpendicular to the plane determined by v and w. This is useful for
obtaining the equation of a plane in R3 (see Math 1052).
7.1.3
Example
ii = jj = kk
=
0
ij =
k
= j i j k = i = k j
ki =
j
= i k.
86
Figure 11: This diagram is an aid in remembering the cross product between unit vectors.
7.1.4
Properties of cross product
It can be shown that

v (w + u) = v w + v u
(v + w) u = v u + w u
v (w) = (v) w = (v w)
R.
We have seen that

v w = w v
so the commutative law fails. Also we do not have the associative law: ie. in general
v (w u) 6= (v w) u
For example, (i i) k = 0 k = 0, but i (i k) = i (j) = k.
Using the above results we can express the cross product of two vectors in terms of their components.
7.1.5
Theorem: Calculating the cross product
v1
Suppose v = v2 ,
v3
ie.
w1
w = w2 . Then
w3
v w = (v2 w3 v3 w2 )i (v1 w3 v3 w1 )j + (v1 w2 v2 w1 )k

i

j
k
v1 v3

v
v
2
3
i
j + v1 v2
v w = v1 v2 v3 =

w1 w2
w2 w3
w1 w3
w1 w2 w3
can be evaluated by expanding above determinant along the first row.
7.1.6
87
Proof
This follows directly by expanding:

v w = (v1 i + v2 j + v3 k) (w1 i + w2 j + w3 k)
= v1 i (w1 i + w2 j + w3 k)
+ v2 j (w1 i + w2 j + w3 k)
+ v3 k (w1 i + w2 j + w3 k)
= v1 w2 (i j) v1 w3 (k i)
v2 w1 (i j) + v2 w3 (j k)
+ v3 w1 (k i) v3 w2 (j k)
= (v2 w3 v3 w2 )i (v1 w3 v3 w1 )j + (v1 w2 v2 w1 )k.
This follows directly by expanding:
v w = (v1 i + v2 j + v3 k) (w1 i + w2 j + w3 k)
= v1 i (w1 i + w2 j + w3 k)
+ v2 j (w1 i + w2 j + w3 k)
+ v3 k (w1 i + w2 j + w3 k)
= v1 w2 (i j) v1 w3 (k i)
v2 w1 (i j) + v2 w3 (j k)
+ v3 w1 (k i) v3 w2 (j k)
= (v2 w3 v3 w2 )i (v1 w3 v3 w1 )j + (v1 w2 v2 w1 )k.
7.1.7
Example
Let v = (1, 2, 3), w = (4, 5, 6). Find a (non-zero) vector orthogonal to both v and w.
7.2 Application: area of a triangle
88
The vector v w will do:

i j k

v w = 1 2 3
4 5 6

2 3
1 3
1 2

j
+k
= i
5 6
4 6
4 5
= i(12 15) j(6 12) + k(5 8)
= 3i + 6j 3k.
This is orthogonal to both v and w.
The vector v w will do:
vw

i
j
k

1
2
3

4
5
6

2

3
j 1
i
4
5
6

1
3
+ k
6
4
i(12 15) j(6 12) + k(5 8)
3i + 6j 3k.

2
5
This is orthogonal to both v and w.
7.2
Application: area of a triangle
Find area of the ABC with vectors v, w along edges as shown in Figure 12.
Figure 12: Find the area of the triangle ABC
Area =
=
Area
1
base perpendicular height
2
1
1
kvk kwk sin = kv wk .
2
2
=
=
1
2
1
2
base perpendicular height

kvk kwk sin =
1
2
kv wk .
7.3 Scalar Triple Product
7.2.1
89
Example
Find the area of a triangle with vertices A = (0, 1, 3), B = (1, 2, 1), C = (4, 1, 0) and obtain a vector
perpendicular to the plane of ABC.
w = AB = 1 ,
2
v = AC = 0
3
Hence a vector perpendicular to the plane of ABC is

i j
k

v w = 4 0 3 = 3i + 5j + 4k.
1 1 2
Area of the triangle
1
kv wk =
2

1
3i + 5j + 4k
2
1
9 + 25 + 16
2
1
50
2
5 2
.
2
=
=
=
w = AB =
1
1 ,
2
v = AC =
4
0
3
Hence a vector perpendicular to the plane of ABC is

i

v w = 4
1
j
0
1
k
3
2

= 3i + 5j + 4k.

Area of the triangle

1
2
kv wk
=
=
=
=
7.3

1

3i + 5j + 4k
2
1
9 + 25 + 16
2
1
50
2
5 2
.
2
Scalar Triple Product
The number u (v w) is called the scalar triple product of u, v, w.

Using the result of the previous theorem, express this product in terms of components.
7.4 Geometrical Interpretation
90
Set u = u1 i + u2 j + u3 k. Then

v2 v3
v1 v3
v1 v2
i

u (v w) = (u1 i + u2 j + u3 k)
w1 w3 j + w1 w2 k
w2 w3

v1 v2
v1 v3
v2 v3

+ u3
u2
= u1
w1 w2
w1 w3
w2 w3

u1 u2 u3

= v1 v2 v3 .
w1 w2 w3
This is a 3 3 determinant.
Set u = u1 i + u2 j + u3 k. Then
u (v w)
=
=

v
v2
v3

i 1
w2
w1
w3

v1
v
v3
v3

+
u3
u
u1 2
2 w
w3
w2
w3
1

u1
u2
u3

v1
v2
v3 .

w1
w2
w3
(u1 i + u2 j + u3 k)

v
v3
j + 1
w3
w1

v1
v2
w1
w2

v2
k
w2
This is a 3 3 determinant.
7.4
Geometrical Interpretation
The volume V of a parallelepiped determined by vectors u, v, w along edges as shown is given by

V = |u (v w)|
vxw
w
A
Figure 13: Parallelepiped determined using vectors.
7.4.1
91
Proof
V = (area of base ABCD) h where

h = perpendicular height
= size of component of u in direction of v w

u (v w)
|u (v w)|

=
(v
=
w)

kv wk
kv wk2
From previous example,
area of the base ABCD = 2 (area ABD)
= kv wk
V = kv wk h = |u (v w)| .
V = (area of base ABCD) h where

h
perpendicular height
size of component of u in direction of v w

u (v w)

|u (v w)|

(v w) =

kv wk2

kv wk
From previous example,

area of the base ABCD
7.4.2
2 (area ABD)
kv wk
V = kv wk h = |u (v w)| .
Example
Find the volume V of the parallelepiped determined by the vectors

1
1
0
2
1
0 .
a=
, b=
, c=
1
1
1
V = |a (b c)|, where

1 2 1

a (b c) = 1 1 1
0 0 1
V = |1| = 1.

=1 1 2

1 1

= 1 2 = 1

(expanded along r3 )
V = |a (b c)|, where
a (b c)

1

1

0
|1| = 1.
2
1
0
1
1
1

=1 1

1

2
= 1 2 = 1
1
(expanded along r3 )
92
The volume V of a parallepiped determined by the vectors u, v, w vanishes if and only if vectors are
co-planar; therefore vectors are co-planar if and only if

u1 u2 u3

v1 v2 v3 = 0.

w1 w2 w3
This gives a useful method for determining when 3 vectors are co-planar.
Notes
93
8.1 Ranking Webpages
94
Eigenvalues and eigenvectors
8.1
Ranking Webpages
Web search engines assign a number to each webpage, called the pagerank. The higher the pagerank, the
higher that webpage is listed in websearches.
If a webpage links to 3 other pages, it passes one third of its rank to each of those pages. The pagerank
of a webpage is calculating by summing all the contributions from in-linking pages.
8.1.1
Example
Four webpages link to each as follows:

1 o
/ 2
@ O
O ^

/ 4
3
Let the ranks of the pages be r1 , r2 , r3 , r4 .
The only page linking in to 4 is page 3. Page 3 links to 3 pages (1, 2, 4), so it passes 1/3 of its rank on
to each. So
1
r4 = r3 .
3
The rank of page 3 is given by:
The only page linking in to 3 is page 1. Page 1 links to 2 pages (2, 3), so it passes 1/2 of its rank on to
each. So
1
r3 = r1 .
2
The only page linking in to 3 is page 1. Page 1 links to 2 pages (2, 3), so it passes 1/2 of its rank on to each. So
r3 =
1
2
r1 .
The pages linking in to page 2 are 1, 3 and 4. So
1
1
r2 = r1 + r3 +
2
3
r1 gets weight 1/2 because page 1 links to 2 pages. r3 gets
r2 =
1
2
r1 +
1
3
r3 +
1
2
1
r4 .
2
weight 1/3 because page 3 links to 3 pages.
r4 .
r1 gets weight 1/2 because page 1 links to 2 pages. r3 gets weight 1/3 because page 3 links to 3 pages.
Altogether we get the following system of equations:
8.1 Ranking Webpages
95
r1 =
r2 =
1
2 r1
r3 =
1
2 r1
r2 +
1
3 r3
1
2 r4
1
3 r3
1
2 r4
1
3 r3
r4 =
r1
r2
r2
1r
2 1
r3
1r
2 1
r4
1r
3 3
1r
2 4
1r
3 3
1r
2 4
1r
3 3
This can be written as Ar = r where
A=
0 1
1
3
1
2
1
3
1
2
0 0
A=
1
2
1
3
0 0
1
2
1
3
1
2
1
3
1
2
1
2
1
3
1
2
Solving this system (exercise) gives r1 = 12, r2 = 9, r3 = 6, r4 = 2 is a solution, so page 1 is ranked

highest, and page 4 lowest.
In general, the pagerank of v is calculated by adding the weight of all the pages w1 , . . . , wn that link in
to v:
X
r(wj )
r(v) =
.
nj
all pages w1 , . . . , wn that link to v
where page wj links to nj pages.

Problem: to find r(v) we need to know r(w1 ), . . . , r(wn ). How do we ever get started?
Idea: given a set of N webpages w1 , . . . , wN , create the N N matrix A = (aij ), where
(
1
,
if page wj links to wi
aij = nj
0,
otherwise
Let
r=
r1
r2
..
.
rN
8.2 Geometry of eigenvectors
96
Then the pagerank equation is

Ar = r.
(8.1)
This is an N N system, but not one that we have seen before because the RHS is not constant. The
unknown vector r appears on both sides of the equation.
Equation (8.1) is called an eigenvector equation. The vector r is called an eigenvector of the matrix A.
Instead of the equation Ax = x, sometimes we often encounter equations like Ax = 2x. This is called
an eigenvector problem with eigenvalue 2. Equation (8.1) is a special case with eigenvalue 1.
8.2
Geometry of eigenvectors

Consider the matrix A =

1
.
j 7
2

A maps i 7
2
1
2 1
1 2

and j 7
1
2

2
. This acts as a linear transformation, mapping i 7
and
1

.
Draw a vector that is unmoved under the action of A.
8.2 Geometry of eigenvectors

The vector
1
1
1
1
Check: A
97

is unchanged.

=
2 1
1 2

For most vectors Av 6= v, but

The vector

Check: A
1
1
1
1
1
1
1
1

=
1
1

.

is special for the matrix A. It is called an eigenvector of A.

is unchanged.

=
2
1
1
2

For most vectors Av 6= v, but
1
1
1
1

=
1
1

.

is special for the matrix A. It is called an eigenvector of A.
Instead of Ax = x we often encounter equations like Ax = 2x. Then 2 is called an eigenvalue, and x a
corresponding eigenvector. Above we looked at the special case of eigenvalue 1.
8.3 Eigenvalues & Vectors
8.2.1
98
Other applications
Eigenvalues and eigenvectors have many other applications e.g. in Quantum Mechanics the energy of a
system is represented by a matrix whose eigenvalues give the allowed energies. This is why discrete energy
levels are observed. Another situation where they are used is in modelling the population dynamics of a
predator-prey system in biology. The eigenvalues and eigenvectors determine the stability and dynamics
of a population in such models.
There are many other applications in Mathematics, Physics, Chemistry and Biology.
8.3
Eigenvalues & Vectors
Let A be an n n matrix and consider the equation

Ax = x
(8.2)
with a number (possibly complex). We are given the matrix A, and need to find and x.
Obviously x = 0 is always a solution to (8.2) (for any ). A value for which (8.2) has solution x 6= 0
is called an eigenvalue of A. The corresponding solution vector x 6= 0 is called an eigenvector of A
corresponding to the eigenvalue .
Note that is allowed to be 0, but x is not allowed to be 0.
8.3.1
Example

1 3
4 2
eigenvector x.
Let A =

1
, = 2 and x =
. Show that is an eigenvalue for A, with corresponding
1
We have to check that Ax = x.

1
2
1
1 3
= x.
= 2
=
Ax =
1
4 2
1
2

3
Check that another eigenvalue is = 5 with corresponding eigenvector
. We have to check that Ax = x.
4

Ax =
1
4
3
2

1
1

Check that another eigenvalue is = 5 with corresponding eigenvector
8.3.2

=
3
4
2
2

= 2
1
1

= x.

.
Note
Usually Ax is related to x in a much more complicated way,

1 3
1
7
1
e.g.
=
6=
.
4 2
2
8
2
8.4 How to find eigenvalues
8.4
99
How to find eigenvalues
We have
Ax
Ax
Ax Ix
(A I)x
=
x
= (Ix)
=
0
=
0.
We seek a non-zero solution x of this equation.

Recall that if B is an n n matrix and det B 6= 0 then Bx = 0 has a unique solution, which must be
x = 0 ( 5.3.2 and 6.3.2). Since we want a non-zero solution x above, we must have det(A I) = 0.
Thus A has eigenvalue only if
|A I| = 0.
(8.3)
We will explore eigenvalues and eigenvectors in a MATLAB module.

8.4.1
Definition: Characteristic polynomial, characteristic equation
The determinant
p() = |A I|
is a polynomial of degree n in called the characteristic polynomial of A. We call equation (8.3) the
characteristic equation.
Thus the eigenvalues of A are given by the roots of the characteristic equation (8.3). This means that
A has at least 1 and at most n distinct eigenvalues (possibly complex). The corresponding eigenvectors
may then be obtained for each in turn by solving
(A I)x = 0
using Gaussian elimination.
8.4.2
Example
Find the eigenvalues and eigenvectors of

A=
5
2
2 2

.
100
We have
5
2
2
2
A I =
So

5
2
p() = |A I| =
2
2

.

= ( + 5)( + 2) 4

= 2 + 7 + 6 = ( + 6)( + 1).
Solving p() = 0 yields the eigenvalues 1 = 1, 2 = 6.

A I =
So
5
2

5
p() = |A I| =
2
We have
2
2

.

2
= ( + 5)( + 2) 4
2
= 2 + 7 + 6 = ( + 6)( + 1).
Solving p() = 0 yields the eigenvalues 1 = 1, 2 = 6.
To get the corresponding eigenvectors, solve (A i )x = 0 for i = 1, 2.

1 12 0
1 12 0
4
2 0
1 = 1 :
2 1 0
2 1 0
0
0 0
with general solution

x=

=

1
1
is an eigenvector corresponding
2
2
to the eigenvalue = 1;

1 2 0
2 = 6 :
0 0 0

2
2
general solution x =
=
.
1

2
is an eigenvector corresponding to eigenvalue = 6.
1

Hence
1 2 0
2 4 0
To get the corresponding eigenvectors, solve (A i )x = 0 for i = 1, 2.

4
2
0
1
1 = 1 :
2
1
0
2
1
2
1
0
0
1
0
1
2
0
0
0
with general solution

x=

1
2

1
2
is an eigenvector corresponding
to the eigenvalue = 1;

2 = 6 :
1
2
2
4
0
0

general solution x =
Hence

2
1
1
0
2
0

2
.
1
0
0
is an eigenvector corresponding to eigenvalue = 6.
8.4.3
101
Example

Find the eigenvalues and eigenvectors of A =

2
p() = |A I| =
2
0 2
2 0

= 2 + 4.

Eigenvalues: p() = 0 2 + 4 = 0 = 2i 1 = 2i; 2 = 2i.

In this case A has two complex
= 2i, exactly as before, using
to 1 = 2i is

2i
2
eigenvalues = 2i. To obtain eigenvectors solve (A I)x = 0,

complex numbers. The augmented matrix (A I | 0) corresponding
2 0
2i 0
x1 + ix2 = 0
2i 2 0
0 0 0
x1 = ix2 = i
1 i 0
0 0 0
x2 =

i
i
x =
=
is the general solution,
1

and
i
1

is an eigenvector corresponding to 1 = 2i.
For = 2i:

2i 2 0
2 2i 0
1 i 0
0
0 0
x1 ix2 = 0 x1 = ix2 = i
x2 =
x=
i
1

is the general solution, and
i
1
is an eigenvector for = 2i.

p() = |A I| =
2

2
= 2 + 4.

Eigenvalues: p() = 0 2 + 4 = 0 = 2i 1 = 2i; 2 = 2i.

In this case A has two complex eigenvalues = 2i. To obtain eigenvectors solve (A I)x = 0, = 2i, exactly as before, using complex numbers. The
augmented matrix (A I | 0) corresponding to 1 = 2i is

2i
2
2
2i
0
0
x1 + ix2 = 0
2i
0
2
0
0
0
1
0
i
0
0
0
x1 = ix2 = i
x2 =

i
i
x =
=
is the general solution,
1

and
i
1

is an eigenvector corresponding to 1 = 2i.
For = 2i:

x=
2i
2
2
2i
0
0
1
0
i
0
0
0
x1 ix2 = 0 x1 = ix2 = i

=
i
1
x2 =

i
is the general solution, and
1
is an eigenvector for = 2i.
8.4.4
102
Check
Solving the system (A I)x = 0 to find eigenvectors always leads to at least one row of zeros when
A I is row reduced. This is because is chosen so that (A I)x = 0 has a solution x 6= 0.
So A I is not invertible and so A I does not reduce to I.
Notes
103
9.1 Linear combinations
104
Vector Spaces
In this chapter we consider sets of vectors with special properties, called vector spaces.
Recall that if A and B are sets then A B means A is a subset of B. The set with no elements, called
the empty set is denoted . This is not the same as the set {0} whose single element is the zero vector.
9.1
Linear combinations

x
x
can be written using the vectors i and j:
= x i + y j. In R3 every
y
y
vector can be written using i, j, k. We generalize this idea.
In R2 , every vector
9.1.1
Definition: Linear combination
If v 1 , v 2 , ..., v m Rn , a vector of the form

v = 1 v 1 + 2 v 2 + + m v m ,
i R
is called a linear combination of the vectors v 1 , v 2 , ..., v m .

9.1.2
Example
Show that each of the vectors
is a linear combination of

1
0
, w2 =
w1 =
0
1

1
1
v1 =
and v 2 =
.
1
1
9.1 Linear combinations
105
Write
w 1 = 1

1
1
.
+ 2
1
1
Equating components:
1 + 2 = 1,
1 2 = 0
Solving, 1 = 2 = 1/2. So
w1 =
=

1 1
1
1 1
+
=
2 1
2 1
0
1
1
v1 + v2
2
2
Similarly we get
w2

0
1 1
1 1
=
=
1
2 1
2 1
=
1
1
v1 v2.
2
2
Write
w 1 = 1

1
1

+ 2

.
Equating components:
1 + 2
1,
1 2
Solving, 1 = 2 = 1/2. So
w1

1
0
1
2
v1 +

1
1
2
1
2
v2
Similarly we get
w2
9.1.3

0
1
1
2
v1

1
1
2
1
2
v2 .
Example
Every v =
v1
v2
..
.
Rn is a linear combination of the coordinate vectors ei :
vn
v = v1 e1 + v2 e2 + + vn en .
(9.1)
9.3 How to test for linear independence
9.2
9.2.1
106
Linear independence
Definition: Linear independence
Consider the linear combination

1 v 1 + 2 v 2 + + m v m
with 1 = 2 = = n = 0. Obviously this gives the 0 vector.
A set of vectors S = {v 1 , v 2 , ..., v m } Rn is linearly dependent if there exist scalars 1 , ..., m not all
zero such that
1 v 1 + 2 v 2 + + m v m = 0.
S is called linearly independent if no such scalars exist, i.e. if the only linear combination adding up to
0 is with all the scalars 0. So S is linearly independent if
1 v 1 + 2 v 2 + + m v m = 0
9.3
1 = 2 = = m = 0
How to test for linear independence
Given vectors v 1 , . . . , v m , solve the vector equation

1 v 1 + + m v m = 0.
If the only solution is 1 , . . . , m = 0 then the vectors are linearly independent.
If there is any other solution, the vectors are linearly dependent.
9.3.1
Example
Show that in
R2 ,

the vectors
1
1
1
1

,
9.3.2

are linearly independent.

0
+
0
1
=
+
=
0
0
1
+ =0
= = 0.
=0

1
1
1
1

+

1
0
+
0
=
=
1
0
0
+ =0
= = 0.
=0
Example
Show that in Rn the coordinate vectors ei are linearly independent.
1
0
..
.
107
+
n en =
0
1 e1+ 2 e2 +
1
0
0
.. 2
2

+ .. + + . = ..
0 .
.
0
0
..
.
1 = 0, 2 = 0, . . . , n = 0.
1
0
.
.
.
0
1 e1 + 2 e2 + + n en
0
0
.
2
+ + .. =
.
.
0
.
0
n
=0
1
2
..
.
n
0
0
= .
.
1 = 0, 2 = 0, . . . , n = 0.
9.3.3
Example

1
1
0
Show that in R3 , the vectors 3 , 1 , 1 are linearly dependent.
1
1
1
108
1
1 3 + 2
1

0
1
1
1 0
3
0
1 1
0
1 1 1
0
0

1
0
0
1 + 3 1 = 0
1
1
0

1 0
0
1 1
0
0
0 0 1 12
0 .
2 1
2 1
0 0
0
0
0
1
2
1
1
1
x3 is not determined. Let x3 = . Then x2 = 2 , x1 = 2 , so x =

for any . If we take
2
1
1
= 2, x = 1 is a solution.
2
Check:

1
1
0
0
3 1 2 1 = 0 .
1
1
1
0
We have found a non-trivial linear combination summing to 0, so the vectors are linearly dependent.
1
1
0
0
1 3 + 2 1 + 3 1 = 0
1
1
1
0
1
3
1
1
1
1
1
0
0 0
0
0
0
1
1
1
2
2
x3 is not determined. Let x3 = . Then x2 = 1

, x1 = 1
, so x =
2
2
1
0
0 0
0
0
0
1
1
1
2
1
2
1
1
0
0
1
2
0
0
0 .
0
1
for any . If we take = 2, x = 1 is a solution.
2
Check:
0
0
1
1
3 1 2 1 = 0 .
0
1
1
1
We have found a non-trivial linear combination summing to 0, so the vectors are linearly dependent.
9.3.4
Example
Show that any finite set of vectors containing the zero vector 0 is linearly dependent.
Write S = {v 1 , ..., v m , 0} Rn . Then

0 v1 + + 0 vm + 1 0 = 0
S linearly dependent.
Write S = {v 1 , ..., v m , 0} Rn . Then

0 v1 + + 0 vm + 1 0 = 0
S linearly dependent.
9.3.5
Special cases of linear dependence
If two vectors v 1 , v 2 are linearly dependent, then there exist scalars 1 , 2 R, not both 0, with
1 v 1 + 2 v 2 = 0.
9.4 Invertible matrices and linear independence

2
If 1 6= 0 then v 1 =
1 v 2 , so v 1 is a scalar multiple of v 2 . If 2 6= 0 then v 2 =
scalar multiple of v 1 .
109
1
2 v 1 ,
so v 2 is a
So two vectors are linearly dependent if and only if one of them is a multiple of the other.
(ii) If one vector v is linearly dependent then v = 0 with 6= 0, so v = 0. Thus a single vector is
linearly dependent if and only if it is the zero vector.
9.4
Invertible matrices and linear independence
There is a relation between the nullspace of A and between linear independence of the columns of A.
b1
b2
Suppose b NS(A). Let b =

. That is b = b1 e1 + b2 e2 + + bn en .
bn
Since b NS(A), Ab = 0 so

0 = Ab = A b1 e1 + b2 e2 + + bn en
= b1 (Ae1 ) + b2 (Ae2 ) + + bn (Aen ).
We saw earlier that Aej is the jth column vector of A. So the right hand side is a linear combination
among the columns of A.
Thus: if there is a vector b NS(A) with b 6= 0 then there is a non-trivial linear combination of columns
of A adding up to 0, so the columns of A are linearly dependent.
Conversely, if the columns of A are dependent there is a non-trivial linear combination of columns of
A adding up to 0, say b1 (Ae
1) +
b2 (Ae2 ) + + bn (Aen ) = 0. This gives a non-zero vector b =
b1
b2
b1 e1 + b2 e2 + + bn en =
in NS(A).
bn
Thus: NS(A) = {0} if and only if the columns of A are linearly independent. But recall (5.3.2) that A
is invertible exactly when NS(A) = {0}.
We have proved:
9.4.1
Theorem (existence of inverse)
Let A be a square matrix. Then A is invertible if and only if the columns of A are linearly independent.
Notice this means swapping the columns of a matrix does not affect whether or not it is invertible.
Furthermore, A is invertible if and only if the rows of A are linearly independent:
A is invertible AT is invertible
columns of AT are linearly independent
rows of A are linearly independent.
9.4 Invertible matrices and linear independence
9.4.2
110
Example
From example 9.3.3, the columns of
1 1 0
A= 3 1 1
1 1 1
are linearly dependent. Therefore A is not invertible.
9.4.3
Let
Example
1 1 1
A= 0 1 1
0 0 1
It may be shown that the columns of A are linearly independent (left as an exercise). Therefore A is
invertible.
9.4.4
Invertibility and the Nullspace
9.4.5
Summary of conditions equivalent to invertibility
Let A be an n n matrix. Give a summary of all the conditions we have found that are equivalent to A
being invertible.
9.5 Vector spaces
111
1. A is invertible.
2. There exists an n n matrix X with AX = XA = I
3. AT is invertible
[ 5.3.4(3)].
4. The columns of A are linearly independent

5. The rows of A are linearly independent
[ 9.4.1].
[ 9.4.1].
6. Ax = b has a unique solution x for any b R

7. N (A) = {0}
[Definition of invertible, 5.2].
[ 5.3.2 ].
[ 5.3.2 ].
8. The row reduced form of A is I

9. det(A) 6= 0
[ 4.4].
[Theorem 6.3.2].
10. rank A = n.
1. A is invertible.
2. There exists an n n matrix X with AX = XA = I
3. A
is invertible
[ 5.3.4(3)].
4. The columns of A are linearly independent

5. The rows of A are linearly independent
[ 9.4.1].
[ 9.4.1].
6. Ax = b has a unique solution x for any b R

7. N (A) = {0}
[ 5.3.2 ].
[ 5.3.2 ].
8. The row reduced form of A is I

9. det(A) 6= 0
[Definition of invertible, 5.2].
[ 4.4].
[Theorem 6.3.2].
10. rank A = n.
9.5
Vector spaces
Consider the following sets of vectors:

V =
x
y

| x + y = 2 . The set V is a set of (infinitely many) vectors in R2 .
The set containing just the zero vector {0}. This is a set with just one vector.

W =
x
y

|
x2
y2

= 0 . Since x2 , y 2 0, the only vector in W is 0, so W = {0}.
Another special set is the empty set, not containing any vectors at all(!) e.g.

x
2
2
Y =
| x + y = 1 . There are no vectors in Y .
y

Let V = b a + b = 2c .
9.5 Vector spaces
112
1
1
2
Is 0 V ? No: 1 + 0 6= 2 3. But v = 1 , w = 4 V :
3
1
3

3
Is v + w V ? Yes: v + w = 5 and 3 + 5 = 2 4.
4
1 + 1 = 2 1 and 2 + 4 = 2 3.
We now consider sets of vectors with certain properties, called vector spaces.
9.5.1
Definition: Vector space, subspace
A subset V Rn is called a vector space if it satisfies the following

(o) V is non-empty.
(i) If v V and w V then v + w V
(closure under vector addition).
(ii) If v V and R then v V (closure under scalar multiplication).

If V is a vector space, W V and W is also a vector space, we call W a subspace of V . Note: V is a
subspace of itself.
If V Rn is a vector space and v V , then by property (ii)
0 = 0v V
that is, the zero vector belongs to every vector space.
9.5.2
Example
Show that
is a vector space.
a
V = b a + b = 2c, a, b, c R
9.5 Vector spaces
113
(o) V is non-empty
V.
since 0
a1
a2
Let v = b1 , w = b2 be in V .
c1
c2
a1 + a2
b1 + b2
(i) v + w =
c1 + c2
v V a1 + b1 = 2c1
w V a2 + b2 = 2c2 .
Is v + w in V ?
(a1 + a2 ) + (b1 + b2 ) = (a1 + b1 ) + (a2 + b2 )
= 2c1 + 2c2 = 2(c1 + c2 )
v + w V . So V is closed under addition.
a1
a2
Consider v = b1 , w = b2 V .
c1
c2
a1
(ii) If is any scalar then v = b1
c1
(o) V is non-empty
since 0 V.
a1
a1
2
1
1
1
1
Let v = b1 , w = b2 be in V .
c1
c2
a + b = (a + b ) = 2c = 2(c1 )
v V. So V is closed under scalar multiplication.
a1 + a2
(i) v + w = b1 + b2
c1 + c2
Hence V is a vector space.
v V a1 + b1 = 2c1
w V a2 + b2 = 2c2 .
Is v + w in V ?
(a1 + a2 ) + (b1 + b2 ) = (a1 + b1 ) + (a2 + b2 )
= 2c1 + 2c2 = 2(c1 + c2 )
v + w V . So V is closed under addition.
a1
Consider v = b1 ,
c1
a2
w = b2 V .
c2
a1
(ii) If is any scalar then v = b1
c1
9.5.3
Example
a1 + b1 = (a1 + b1 ) = 2c1 = 2(c1 )
Is V ={0}
a vector space?
v V. So V is closed under scalar multiplication.
Hence V is a vector space.
(o) 0 V V is non-empty.
(i) If v, w V then v = w = 0 and v + w = 0 V .
(ii) If R, then 0 = 0 V .
(o) 0 V V is non-empty.
(i) If v, w V then v = w = 0 and v + w = 0 V .
(ii) If R, then 0 = 0 V .
9.5 Vector spaces
9.5.4
114
Example
Show that V = Rn is a vector space.
(o) 0 V
(i) v, w V v + w Rn = V
(ii) v V v Rn = V
So V is a vector space.
(o) 0 V
(i) v, w V v + w Rn = V
(ii) v V v Rn = V
So V is a vector space.
9.5.5
Example

Is V =
1
2

a vector space?
No, since 0
/ V.
/ V.
No, since 0
9.5.6
Example

1
Is V =
| R a vector space?
2
9.5 Vector spaces
115
(o) = 0 0 V .

(i) If v, w V then v =
1
2

and w =
1
2

for suitable , R.

v+w =
+
2

+
=
2( + )

1
= ( + )
V.
2
Then

(ii) If R, then v =
1
2

V.
Therefore V is a vector space.

(o) = 0 0 V .

(i) If v, w V then v =
1
2

and w =
1
2

for suitable , R.

Then
v+w
=

=
=

(ii) If R, then v =
1
2

+
2

+
2( + )

1
( + )
V.
2

V.
Geometrically, all these vectors can be viewed as lying along a single line through the origin.
9.5.7
Example
Show that for v 1 , v 2 R3 ,

V = {1 v 1 + 2 v 2 | 1 , 2 R}
is a vector space.
(o) If we choose 1 = 2 = 0 we see that 0v 1 + 0v 2 = 0 V , so V is non-empty.

(i) Let v, w V there exists 1 , 2 , 1 , 2 such that v = 1 v 1 + 2 v 2 and w = 1 v 1 + 2 v 2 .
Then v + w = (1 + 1 )v 1 + (2 + 2 )v 2 V .
(ii) For any scalar R, v = 1 v 1 + 2 v 2 V .
(o) If we choose 1 = 2 = 0 we see that 0v 1 + 0v 2 = 0 V , so V is non-empty.
(i) Let v, w V there exists 1 , 2 , 1 , 2 such that v = 1 v 1 + 2 v 2 and w = 1 v 1 + 2 v 2 . Then v + w = (1 + 1 )v 1 + (2 + 2 )v 2 V .
(ii) For any scalar R, v = 1 v 1 + 2 v 2 V .
9.6 Eigenspace
116
If we draw v 1 and v 2 starting from the origin, then all these combinations 1 v 1 + 2 v 2 will lie in a single
plane in R3 , through the origin.
For example, if v 1 = i and v 2 = j then V = {1 i + 2 j | 1 , 2 R} is the set of vectors with no
k-component i.e. the xy plane.
Thus geometrically, a vector space can be a single point {0}, or can be all vectors along a single line
through the origin, or all vectors lying in a single plane through the origin, or similarly in higher dimensions.
Theorem: If A is a matrix, then NS(A) is a vector space.
Proof:
(o) A0 = 0
0 NS(A)
NS(A) is non-empty.
(i) v, w NS(A) Av = 0, Aw = 0
A(v + w) = Av + Aw = 0 + 0 = 0
v + w NS(A)
(ii) If v NS(A), R then Av = 0

A(v) = Av = 0 = 0 v NS(A).
Hence NS(A) is a vector space.
(o) A0 = 0
0 NS(A)
NS(A) is non-empty.
(i) v, w NS(A)
Av = 0, Aw = 0
A(v + w) = Av + Aw = 0 + 0 = 0
v + w NS(A)
(ii) If v NS(A), R then Av = 0

A(v) = Av = 0 = 0
v NS(A).
Hence NS(A) is a vector space.
This explains why either NS(A) = {0} or else it is infinite. If there is any solution v of Av = 0, there will
be infinitely many solutions. This is why a consistent linear system can have one solution, or infinitely
many, depending on the size of NS(A).
9.6
9.6.1
Eigenspace
Proposition (eigenspace is a vector space)
Let A be a square matrix. Let be an eigenvalue of A. Then the set of solutions of the equation Ax = x
forms a vector space called the eigenspace of A corresponding to eigenvalue , denoted N . So
N = {x Rn | Ax = x}.
Special case: if = 0 is an eigenvalue of A then N0 = {x Rn | Ax = 0} = NS(A), the nullspace of A
as above.
9.7 The span of a set of vectors
9.6.2
117
Proof
(o) Note that 0 is always a solution to Ax = x, so 0 N .

Take x, y N , R.
(i) Then
A(x + y)
Ax + Ay
x + y
(x + y)
x + y N .
(ii)
A(x)
Ax
(x)
x N .
Hence the set of solutions to Ax = x forms a vector space.
(o) Note that 0 is always a solution to Ax = x, so 0 N .
Take x, y N , R.
(i) Then
A(x + y)
Ax + Ay
x + y
(x + y)
x + y N .
Ax
(x)
x N .
(ii)
A(x)
Hence the set of solutions to Ax = x forms a vector space.
This is why when we found eigenvectors previously, we actually found infinitely many eigenvectors for
each .
9.7
The span of a set of vectors
We would like a compact way to describe all the vectors in a vector space V , and a way of generating a
vector space from a given set of vectors.
9.7.1
Definition: Span
The span of vectors v 1 , v 2 , ..., v m Rn is the set of all possible linear combinations of these vectors:
n
o
span{v 1 , v 2 , ..., v m } = 1 v 1 + 2 v 2 + + m v m | 1 , 2 , . . . , m R .
118
Note that the span is a set of vectors.

9.7.2
Example

3
5
Show that v =

is in span
1
1

1
,
.
1

3
1
1
We want
=
+
.
5
1
1

Solving, = 4, = 1. We want 35 =
1
1

+
1
1

.
Solving, = 4, = 1.
9.7.3
Example
In fact, every vector in R2 is in the previous spanning set:

1
1
Show that span
,
= R2 .
1
1

Let v =
v1
v2
R2 be arbitrary. We need to find 1 and 2 R such that

1
+ 2
v = 1
1
So we need:
(
v 1 = 1 + 2
v 2 = 1 2

1 + 2
1
=
.
1
1 2
1 = (v1 + v2 )
2
.
2 = (v1 v2 )
2
Hence

1
1
1
1
v1
v1 + v2
v1 v2
v=
=
+
.
1
1
v2
2
2

1
1
2
.
So any vector in R can be written as a combination of
and
1
1

3
The previous example was v =
, so v1 = 3, v2 = 5 so 12 (v1 + v2 ) = 4,
5

v=
v1
v2
R2 be arbitrary. We need to find 1 and 2 R such that

1
+ 2
1
So we need:
v 1 = 1 + 2
v 2 = 1 2
v = 1
Hence
1
1

=
1 + 2
1 2

.
1 = (v1 + v2 )
2
.
2 = (v1 v2 )
2

1

1
1
v1 v2
+
.
1
1
2

1
1
So any vector in R2 can be written as a combination of
and
.
1
1

3
1
1 (v v ) = 1.
The previous example was v =
, so v1 = 3, v2 = 5 so 2 (v1 + v2 ) = 4, 2
1
2
5

v=
v1
v2
1
2
v1 + v2
1
2 (v1
v2 ) = 1.
Let
9.7.4
119
Example

1
1
1
Show that span 0 , 1 , 1 = R3 .
0
0
1
a
Let v = b R3 be arbitrary. We need to find , , R such that
c

1
1
1
v = 0 + 1 + 1 .
0
0
1
That is,
+ + = a
+ + = b
= c
Back substitution gives = c, = b c, = a b. So

a
1
1
1
v = b = (a b) 0 + (b c) 1 + c 1 .
c
0
0
1
a
Let v = b R3 be arbitrary. We need to find , , R such that
c
1
1
1
v = 0 + 1 + 1 .
0
0
1
That is,
+
+
=
+
+
=
=
a
b
Back substitution gives = c, = b c, = a b. So
1
a
1
1
v = b = (a b) 0 + (b c) 1 + c 1 .
1
c
0
0
9.7.5
The Span of vectors is a Vector Space
The span of any non-empty set of vectors is a vector space.

We proved this for two vectors in Example 9.5.7. The general proof is similar.
If span{v 1 , v 2 , ..., v m } = V then we say that v 1 , . . . , v m span V , or that they form a spanning set for
V.
9.7.6
Example

1
Find a spanning set for V =
|R .
2
9.8 Bases
120

1
1
Every vector in V is a linear combination of
, so a spanning set is
.
2
2

1
This answer is not unique. For example
is another spanning set for V .
2

Every vector in V is a linear
1
, so a spanning set is
.
2

1
This answer is not unique. For example
is another spanning set for V .
2
combination of
9.7.7
1
2
Example
0
..
.

0

The coordinate vectors ei =

1 i
0

..
.
span Rn since every vector v Rn is a linear combination of
0
these vectors (see equation 9.1).
9.7.8
Example
The vector space {0} is spanned by the zero vector 0 Rn is and called the zero vector space.
9.8
Bases
We have seen that we can describe a vector space by giving a spanning set. When we do this, we usually
like to choose a spanning set that is as small as possible (contains no redundancies). We ensure this as
follows.
9.8.1
Definition: Basis
Let V be a vector space. We say that a set of vectors B = {v 1 , ..., v m } V is a basis for V if:
1. B spans V ;
2. B is linearly independent.
To prove that a set of vectors is a basis, we must prove both properties.
Sometimes we are just given a vector space V , and asked to find a basis. Problems like this are solved
by finding a spanning set for V , and then checking that the set is linearly independent.
A single vector space V may have many different bases.
9.8.2
Example
We have seen that the coordinate vectors ei , i = 1, ..., n, are linearly independent (section 9.3.2) and
span Rn (section 9.7.7). Therefore they form a basis for Rn , called the standard basis.
9.9 Dimension
9.8.3
121
Example

Show that the vectors
1
1

,
1
1
form a basis for R2 .
We must check: (i) linear independence; (ii) the set spans R2 .
(i) was example 9.3.1.

(ii) was example 9.7.3.
We must check: (i) linear independence; (ii) the set spans R2 .
(i) was example 9.3.1.
(ii) was example 9.7.3.

Note that this basis differs from the standard
9.8.4
1
0

0
,
basis of R2 , but still has two elements.
1
Example
Find a basis for the vector space of 9.5.2

V = b a + b = 2c , a, b, c R
Any vector in V is of the form
2c b
2c
b
= 0 + b = cv 1 + bv 2
b
v=
c
c
0
with c, b R. Thus {v 1 , v 2 } is a spanning set for V . For {v 1 , v 2 } to be a basis, we need to show v 1 and
v 2 are linearly independent. If 1 v 1 + 2 v 2 = 0 then from the second coordinate 2 = 0 and from the
first 1 = 0, so {v 1 , v 2 } is also linearly independent. Any vector in V is of the form
v=
2c b
2c
b
= 0 + b = cv + bv
b
1
2
c
c
0
with c, b R. Thus {v 1 , v 2 } is a spanning set for V . For {v 1 , v 2 } to be a basis, we need to show v 1 and v 2 are linearly independent. If 1 v 1 + 2 v 2 = 0
then from the second coordinate 2 = 0 and from the first 1 = 0, so {v 1 , v 2 } is also linearly independent.
9.9
Dimension
Suppose B = {v 1 , ..., v m } is a basis for V . Then it can be shown that every basis for V has m vectors.
We call m the dimension of V and write m = dim V . (By convention, the zero vector space {0} has
dimension zero.)
9.10 Components
9.9.1
122
Example
We have seen that the coordinate vectors ei , form a basis for Rn , so

dim Rn = n.
9.9.2
Example
The vector space

1
V =
| R R2
2

is spanned by a single vector
1
2

(example 9.7.6) which thus constitutes a basis for V . Hence dim V =
1.
Geometrically, V consists of vectors along a single line, which is a one dimensional object.
9.9.3
Example
Consider the vector space of 9.5.2

b
V =
a + b = 2c , a, b, c R
c
In example 9.8.4 we found a basis with 2 vectors, so dim V = 2. Note that V is a subspace of R3 , but it
consists of a plane through the origin, which geometrically is a two dimensional object.
9.10
Components
Suppose {v 1 , . . . , v m } is a basis for V . Then every v V may be written as a linear combination of

elements in B
v = 1 v 1 + + m v m
and this representation is unique: if we also have
v = 1 v 1 + + m v m
then 1 = 1 , . . . , m = m .
Proof:
123
Since B is a basis, it spans V .

Hence every v can be written as some linear combination of v 1 , . . . , v m .
We need to prove uniqueness. But
0 = v v = (1 v 1 + + m v m ) (1 v 1 + + m v m )
= (1 1 )v 1 + + (m m )v m .
Since B is linearly independent, i i = 0, i = 1, ..., m.
So i = i meaning that the representation is unique.
Since B is a basis, it spans V .
Hence every v can be written as some linear combination of v 1 , . . . , v m .

We need to prove uniqueness. But
0
v v = (1 v 1 + + m v m ) (1 v 1 + + m v m )
(1 1 )v 1 + + (m m )v m .
Since B is linearly independent, i i = 0, i = 1, ..., m.

So i = i meaning that the representation is unique.
If v = 1 v 1 + + m v m we call i the ith component of v with respect to the basis B.
9.11
Further properties of bases
With more work, one can also prove the following:

(i) Every vector space has a basis.
(ii) If m = dim V , any m linearly independent vectors in V will form a basis.
(iii) If m = dim V , any set of more than m vectors in V will be linearly dependent (and so not be a
basis).
(iv) If m = dim V , any set of fewer than m vectors in V will not span V (and so not be a basis).
(v) If W is a subspace of V , then dim W dim V and dim W = dim V if and only if V = W .
(vi) The only m-dimensional subspace of Rm is Rm itself.
9.11.1
Example
Show that in R3 , the vectors
1
v1 = 0 ,
0
1
v2 = 1
0
are linearly independent and hence form a basis for R3 .
1
v3 = 1
1
124
Setting v 1 + v 2 + v 3 = 0 implies

1
1
1
0
0
1
1
0
+
+
=
0
0
1
0

++
0
+
0
0
=0=0=0
Therefore the vectors are linearly independent.
Because we already know dim R3 = 3 and we have found 3 linearly independent vectors, they must form
a basis, by (vi) above.
(Alternatively: we proved they span in example 9.7.4.) Setting v1 + v2 + v3 = 0 implies
1
1
1
0 + 1 + 1
0
0
1
++
0
0
0
0
0
0
=0 =0=0
Therefore the vectors are linearly independent.
Because we already know dim R3 = 3 and we have found 3 linearly independent vectors, they must form a basis, by (vi) above.
(Alternatively: we proved they span in example 9.7.4.)
The shortcut in this example is fine if we already know the dimension of V . If we are just given a space
like

b
V =
a + b = 2c , a, b, c R
c
it may not be obvious what the dimension is.
Notes
125
10.3 Definition: Intervals
10
126
Numbers
Mathematics uses structures at a fundamental level. These basic structures include numbers, sets (e.g.
intervals), shapes, vectors, complex numbers, quaternions and many more strange and exotic things.
Numbers are very important in mathematics as building blocks as they lead to equations and inequalities
which we can use as a tool to solve many abstract and applied problems.
10.1
Number systems
R:
The following are commonly used subsets of R:
N:
Z:
Q:
Irrational
numbers are real numbers which cannot be represented as a ratio of integers. For example,
2, 3, 5, = 3.14159 . . ., e = 2.71828 . . . are all irrational. Note: proving rationality or irrationality
of a given number can be quite subtle!
10.2
Real number line and ordering on R
The real number system can be visualised by imagining each real number as a point on a line, with the
positive direction being to the right, and an arbitrary origin being chosen to represent 0 (see Figure 14).
1.5
3
e
1
Figure 14: The real number line.

The real numbers are ordered, i.e. given any 2 real numbers a and b there holds precisely one of the
following: a > b, a < b or a = b.This means we can use the symbols <, >, and to write
statements such as 1 2 and 3 > 3. Geometrically, a < b means that a lies to the left of b on the real
number line. Note also that a b means that either a < b or a = b.
10.3
Definition: Intervals
An interval is a set of real numbers that can be thought of as a segment of the real number line. For
a < b, the open interval from a to b is given by
(a, b) = {x R| a < x < b}.
This notation is interpreted as
10.3 Definition: Intervals
127
Note here that the end points are not included. If we wanted to include the endpoints, we would have a
closed interval from a to b, denoted
In this notation it is important to distinguish between the round brackets (a, b) for an open interval
(excluding the end points) and the square brackets [a, b] for a closed interval (including the end points).
We also have half-open intervals (or half-closed depending on how you feel), denoted
10.4 Absolute value
10.4
128
Absolute value
We define

|x| =
10.4.1
x,
if x 0
x, if x < 0
Examples
|3| =
| 4| =
|1 | =
10.4.2
Properties of absolute value
It is worth emphasising that in general |a + b| =

6 |a| + |b|, but the triangle inequality does hold:
|a + b| |a| + |b|.
For example, set a = 1 and b = 2. We have l.h.s = |1 + (2)| = | 1| = 1,
r.h.s = |1| + | 2| = 1 + 2 = 3.
10.4.3
Convention for
For a > 0, a always denotes the positive solution of x2 = a. Thus 4 = 2 and so on. This means that
we can now solve x2 = a. For a > 0, solutions to x2 = a are x = a.
10.4 Absolute value
Notes
129
10.4 Absolute value
Notes
130
11.2 Definition: Function, domain, range
11
11.1
131
Functions
Notation
Let X and Y be subsets of R. The notation f : X Y means f maps from X into Y .
11.2
Definition: Function, domain, range
A function f : X Y is a rule which assigns to every element x X exactly one element f (x) Y
called the value of f at x. X is called the domain of f and
f (X) = {f (x)| x X}
is called the range of f .
The range of f , f (X), is a subset of Y . In other words, you can take any x value from X, plug it into f ,
and return the value f (x), which is in Y . The range is the set of all possible values of f (x) as x varies
throughout the domain. Note that f (X) is not necessarily equal to all of Y .
11.2.1
Example
The function f : R R, such that f (x) = x2 is a function with domain R. The range is given by
f (R) = {f (x) | x R}
=
=
11.2.2
Example
The following is a perfectly good example of a function, g : (6, 7) R
5 , 6 < x < 0
, x = 0
g(x) =
x , 0<x<7
Piecewise defined functions are discussed in Stewart, 7ed. p. 16; 6ed. p. 17.
11.3 Graphs
132
Clearly the domain of g is the open interval (6, 7), but what about the range? Looking at the algebraic
expression for g given above, the range can be expressed as
11.3
Graphs
We can represent a function by drawing its graph which is the set of all points (x, y) in a plane where
y = f (x).
11.3.1
Example
Lets plot the previous two functions:

y
Figure 15: Graph of f (x) = x2 from Example 11.2.1
Figure 16: Graph of g(x) from Example 11.2.2.
11.5 Caution (non-functions)
11.3.2
133
Example
Graph the function f (x) = x2 4x + 3 on the axes in Figure 17.
Figure 17: f (x) = x2 4x + 3 = (x 1)(x 3)
11.4
Convention (domain)
An expression like the function y = 1 x2 means the function f with y = f (x) = 1 x2 . When
the domain is not specified it is taken to be the largest set on which the rule is defined. In this example,
the domain would be [1, 1].
11.5
Caution (non-functions)
Sometimes we see equations for curves which are not functions. Consider the equation of a circle of radius
1, centred at the origin: x2 + y 2 = 1. The crucial function property states that for each value x in the
domain there must correspond exactly one value y in the range. If there are any more y values, then it
is not a function.
11.6 Vertical line test
11.6
134
Vertical line test
In the graph of a function, the vertical line x =constant must cut the graph in at most one point. The
graph for the circle x2 + y 2 = 1 is given in Figure 18. We can clearly seethat the vertical line intersects
the circle at two points. In this case the two y values are given by y = 1 x2 . Therefore x2 + y 2 = 1
does not give rise to a function on any domain intersecting (1, 1).
y
Figure 18: The circle x2 + y 2 = 1 with a vertical line intersecting at two points. This shows that the
equation for a circle does not give rise to a function.
11.6.1
Example
Both y = 1 x2 (the top half of the circle) and y = 1 x2 (the bottom half of the circle) are
functions of x. What are the domains and ranges of these two functions?
11.7 Exponential functions
11.6.2
135
Example
Look at the graphs in Figures 19(a) and 19(b). Do either of these graphs represent functions?
y
(a)
(b)
Figure 19: Do either of these two graphs represent functions?
11.7
Exponential functions
An exponential function is one of the form f (x) = ax , where the base a is a positive constant, and
x is said to be the exponent or power. One very common exponential function which we shall see often in this course is given by f (x) = ex . It has the graph shown in Figure 20. Notice that it cuts
the y-axis (the line x = 0) at y = 1. This graph was made using MATLAB with the command
ezplot(exp(x),[-1.75,1.75,-4,6]).
exp(x)
6
5
4
3
2
1
0
1
2
3
4
1.5
0.5
0
x
0.5
1.5
Figure 20: The function f (x) = ex .
11.7 Exponential functions
11.7.1
136
Properties of exponents
Exponential functions are very useful for modelling many natural phenomena such as population growth
(base a > 1) and radioactive decay (base 0 < a < 1).
11.7.2
Example
This example is from Stewart, 5ed. p. 426. It does not appear in the 6th or 7th editions.
The half-life of the isotope strontium-90, 90 Sr, is 29 years. This means that half of any quantity of 90 Sr
will disintegrate in 29 years. Say the initial mass of a sample is 24mg, write an expression for the mass
remaining after t years.
11.8 Trigonometric functions (sin, cos, tan)
11.8
137
Trigonometric functions (sin, cos, tan)
In calculus, we measure angles in radians. One radian is defined as the angle subtended at the centre of
a circle of radius 1 by a segment of arc length 1.
1 rad
1
Figure 21: Angle in radians.

By definition, 2 radians = 360 . To convert between the two measures use the following formulae:
For x angle in radians, angle in degrees,
x
360
= ,
2
2
= x.
360
The point on the unit circle x2 + y 2 = 1 making an angle (in radians) anti-clockwise from the x-axis has
coordinates (cos , sin ), so cos2 + sin2 = 1. This defines the cos and sin functions. This relationship
is shown in Figure 22.
y
1
(cos , sin )
sin
x
cos
Figure 22: The relationship between the cosine and sine functions.
138
The graphs of the functions cos and sin are shown below in Figure 23. Note how the two functions
have the same behaviour, but are shifted (i.e. phase shifted)) by 2 . This leads to the relationship

cos = sin
= sin +
.
2
2
cos(x), sin(x)
cos(x)
sin(x)
0.5
0.5
1
6
0
x
Figure 23: The relationship between the cosine and sine functions.
In both cases in Figure 23 the range is [1, 1]. There are some unmarked vertical dotted lines. What
values do they represent? (This graph was made using MATLAB. To get both curves on the same axes,
we used the hold on command.)
The tan function is related to the sine and cosine functions by
It is defined for cos 6= 0. It is therefore not defined at values of which are odd integer multiples of
5
(eg 2 , 3
2 , 2 etc).
The graph of tan is given in Figure 24 below. Mark the x values in the graph where the function is not
defined.
tan(x)
6
4
2
0
2
4
6
6
0
x
Figure 24: Graph of tan .
139
Figure 25 shows another common function,

cos x
.
sin x
cot x =
cot (x)
6
6
6
Figure 25: Graph of cot .

The periodic nature of all the trigonometric functions we have seen makes them ideal for modelling
repetitive phenomena such as tides, vibrating strings, and various types of natural wave-like behaviour.
Both the sine and cosine functions have period 2. Hence
sin(x + 2) = sin x, and
cos(x + 2) = cos x.
Furthermore, both sin(x) and cos(x) have an amplitude of one.
We can speed up or slow down these
functions by tinkering with their periodicity. For example,
x
compare the functions sin(x) and sin 2 in Figure 26. Compare also the functions sin(x) and sin(2x) in
Figure 27.
sinx, sin(x/2)
sin(x)
sin(x/2)
1
0.8
0.6
0.4
0.2
0
0.2
0.4
0.6
0.8
1
8
0
x
Figure 26: Graph of sin(x) and sin
x
2
140
sinx, sin(2 x)
sin(x)
sin(2x)
1
0.8
0.6
0.4
0.2
0
0.2
0.4
0.6
0.8
1
8
0
x
Figure 27: Graph of sin(x) and sin(2x).

In addition, we can stretch or shrink trigonometric functions by multiplying these functions by constants
other than one. For example, Figure 28 shows the functions 21 cos(x) and 3 cos(x). What amplitudes do
these have?
1/2cos(x), 3 cos(x)
3
1/2 cos(x)
3cos(x)
3
8
Figure 28: Graph of
0
x
1
2
cos x and 3 cos(x).
Naturally, we can change both the

period and amplitude of trigonometric functions simultaneously. Figure
x
29 show the graphs of 4 cos 3 and 2 sin(5x).
4cos(x/3) and 2sin(5x)
5
4cos(x/3)
2sin(5x)
4
3
2
1
0
1
2
3
4
5
20
15
10
0
x
Figure 29: Graphs of 4 cos
10
x
3
15
20
and 2 sin(5x).
11.8.1
141
Example
Suppose that the temperature in Brisbane is periodic, peaking at 32 degrees Celsius on January 1, and
sinking to 8 degrees Celsius in June; see Figure 30. Use the cosine function to model temperature as a
function of time (in days).
35
30
temperature
(T)
25
20
15
10
50
100
150
200
time, (t)
250
300
350
Figure 30: Graph of temperature fluctuations in Brisbane.
11.10 One-to-one (1-1) functions - Stewart, 7ed. pp. 384-387; 6ed. pp. 385-388
11.9
142
Composition of functions
Let f and g be two functions. Then the composition of f and g, denoted f g, is the function defined by
11.9.1
Example
f (x) = x2 + 1, g(x) =
1
. Their compositions f g and g f are given by
x
Notice how these two composite functions are not equal. In general, f g 6= g f .
11.10
One-to-one (1-1) functions - Stewart, 7ed. pp. 384-387; 6ed. pp. 385-388
A function f : X Y is said to be one-to-one (usually written as 1-1) if, for all x1 , x2 X,
On the graph y = f (x), 1-1 means that any horizontal line y =constant cuts through the curve in at
most one place; see Figure 31.
11.10 One-to-one (1-1) functions - Stewart, 7ed. pp. 384-387; 6ed. pp. 385-388
y
143
y
(a)
(b)
Figure 31: Are either of these functions 1-1?
11.10.1
Example
Show from the definition that the function f (x) = e2x5 is 1-1 on its domain.
11.11 Inverse Functions
11.11
144
Inverse Functions
Let f : X Y be a 1-1 function. Then we can define the inverse function, f 1 : f (X) X by
By definition,
Note that y = f (x) x = f 1 (y), and that (f 1 (y))1 = f (x). The function f must be 1-1 in order
that f 1 satisfy the crucial function property. In order to obtain the graph of f 1 (x), we reflect the
graph of f (x) about the line y = x; for example, see Figure 32 below. The dashed line represents the
graph of the inverse function, having been reflected about the line y = x.
y
y=f
y=x
(x)
y=f(x)
x
Figure 32: Graph of f 1 (x).
11.11 Inverse Functions
11.11.1
145
Example
Draw the graph of the inverse function of f (x) = x3 (the cube root function f 1 (x) = x1/3 ) on the same
axes given below in Figure 33.
x3
2
1.5
1
0.5
0
0.5
1
1.5
2
2
1.5
0.5
0
x
0.5
1.5
Figure 33: f (x) = x3 . Draw in the graph of f 1 (x) = x1/3 .
11.11.2
Example
f : R R, f (x) = x2 is not 1-1 and therefore has no inverse. However, x 0 gives a 1-1 function
f : [0, ) R, f (x) = x2 with range [0, ). The inverse of this function is then f 1 : [0, ) [0, ),
f 1 (x) = x. Similarly the negative half of the function f (x) = x2 is 1-1, with inverse f 1 : [0, )
(, 0], f 1 (x) = x.
This technique is often used when the function is not 1-1 over its entire domain: just take a part where
it is 1-1 and determine the inverse for that part.
For further discussion on how to find the inverse function of a 1-1 function, see Stewart, 7ed. p. 387; 6ed.
p. 388 box 5.
11.13 Natural logarithm
11.12
146
Logarithms - Stewart, 7ed. p. 404; 6ed. p. 405
Logarithms are the inverse functions of the exponential functions.

ax
10
4x
3x
8
7
6
5
2x
4
3
2
1
0
1
4
0
x
Figure 34: Three plots of y = ax with a = 2, 3, 4.

From the graph of y = ax (a 6= 1 a positive constant), we see that it is 1-1 and thus has an inverse,
denoted loga x. From this definition we have the following facts:
loga (ax ) = x
x R,
aloga x = x
x > 0.
What is the domain and range of f (x) = loga (x)?
11.13
Natural logarithm
Now we set a = e (Eulers number = 2.71828. . .). The inverse function of f (x) = ex is
loge (x) ln x.
147
6
exp(x)
5
3
log(x)
2
3
3
Figure 35: Graphs of f (x) = ex and g(x) = ln x.
11.13.1
Properties
Using exponent laws, together with the fact that ln x is the inverse function of ex , we can prove the
following.
11.13.2
148
Example: Bacteria population - Stewart, 7ed. p. 409 qn 45, 6ed. p. 410 qn 43
If a bacteria population starts with 100 bacteria and doubles every 3 hours, then the number of bacteria
n after t hours is given by the formula
n = f (t) = 100 2t/3 .
(a) Find the inverse of this function and explain its meaning.
(b) When will the population reach 50000?
11.13.3
149
Example
Suppose that there are 1000 rabbits on January 1. Thirty days later there are 1500 rabbits. Assuming
that the population grows exponentially, find an expression for the rabbit population at time t.
11.14 Inverse trigonometric functions
11.14
150
Inverse trigonometric functions
The function y = sin x is 1-1 if we just define it over the interval [/2, /2]; see Figure 23. The inverse
function for this part of sin x is denoted arcsin x. Thus arcsin x is defined on the interval [1, 1] and takes
values in the range [/2, /2]. The graph can easily be obtained by reflecting the graph of sin x about
the line y = x over the appropriate interval; see Figure 36.
sin x, arcsin x
1.5
0.5
0.5
sin x
1
arcsin x
1.5
1.5
0.5
0.5
1.5
Figure 36: f (x) = sin x defined on [/2, /2] reflected about the line y = x to give f 1 (x) = arcsin x.
Similarly y = cos x is 1-1 on the interval [0, ] and its inverse function is denoted arccos x. The function
arccos x is defined on [1, 1] and takes values in the range [0, ]. Figure 37 below shows the graphs of
f (x) = cos x before and after reflection about the line y = x. This gives the graph of f 1 (x) = arccos x.
cos x, acos x
3
2.5
arccos x
2
1.5
1
0.5
0
cos x
0.5
1
1
0.5
0.5
1
x
1.5
2.5
Figure 37: f (x) = cos x defined on [0, ] reflected about the line y = x to give f 1 (x) = arccos x (written
in MATLAB as acos(x)).
151
Also, tan x is 1-1 on the open interval (/2, /2) with inverse function denoted by arctan x. Hence
arctan has the domain (, ) with values in the range (/2, /2).
tan x, atan(x)
10
tan x
8
6
4
2
0
arctan x
2
4
6
8
10
10
Figure 38: f (x) = tan x.
10
Notes
152
Notes
153
Notes
154
Notes
155
12.1 Definition: Limit - Stewart, 7ed. p. 50; 6ed. p. 66
12
156
Limits
Limits arise when we want to find the tangent to a curve or the velocity of an object, for example. Once
we understand limits, we can proceed to studying continuity and calculus in general. You should be
aware that limits are a fundamental notion to calculus, so it is important to understand them clearly.
12.1
Definition: Limit - Stewart, 7ed. p. 50; 6ed. p. 66
Let f (x) be a function and ` R. We say f (x) approaches the limit ` (or converges to the limit `) as x
approaches a if we can make the value of f (x) arbitrarily close to ` (as close to ` as we like) by taking x
to be sufficiently close to a (on either side of a) but not equal to a.
We write
lim f (x) = `.
xa
Roughly speaking, f (x) is close to ` for all x values sufficiently close to a, with x 6= a. The limit predicts
what should happen at x = a by looking at x values close to but not equal to a.
12.1.1
Precise Definition
Let f be a function defined on some open interval that contains the number a, except possibly a itself.
Then we write
lim f (x) = `
xa
if for every number > 0 there is a number > 0 such that

|f (x) `| < whenever 0 < |x a| < .
Detailed discussion on the precise definition of a limit can be found in Stewart, 7ed. p. 72; 6ed. p. 88.
We can sometimes determine the limit of a function by looking at its graph. Figure 39 shows two functions
with the same limit at x = 5.
157
14
14
12
12
10
10
10
Function 1
10
Function 2
Figure 39: Two functions with limit equal to 10 as x 5.

Function 1 is described by

2x
2x + 20
for 0 x 5,
for 5 < x 10,
2x
2
g(x) =
2x + 20
for 0 x < 5,
for x = 5,
for 5 < x 10.
f (x) =
while function 2 is described by
Each function has a limit of 10 as x approaches 5. It does not matter that the value of the second
function is 2 when x equals 5 since, when dealing with limits, we are only interested in the behaviour of
the function as x approaches 5.
12.1.2
158
Properties
Suppose that c is a constant and the limits ` = lim f (x) and m = lim g(x) exist for some fixed a R.
xa
xa
Then
12.1.3
Example
Find the value of lim (x2 + 1).

x1
12.1.4
Example: Stewart, 7ed. p. 51 ex. 1; 6ed. p. 67 ex. 1

x1
.
x1 x2 1
Determine the value of lim
159
12.2 One-sided limits
12.1.5
160
Definition: Infinite Limits
Let f be a function defined on both sides of a, except possibly at a itself. Then

lim f (x) =
xa
means that the values of f (x) can be made arbitrarily large by taking x sufficiently close to a, but not
equal to a.
Similarly,
lim f (x) =
xa
means that the values of f (x) can be made arbitrarily large negatively by taking x sufficiently close to
a, but not equal to a.
In these cases, we say that f (x) diverges to . We also say that the limit does not exist in these cases.
Note that the limit properties in section 12.1.2 do not necessarily apply if the limits diverge.
12.2
One-sided limits
Consider the piecewise function

1, x 0
2, x < 0.
f (x) =
This function has the graph depicted in Figure 40.

y
Figure 40: The limit of this function as x 0 does not exist.
Notice that
lim
x0,x>0
f (x) = 1, but
lim
x0,x<0
f (x) = 2. Therefore, the limit as x 0 does not exist. We
can, however, talk about the one-sided limits.

In the above example, we say that the limit as x 0 from above (or from the right) equals 1 and we
write
lim f (x) = 1.
x0+
12.2 One-sided limits
161
Similarly, we say that the limit as x 0 from below (or from the left) equals 2 and we write
lim f (x) = 2.
x0
In general, for lim f (x) = `, just consider x with x > a and similarly for lim f (x) = `, consider only
xa+
xa
x < a.
12.2.1
Example
Consider both lim
x2+
12.2.2
x 2 and lim
x2
x 2. Based on your conclusion, does lim
Theorem
lim f (x) = ` if and only if
xa
12.2.3
Example

Find lim f (x) where f (x) =

x1
x2 , x 1,
2 x, x < 1.
x2
x 2 exist?
12.3 Theorem: Squeeze principle
12.3
162
Theorem: Squeeze principle
Suppose
lim g(x) = ` = lim h(x)
xa
xa
and, for x close to a (x 6= a)

h(x) f (x) g(x).
Then
See the graph in Figure 41.

y
g(x)
f(x)
l
h(x)
a
Figure 41: The squeeze principle
12.4 Limits as x approaches infinity
12.3.1
Example
Prove that lim x2 sin

x0
12.4
163

1
= 0.
x
Limits as x approaches infinity
We say that f (x) approaches ` as x if f (x) is arbitrarily close to ` for all large enough values of x.
Formally, we write:
That is, f (x) can be made arbitrarily close to ` by taking x sufficiently large. Similarly, we write
if f (x) approaches ` as x becomes more and more negative.
164
Note:
1
= 0 and
x x
lim
1
= 0.
x x
lim
1/x
10
8
6
4
2
0
2
4
6
8
10
10
0
x
10
Figure 42: Graph of f (x) = 1/x.
12.4.1
Example
Determine lim sin x, or show that it does not exist.

x
12.4.2
Example: ratio of two polynomials, numerator degree = denominator degree
2x2 + 3
.
x 3x2 + x
Find lim
12.4.3
Example
8x3 + 6x2
.
x 4x3 x + 12
Find lim
12.4.4
Example: ratio of two polynomials, numerator degree > denominator degree
x2 + 5
.
x x + 1
Find lim
165
12.5 Some important limits
12.4.5
166
Example: ratio of two polynomials, numerator degree < denominator degree
x+1
x x2 + 1
Find lim
The general method in the four previous examples has been to:
1. divide the numerator and denominator by the highest power of x in the denominator; and then
2. determine the limits of the numerator and denominator separately.
12.5
Some important limits
The following limits are fundamental. Combined with the properties given in 12.1.2 and the Squeeze
Principle given in 12.3, these will enable you to compute a range of other limits.
Here a R is arbitrary.
(1)
(4)
lim 1 = 1
(2)
1 cos(x)
1
=
2
x0
x
2
(5)
xa
lim
lim x = a
xa
ex 1
= 1.
x0
x
lim
(3)
sin x
=1
x0 x
lim
Notes
167
Notes
168
Notes
169
13.2 Representations
13
170
Sequences
A sequence {an } is a list of numbers with a definite order:

a0 , a1 , a2 , a3 , . . . , an , . . .
In fact, sequences can contain a finite numbers of terms or may continue forever. Further, a sequence
does not have to consist of numbers, but might contain any sort of mathematical structure, or indeed
any other objects.
13.1
Formal Definition: Sequence
More formally, a sequence is a function, with domain being N0 = {0, 1, 2, 3, . . . }. We can also take the
domain as N = {1, 2, 3, ...} and start the sequence at a1 rather than a0 .
13.1.1
Motivation
Infinite sequences of numbers are useful in many applications. The sequence might represent approximations to the solution of a problem such as the bisection method. Or the sequence might represent a time
series, where the numbers represent the population each year. Mathematically, we are interested in how
the sequence behaves as the number of terms becomes large: does it converge to a solution, or how fast
does the population grow? Sequences can also appear from the partial sums in an infinite series, which
comprises our next major topic.
13.2
Representations
There are two main ways to represent the nth term in a sequence. Firstly, there is a direct (or closed
form or functional) representation. Secondly, there is a recursive (or indirect) representation.
13.2 Representations
171
The recursive form can have a simpler structure. It might be that an = f (an1 , an2 ) or even an =
f (an1 ). Depending on the problem, it can be natural to express the formula in one or the other way.
13.2.1
Example
List the first terms of the sequence an =
13.2.2
n
.
n+1
Example
List the first terms of the recursive sequence an+1 =
1
with a1 = 2.
3 an
Note that for the recursive form, you must express one or more initial conditions or starting values. The
recursive form is actually a difference equation, and there are similarities with differential equations that
are studied in MATH1052.
13.2.3
Example: Fibonacci sequence
The Fibonacci sequence is defined through
an = f (an1 , an2 ) = an1 + an2

with a0 = 0, a1 = 1. List the first 9 terms.
13.3 Application: Rabbit population - Stewart, 7ed. p. 726 ex. 83; 6ed. p. 722 ex. 71
13.3
172
Application: Rabbit population - Stewart, 7ed. p. 726 ex. 83; 6ed. p. 722
ex. 71
Suppose that rabbits live forever and that every month each pair produces a new pair which becomes
productive at age 2 months. Starting with one newborn pair, how many pairs of rabbits will be there in
the nth month?
13.4 Limits
13.4
173
Limits
Let {an }
n=0 be a sequence. Then
lim an = `
` R,
means :
an approaches ` as n gets larger and larger, ie. an is close to ` for n sufficiently large.
This is similar to lim f (x) = `, except here an is defined only for natural numbers n.
x
13.4.1
Convention
If a sequence {an }
n=0 has limit ` R, we say that an converges to ` and that the sequence {an }n=0 is
convergent. Otherwise the sequence is divergent.
13.4.2
Examples
Determine if the sequences {an }, with an as given below, are convergent and if so find their limits.
1. an =
1
n
1
2. an = an1 , a0 = 1.
2
13.5 Theorem: Limit laws
174
3. an = (1)n
1
4. an = rn , for r = , r = 1, r = 2
2
13.5
Theorem: Limit laws
The following limit laws apply provided that the separate limits exist (that is {an } and {bn } are convergent):
13.7 Useful sequences to remember
13.6
175
Theorem: Squeeze
If an bn cn for n n0 for some n0 N and lim an = lim cn = `, then

n
13.6.1
Example
Use the squeeze theorem on {an }, where an =
13.7
1
sin(n).
n
Useful sequences to remember

n
(1) For constant c, lim c =

n
0,
1,
if |c| < 1
if c = 1.
Sequence {cn }
n=0 is divergent if c = 1 or |c| > 1.
(2) For constant c > 0, lim c1/n = lim n c = 1.

n
(3) lim
n nr
= 0 for r > 0.
(4) lim n1/n = lim

n
(5) lim
n = 1.
1
= 0.
n!
cn
= 0.
n n!
(6) For constant c, lim

1 n
(7) lim 1 +
= e.
n
n

a n
(8) lim 1 +
= ea .
n
n
13.8 Application: Logistic sequence
13.8
176
Application: Logistic sequence
Sequences which are bounded, but not convergent can be very interesting. The study of chaos appears
in these sequences. A famous recursive sequence is the logistic sequence from population dynamics:
For certain values of the coefficient k, the behaviour is regular: convergent, or periodic. But other values
of k yield chaotic sequences which cannot be predicted. Figure 43 has some examples of the logistic
sequence when k = 2.9, 3.5 and 3.9.
Figure 43: Logistic sequence, a0 = 12 , k = 2.9, 3.5, 3.9
Notes
177
Notes
178
14.1 Definition: Continuity
14
14.1
179
Continuity
Definition: Continuity
We say that a function f is continuous at a if
If f is not continuous at a, we say that f is discontinuous at a, or f has a discontinuity at a.
y
y = f(x)
f(x)
approaches
f(a)
f(a)
a
as x approaches
x
a
Figure 44: Graphical representation of continuity at x = a.
A function may not be continuous at x = a for a number of reasons.

14.1.1
Example
Let

f (x) =
x, x 6= 0
1, x = 0.
Then f (x) has a discontinuity at x = 0. This is because f (0) = 1 while lim f (x) = 0. Hence Condition
x0
(iii) of Definition 14.1 doesnt hold. See Figure 45.
14.1 Definition: Continuity
180
y
Figure 45: An example of a discontinuous function with discontinuity at x = 0.
14.1.2
Example
The function f (x) =
1
(see Figure 46) is not continuous at x = 0, since
x2
1/(x2)
5
f(x)
0
2
0
x
Figure 46: The function f (x) = 1/x2 has a discontinuity at x = 0.
14.1.3
Example

f (x) =
is not continuous at x = 0 since
x + 1, x 0
x2 , x < 0
14.2 Definition: Continuity on intervals
14.2
181
Definition: Continuity on intervals
We say that f is continuous on the open interval (a, b), if
If f is continuous on the closed interval [a, b], then f is continuous on (a, b) and
You could think of a continuous function being one that on an interval can be drawn without lifting your
pen.
14.2.1
Examples
Any polynomial in x is continuous on R. For example, f (x) = ax2 + bx + c is continuous on R.

ex , |x|, sin x, cos x and arctan x are continuous on R.
f (x) = ln x is continuous on (0, ).
14.3 Properties
14.3
182
Properties
If f (x) and g(x) are continuous at x = a and c is a constant, then
14.3.1
Theorem: Limit of a composite function - Stewart, 7ed. p. 87 Thm 8; 6ed. p. 103

Thm 8
If f is continuous at b and lim g(x) = b, then lim f (g(x)) = f (b). In other words,
xa
xa
Similarly if lim g(x) = b and f is continuous at x = b, then

x
lim f (g(x)) = f (b).
14.3.2
Corollary: Stewart, 7ed. p 88 Thm 9; 6ed. p. 103 Thm 9
If g is continuous at a and f is continuous at g(a) then f g is continuous at a.
14.5 Application of IVT (bisection method)
14.4
183
Intermediate Value Theorem (IVT) - Stewart, 7ed. p. 89; 6ed. p. 104
Suppose that f is continuous on the closed interval [a, b] and let N be any number between f (a) and
f (b), where f (a) 6= f (b). Then there exists a number c (a, b) such that f (c) = N .
y
f(b)
N
f(a)
x
a
Figure 47: The IVT states that a continuous function takes on every intermediate value between the
function values f (a) and f (b). That is, we can choose any N between f (a) and f (b) which corresponds
to a value c between a and b. Note that this is not necessarily true if f is discontinuous.
14.4.1
Example
Suppose that a function f is continuous everywhere and that f (2) = 3,

f (1) = 1, f (0) = 4, f (1) = 1, and f (2) = 5. Does the Intermediate-Value Theorem guarantee that f
has a root on the following intervals?
a) [2, 1]
14.5
b) [1, 0]
c) [1, 1]
d) [0, 2]
e) [1, 3]
Application of IVT (bisection method)
The bisection method is a procedure for approximating the zeros of a continuous function. It first cuts
the interval [a, b] in half (say, at a point c), and then decides in which of the smaller intervals ([a, c]
or [c, b]) the zero lies. This process is repeated until the interval is small enough to give a significant
approximation to the zero itself.
We can present this bisection method as an algorithm:
(1) Given [a, b] such that f (a)f (b) < 0, let c =
(2) If f (c) = 0 then quit; c is a zero of f .
a+b
.
2
184
(3) If f (c) 6= 0 then:

(a) If f (a)f (c) < 0, a zero lies in the interval [a, c). So replace b by c.
(b) If f (a)f (c) > 0, replace a by c.
(4) If the interval [a, b] is small enough to give a precise enough approximation then quit. Otherwise,
go to step (1).
y
f(b)
a
b
f(a)
Figure 48: The bisection method is an application of the IVT.

Note that:
f (a)f (c) < 0
f (a)f (c) > 0
f (a) < 0 and f (c) > 0
or
f (a) > 0 and f (c) < 0
f (a) > 0 and f (c) > 0
or
f (a) < 0 and f (c) < 0
We will investigate the bisection method in more detail in the lab sessions with MATLAB.
Notes
185
Notes
186
15.1 Tangents
15
187
Derivatives
Finding the instantaneous velocity of a moving object and other problems involving rates of change are
situations where derivatives can be used as a powerful tool. All rates of change can be interpreted as
slopes of appropriate tangents. Therefore we shall consider the tangent problem and how it leads to a
precise definition of the derivative.
15.1
Tangents
y
y=f(x)
B=(b, f(b))
f(b) f(a)
A=(a, f(a))
h
x
a
b=a+h
Figure 49: We want to determine the tangent line at the point A.

Consider the graph in Figure 49. We want to determine the tangent line at the point A (x = a) on the
graph of y = f (x), where f (x) is nice enough. We approximate the tangent at A by the chord AB
where B is a point on the curve (close to A) with x-value a + h, where h is small. By looking at the
graph, we can see that the slope m of the chord AB is given by
As the point B gets closer to A, m will become closer to the slope at the point A. To obtain this value,
we take the limit as h 0:
15.2 Definition: Derivative, differentiability
15.2
188
Definition: Derivative, differentiability
The derivative of f at x is defined by
We say that f is differentiable at some point x if this limit exists. Further, we say that f is differentiable
on an open interval if it is differentiable at every point in the interval. Note that f 0 (a) is the slope of the
tangent line to the graph of y = f (x) at x = a.
We have thus defined a new function f 0 , called the derivative of f . Sometimes we use the Leibniz notation
dy
df
or
in place of f 0 (x).
dx
dx
Note that if f is differentiable at a, there holds:
f 0 (a) = lim
xa
15.2.1
f (x) f (a)
.
xa
Example
Using the definition of the derivative (i.e., calculating by first principles), find the derivative of f (x) =
x2 + x.
15.2 Definition: Derivative, differentiability
15.2.2
Example
Find the derivative from first principles of f (x) = ex .
15.2.3
Some useful derivatives
Other useful derivatives are
189
15.4 Rates of change - Stewart, 7ed. p. 108; 6ed. p. 117
15.3
190
Theorem: Differentiability implies continuity
If f is differentiable at a, then
The converse, however, is false. That is, if f is continuous at a, then it is not necessarily differentiable.
(See the following example.)
15.3.1
15.4
Example
Rates of change - Stewart, 7ed. p. 108; 6ed. p. 117
Recall that the slope of the tangent line to y = f (x) at x = a measures the rate of change of y with
respect to x at the point x = a.
15.4.1
Example: Velocity
Velocity v is the rate at which displacement s changes with respect to time t, so
Similarly acceleration a is the rate at which velocity v changes with respect to time t, so
15.4 Rates of change - Stewart, 7ed. p. 108; 6ed. p. 117
15.4.2
191
Example (blood flow in an artery) - Stewart, 7ed. p. 170 ex. 7; 6ed. p. 176 ex. 7
Blood flowing through a blood vessel can be described by the law of laminar flow. We model the
artery or vein as a cylinder. The closer the blood is to the artery wall, the slower it will flow due to
friction. Conversely, the velocity of the blood is greatest at the centre of the blood vessel (cylinder). In
mathematical terms, this law can be stated
v=
P
(R2 r2 ),
4l
where is the viscosity of the blood, P is the pressure difference between the ends of the artery, l is the
length of the blood vessel, R is the radius of the vessel and r is the distance from the centre of the vessel
(the only variable in this case); see Figure 50.
We now seek to find a general expression for the velocity gradient of the blood, that is, an expression for
the rate of change of velocity with respect to r:
Figure 50: The closer the blood is to the vein wall, the smaller its velocity.
This tells us at what rate the velocity of the blood decreases the further it is from the centre of the blood
vessel.
These types of problems are studied in later courses on fluid mechanics and other related topics.
15.5 Higher derivatives
15.5
192
Higher derivatives
Differentiating a function y = f (x) n times, if possible, gives the nth derivative of f , usually denoted
dn f
dn y
f (n) (x),
or
.
dxn
dxn
Therefore acceleration can be represented as
Sometimes when we deal with time derivatives we write v(t) = s(t)
or a(t) = s(t) and so on.
15.6 Rules for differentiation
15.6
193
Rules for differentiation
Suppose that f and g are differentiable functions, and let c be a constant. Then we have the following
rules:
In the Leibniz notation, if f and g are both differentiable functions, and c is a constant, then the rules
for differentiation can be expressed as:
15.7 The chain rule - Stewart, 7ed. p. 148; 6ed. p. 156
15.6.1
Example
Use the rules above to find the derivative of f (x) = tan x =
15.7
194
sin x
.
cos x
The chain rule - Stewart, 7ed. p. 148; 6ed. p. 156
If g and h are both differentiable and f = g h is the composite function defined by f (x) = g(h(x)), then
f is differentiable and f 0 is given by the product
In the Leibniz notation, if y = g(u) and u = h(x) are both differentiable functions, then
15.7 The chain rule - Stewart, 7ed. p. 148; 6ed. p. 156
15.7.1
Example
Differentiate y =
15.7.2
195
sin x.
Example: Rocket tracking application
A rocket is launched vertically and tracked by an observation station 5km from the launch pad. The
angle of elevation is observed to be increasing at 3 /s when = 60 . What is the velocity of the rocket
at that instant?
15.8 Derivative of inverse function
15.8
196
Derivative of inverse function
Suppose y = f 1 (x), where f 1 is the inverse of f . To obtain
dy
we use
dx
x = f (f 1 (x)) = f (y).
Differentiating both sides with respect to x using the chain rule gives
15.8.1
Example
Find the derivative of y = ln x.
15.8 Derivative of inverse function
15.8.2
Example
Find the derivative of y = arcsin x
15.8.3
Other useful derivatives
197
15.9 LH
opitals rule - Stewart, 7ed. p. 470; 6ed. p. 472
15.9
198
LH
opitals rule - Stewart, 7ed. p. 470; 6ed. p. 472
Suppose that f and g are differentiable and g 0 (x) 6= 0 near a (except possibly at a). Suppose that
lim f (x) = 0 and lim g(x) = 0
xa
xa
or
lim f (x) = and lim g(x) = .
xa
Then
if the limit on the right exists.

15.9.1
Example
Find lim
ln x
.
x1
x1
xa
15.10 Theorem: Continuous extension
15.9.2
199
Example
Find lim x ln x.
x0+
15.10
Theorem: Continuous extension
Sometimes LH
opitals rule can be used to evaluate limits of sequences. Let f be a function on the real
numbers such that lim f (x) exists. Let f (n) = an for natural numbers n. Then:
x
15.10.1
Example
ln n
.
n n
Evaluate lim
15.11 The Mean Value Theorem (MVT)
15.11
200
The Mean Value Theorem (MVT)
Let f be continuous on [a, b] and differentiable on (a, b). Then

f (b) f (a)
= f 0 (c)
ba
for some c, where a < c < b.
(a)
is the slope of the chord joining A = (a, f (a))
Note f 0 (c) is the slope of y = f (x) at x = c and f (b)f
ba
to B = (b, f (b)). (Stewart, 7ed. p. 208; 6ed. p. 216)
15.11.1
Example
Without actually finding them, show that the equation

x4 + 4x + 1 = 0
has exactly two real solutions for x.
15.13 Increasing/Decreasing test - Stewart, 7ed. p. 214; 6ed. p. 221
15.12
Definition: Increasing/Decreasing
A function f is called strictly increasing on an interval I if
while f is called strictly decreasing on I if
This leads to the following test.
15.13
Increasing/Decreasing test - Stewart, 7ed. p. 214; 6ed. p. 221
Suppose that f is continuous on [a, b] and differentiable on (a, b).
201
15.14 Definition: Local maxima and minima
15.13.1
Example
Test to see on what intervals the function f (x) = x3 + x is increasing or decreasing.
15.14
Definition: Local maxima and minima
A function f has a local maximum at a if
Similarly, f has a local minimum at b if
15.14.1
Definition (critical point)
A function f is said to have a critical point at x = a, a dom(f ) if
202
15.14 Definition: Local maxima and minima
15.14.2
203
Definition: Global maximum and minimum
Let f be a function defined on the interval [a, b], and c [a, b].
Then f has a global maximum at c if
Similarly f has a global minimum at c if
15.14.3
Example
Figure 51: What do the points x = p, q, r, s, t, u represent?

Consider the function in Figure 51. The domain of this function is the interval [p, u]. Which points represent the critical points? Which points represent the maximum and minimum? Which points represent
local maxima and minima?
15.15 First derivative test
15.14.4
Theorem
If f has a local maximum/minimum at x = c and f 0 (c) exists then f 0 (c) = 0.

Proof (for max):
15.15
First derivative test
Let f be continuous on [a, b] and differentiable on (a, b), and let c (a, b).
(a) If f 0 (x) > 0 for a < x < c and f 0 (x) < 0 for c < x < b
(b) If f 0 (x) < 0 for a < x < c and f 0 (x) > 0 for c < x < b
Proof:
204
15.16 Second derivative test
15.16
205
Second derivative test
Suppose f 00 exists and is continuous near c.

(a) If f 0 (c) = 0 and f 00 (c) > 0, then
(b) If f 0 (c) = 0 and f 00 (c) < 0, then
(c) If f 0 (c) = 0 and f 00 (c) = 0, then the test fails and we get no useful information. In this case the first
derivative test should be used.
15.16.1
Example
Find local minima and maxima of f (x) = 2x3 + 3x2 12x + 4 on R.
15.16.2
Example
Find local extrema of f (x) = x4 .
206
15.16.3
207
Example: Beehive cell - Stewart, 7ed. p. 258 qn 45; 6ed. p. 264 qn 43
In a beehive, each cell is a regular hexagonal prism, open at one end with a trihedral angle at the other
end (see Figure 52). It is believed that bees form their cells in such a way as to minimise the surface area
for a given volume, thus using the least amount of wax in cell construction. Examination of these cells
has shown that the measure of the apex angle is amazingly consistent. Based on the geometry of the
cell, it can be shown that the surface area S is given by
S() = 6sh +
3s2
( 3 cos )
2 sin
where s, the length of the sides of the hexagon, and h, the height, are constants.
Figure 52: The beehive cell. The open front is pictured on the bottom and the closed end with the
trihedral angle on top.
(a) Calculate
dS
d
(b) What angle should the bees prefer?

Actual measurements of the angle in beehives have been made, and the measures of these angles seldom
differ from the calculated value by more than 2 .
208
15.17 Extreme value theorem - Stewart, 7ed. p. 199; 6ed. p. 206
15.17
Extreme value theorem - Stewart, 7ed. p. 199; 6ed. p. 206
If f is continuous on a closed interval [a, b], then
15.17.1
Find the global maximum and minimum of the function f (x) = x3 3x2 + 1 on the interval [ 12 , 4].
209
15.18 Optimization problems
15.18
15.18.1
210
Optimization problems
Example: Optimal rowing
A rowing team launches a boat from point A on a bank of a straight river 3km wide, and wants to reach
point B, 8km downstream on the opposite bank as quickly as possible; see Figure 53. They could row
their boat directly across the river to a point C and then run to B, or they could row directly to B, or
they could row to some point D between C and B and then run to B. If they can row 6km/h and run
8km/h, where should they land to reach B as soon as possible? (Assume that the speed of the water is
negligible compared to the speed at which they row.)
8km
C
3km
Figure 53: How should the team travel from A to B in the shortest possible time?
211
15.18.2
212
Example: Optimal farming
Some farmers have a certain amount of fencing to build a rectangular enclosure which is to be subdivided
into 3 parts by 2 fences parallel to one of the sides. How should they do it so the area enclosed is
maximised?
15.18.3
213
Example: Newtons method
Newtons method is a numerical procedure to approximate the roots of functions. Most programs which
find roots of functions are based on this method. For example, consider the equation
tan x = x
How would you find the roots of this equation? It is impossible to solve analytically (ie, precisely), but
Newtons method gives us a way of approximating the roots to some degree of accuracy.
The idea behind the geometry of this method is given in Figure 54. We must have a starting guess (the
value x1 ), then we find the tangent line to the curve at that point, and find where the tangent line cuts
the x-axis. This gives us the next point. We repeat the process until we have a close enough estimate.
x2
x1
Figure 54: Starting at x1 , this gives you an idea of how Newtons method iterates to estimate the root p.
Mathematically, the slope of the tangent line to the curve at the point (x1 , f (x1 )) is given by
f 0 (x1 ) =
f (x1 ) 0
x1 x2
x2 = x1
f (x1 )
.
f 0 (x1 )
Repeating this procedure leads to the iterative formula
If the numbers xn become closer and closer to p as n becomes large, we say the sequence converges to p.
This method is very sensitive to the type of curve and the choice of starting point.
We will investigate Newtons method in more detail in the lab sessions using MATLAB.
Notes
214
Notes
215
Notes
216
16.2 Motivation
16
16.1
217
Series
Infinite sums (notation)
If we have an infinite sum we write

a0 + a1 + a2 + . . . + an + . . . =
an .
n=0
Note that the lower bound (n = 0) of the sum may vary.
16.2
Motivation
Series come from many fields.

1. Approximation to problem solutions:
a0 zeroth order approximation
a0 + a1 (a1 small) first order approximation
a0 + a1 + a2 (a2 very small) second order approximation
a0 + a1 + a2 + . . . + an nth order
a0 + a1 + a2 + . . . + an + . . . exact solution, provided the series converges
2. Current state of a process over infinite time horizon
3. Approximating functions via Taylor/Fourier Series, e.g.,
1
1
e x = 1 + x + x2 + x3 + . . . ,
2
3!
f (t) = c0 + c1 sin t + c2 sin(2t) + c3 sin(3t) + . . .
4. Measure the volume in a fractal process, like the amount of material in a botanic structure (Figure
55).
Figure 55: This is a graphical representation of a botanic fractal.

5. Riemann sums
16.3 Definition: Convergence - Stewart, 7ed. p. 729; 6ed. p. 724
16.3
218
Definition: Convergence - Stewart, 7ed. p. 729; 6ed. p. 724
Given a series
an = a0 + a1 + a2 + . . ., let sn denote its nth partial sum:
n=0
If the sequence {sn } is convergent (i.e.
lim sn = s with s R), then the series
convergent and we write

lim sn =
an is said to be
n=0
an = s.
n=0
The number s is called the sum of the series. Otherwise the series is said to be divergent.
16.3.1
Example
Does the series
(1)n converge or diverge?
n=0
16.4 Theorem: p test
16.3.2
219
Example
Show that the series
X
n=0
16.4
1
is convergent.
(n + 1)(n + 2)
Theorem: p test
For p R, the p-series
X
1
is
np
n=1
Note the above sum is from n = 1. This is just a matter of taste since
X
X
1
1
=
p
n
(n + 1)p
n=1
n=0
16.6 Test for divergence (nth term test)
16.5
Theorem: nth term test - Stewart, 7ed. p. 733; 6ed. p. 728
16.6
Test for divergence (nth term test)
If lim an 6= 0 then the series is divergent.

n
16.6.1
Examples
Verify that these series diverge by the nth term test.

(i)
(1)n
n=0
(ii)
X
n=1
n2
5n2 + 4
220
16.8 Example: Geometric series
16.6.2
221
Caution
If lim an = 0 then we cannot say whether it is convergent or divergent.

n
For example,
1
=0
n n
lim
but the harmonic series
X
1
1 1 1
= 1 + + + + ...
n
2 3 4
n=1
diverges (p-series with p = 1).
16.7
Proof that the Harmonic Series Diverges
Write the partial sums

s1 = 1
1
2
1 1
= 1+ +
2 3
1 1 1
= 1+ + +
2 3 4
1 1 1
1+ + +
2 |4 {z 4}
1
1
= 1+ +
2
2
s2 = 1 +
s3
s4
s8

1 1 1
1 1 1 1
= 1+ + + +
+ + +
2 3 4
5 6 7 8

1 1 1
1 1 1 1
1+ + + +
+ + +
2 4 4
8 8 8 8
1
1
1
= 1+ +
+
.
2
2
2
In general s2n 1 + n/2, so the partial sums approach infinity as n .
16.8
Example: Geometric series
The series
arn
n=0
is convergent if |r| < 1 with
X
n=0
arn =
a
and divergent if |r| 1.
1r
16.8 Example: Geometric series
16.8.1
Proof
222
16.9 Application: Bouncing ball - Stewart, 7ed. p. 736 qn 72; 6ed. p. 731 qn 58
16.8.2
223
Intuition
If all the terms in the series stay large, then the series is divergent. If the individual terms are becoming
smaller then the series may converge, but not necessarily.
16.9
Application: Bouncing ball - Stewart, 7ed. p. 736 qn 72; 6ed. p. 731 qn 58
A certain ball has the property that each time it falls from a height h onto a hard, level surface, it
rebounds to a height rh, where 0 < r < 1. Suppose that the ball is dropped from an initial height of H
metres.
(a) Find the total distance that the ball travels.
16.9 Application: Bouncing ball - Stewart, 7ed. p. 736 qn 72; 6ed. p. 731 qn 58
224
(b) Calculate the total time that the ball travels. Use the fact that the ball falls gt2 /2 metres in t seconds
(from classical physics).
16.11 Theorem: Comparison test - Stewart, 7ed. p. 746; 6ed. p. 741
16.10
225
Comparison tests
If we are trying to determine whether or not
an converges, where an contains complicated terms,
n=0
then it may be possible to bound the series by using simpler terms. If |an | is always smaller than |bn |
X
X
and
|bn | converges, then
an should also. On the other hand, if an is always large - in fact larger
n=0
than |bn |, and
n=0
bn diverges, then
n=0
16.11
an should diverge too.
n=0
Theorem: Comparison test - Stewart, 7ed. p. 746; 6ed. p. 741
Suppose that
an and
n=0
16.11.1
bn are series with all non-negative terms.
n=0
Example
Test the series
X
ln n
n=1
for convergence.
16.12 Alternating series
16.11.2
Example
Test the series
X
n=0
16.12
226
5
for convergence.
2 + 3n
Alternating series
The tests for convergence have so far required terms to be positive. However, if terms alternate in sign
+, , +, , . . . (or , +, , +, . . .) there are some special results. We call such a series an alternating
series.
Here is quite an interesting result:
We have seen that the harmonic series
1+
1 1 1
+ + + ...
2 3 4
is divergent. However, the series

1 1 1
+ + ...
2 3 4
is convergent. Why? Later, we shall see that this series converges to ln 2.
1
16.13 Theorem: Alternating series test - Stewart, 7ed. p. 751; 6ed. p. 746
16.13
Theorem: Alternating series test - Stewart, 7ed. p. 751; 6ed. p. 746
If the alternating series
(1)n bn = b0 b1 + b2 b3 + . . . (all bn > 0)
n=0
satisfies
then the series is convergent.

16.13.1
Example
(1)n
n=0
1
n+1
227
16.13 Theorem: Alternating series test - Stewart, 7ed. p. 751; 6ed. p. 746
16.13.2
Example
(1)n
n=1
3n
4n 1
228
16.14 Definition: Absolute/conditional convergence
16.14
16.14.1
229
Definition: Absolute/conditional convergence
Examples
Are the following series conditionally or absolutely convergent?
(i)
(1)n
n=1
1
n
16.14 Definition: Absolute/conditional convergence
230
(ii)
(1)n
n=1
(iii)
1
n4
X
cos n
n=1
n2
16.15 Theorem: Ratio test - Stewart, 7ed. p. 758; 6ed. p. 752
16.14.2
231
Theorem: Absolute convergence
If a series is absolutely convergent, then it is convergent.

16.14.3
16.15
Proof
Theorem: Ratio test - Stewart, 7ed. p. 758; 6ed. p. 752
A powerful test for convergence of series is the ratio test. We will use it extensively in the following
sections.

X
an+1

= L < 1, then the series
1. If lim
an is absolutely convergent (and therefore convergent).
n
an
n=1
X
an+1
an+1

2. If lim
= L > 1 or lim
= , then the series
an is divergent.
n
n
an
an
n=1

an+1

= 1, then the ratio test is inconclusive.
3. If lim
n
an

an+1

is not defined, then the ratio test is inconclusive.
4. If lim
n
an
Note that the ratio test only tests for absolute convergence. The ratio test does not give any estimates
of the sum.
16.15.1
232
Examples
Use the ratio test to determine, if possible, whether the following series are absolutely convergent.
X
(i)
arn
n=0
(ii)
(1)n
n=1
n3
3n
Notes
233
17.1 Definition: Power series
17
17.1
234
Power series and Taylor series

Definition: Power series
A power series is a series of the form
Such series have many applications, for example, solutions of higher order O.D.Es (ordinary differential
equations) with non-constant coefficients, which in turn are very important to many areas of science and
engineering.
17.1.1
Example
Use the ratio test to determine for which values of x the following series converges:
X
(x 3)n
n=1
17.1 Definition: Power series
17.1.2
235
Example: Bessel function
Use the ratio test to determine for which values of x the Bessel function of order zero (denoted J0 (x))
converges, where
X
(1)n x2n
J0 (x) =
.
22n (n!)2
n=0
Bessel functions first arose when Bessel solved Keplers equation for describing planetary motion. Since
then, they have been used in applications such as temperature distribution on a circular plate and the
shape of a vibrating membrane (such as the human ear).
The graph of J0 (t) is given in Figure 56.
besselj(0,t)
1.5
0.5
0.5
1
0
10
15
t
20
25
30
Figure 56: The Bessel function J0 (t) is represented as besselj(0,t) in MATLAB.
17.2 Theorem: Defines radius of convergence - Stewart, 7ed. p. 767; 6ed. p. 761
17.2
236
Theorem: Defines radius of convergence - Stewart, 7ed. p. 767; 6ed. p. 761
For a given power series
cn (x a)n there are only three possibilities:
n=0
In case (1), we say that the radius of convergence is 0, while in case (2) we say that the radius of
convergence is .
The interval of convergence of a power series is the interval that consists of all values of x for which the
series converges. In case (1) the interval is a single point, a. In case (2) the interval is (, ). In case
(3) the interval is either (a r, a + r), (a r, a + r], [a r, a + r) or [a r, a + r].
17.3 Taylor series
17.2.1
237
Example
What is the radius of convergence for the series
X
(3)n xn
? For which values of x does the series

n+1
n=0
converge?
17.3
Taylor series
Given a function f (x) it is sometimes possible to expand f (x) as a power series

f (x) =
cn (x a)n = c0 + c1 (x a) + c2 (x a)2 + c3 (x a)3 + . . .
n=0
How do we find the coefficients cn in general? Let us assume that f and all its derivatives are defined at
x = a.
First note that
f (a) = c0 .
Now take the derivative of both sides. Assuming that we are justified in differentiating term by term on
the right-hand side, this gives
17.4 Theorem: Formula for Taylor series
238
so that
f 0 (a) = c1 .
Taking the derivative again and again gives
f 00 (a) = 2c2
c2 = f 00 (a)/2
f 000 (a) = 6c3
c3 = f 000 (a)/3!
and, in general
cn = f (n) (a)/n!, with f (0) (x) f (x).
This leads to
The right-hand side is called the Taylor series of f (x) about x = a. So this result says that if f (x) has
a power series representation (expansion) at x = a, it must be given by the formula above.
17.3.1
Example: MacLaurin series
When a = 0 in the Taylor series we call the series a MacLaurin series.

17.3.2
Note
The function f and all its derivatives must be defined at x = a for the series to exist. In this case we say
that f is infinitely differentiable at x = a.
17.4
Theorem: Formula for Taylor series
Suppose f (x) is infinitely differentiable at x = a. Then for f sufficiently well behaved,

f (x) =
X
f (n) (a)
n=0
n!
(x a)n , for |x a| < r,
where r is the radius of convergence.

(Note that if we define a sequence of numbers
bn = max |f (n) (x)|, n N,
x[r,r]
then the sufficiently well behaved means, among other things,

bn rn
= 0.
n n!
lim
Unless otherwise noted, all functions we consider in this section are sufficiently well behaved.)
239
For |x a| < r, we have approximately

f (x) '
k
X
f (n) (a)
n=0
n!
(x a)n
which is a useful polynomial approximation. The smaller |x a|, the better the approximation. This
approximation can be made accurate to arbitrary order by taking k sufficiently large provided |x a| < r.
This begins to address the question of whether or not f (x) has a power series representation. The details
are beyond the scope of this course, but for those interested there is a good discussion about this issue
and Taylors inequality in Stewart, 7ed. pp. 778-780; 6ed. pp. 772-774.
17.4.1
Example
Find the Taylor series of f (x) = ln x about x = 1. Determine its radius of convergence.
240
17.4.2
Example
Find the Taylor series of f (x) = ex about x = a. Determine its radius of convergence.
241
17.4.3
Example
Find the MacLaurin series for f (x) = sin x. Determine its radius of convergence.
242
17.5 Binomial series
17.5
243
Binomial series
For a non-negative integer m, the binomial series is

m
(1 + x)
m
X
m n
=
x ,
n
n=0

m
where
are the binomial coefficients defined by
n

m
m
m(m 1) . . . (m n + 1)
= 1,
=
0
n
n!
for n = 1, 2, ....
Note that in this expression for the binomial coefficient, m can be any real number. In the case that m is
a positive integer and 0 n m, such as in the above expansion, it can be written in the more familiar
form

m
m!
=
.
n
(m n)!n!
Also note that when both m and n are positive integers, with n m + 1 the binomial coefficient is zero.
The above expansion can then be written
(1 + x)

X
m
n=0
xn .
244
Generalize the binomial theorem to arbitrary real powers by finding the MacLaurin series expansion of
the function f (x) = (1 + x) where R. Find also the radius of convergence of the series.
17.5.1
245
Relationship to geometric series
A familiar series is the geometric series

1
= 1 + x + x2 + ,
1x
|x| < 1.
for
This result follows from the binomial theorem. Taking = 1 and replacing x with x we have
f (x) = (1 x)1 =
1
.
1x
The binomial coefficients in this case are

1
(1)(2) . . . (n)
=
= (1)n ,
n!
n
for
The binomial theorem then gives, for |x| < 1,
n=1
n=0
X
X
1
=1+
(1)n (x)n =
xn ,
1x
which is the geometric series.
n 1.
Notes
246
Notes
247
18.1 Definition: antiderivative - Stewart, 7ed. p. 269; 6ed. p. 274
18
18.1
248
Integration
Definition: antiderivative - Stewart, 7ed. p. 269; 6ed. p. 274
A function F is called an antiderivative of f on an interval I if
See the table below for some examples of antiderivatives. We use the notation F 0 = f , G0 = g and c is a
constant.
Function
Antiderivative
Function Antiderivative
cf (x)
cF (x)
cos x
sin x
f (x) + g(x)
F (x) + G(x)
sec2 x
tan x
x , ( 6= 1)
x+1
+1
1
x
ln x
sin x
cos x
ex
ex
The most general antiderivative can be obtained from those given in the table by adding a constant. Notice
d
d
that if
F (x) = f (x) then
(F (x) + C) = f (x) is also true where C is any constant (independent of
dx
dx
x).
18.1.1
Example
Find the most general antiderivative of the function f (x) = x2 + 3x.
18.3 Area under a curve - Stewart, 7ed. pp. 284-291; 6ed. pp. 289-296
18.1.2
Example
If f 00 (x) = x
18.2
249
x, find f (x).
Definition: Indefinite integral

Z
The indefinite integral
f (x)dx, of a suitable function f (x), is defined by
whereRF (x) is any antiderivative of f (x) and c is an arbitrary constant called the constant of integration.
Thus f (x)dx gives all the antiderivatives of f (x).
18.3
Area under a curve - Stewart, 7ed. pp. 284-291; 6ed. pp. 289-296
Consider the problem of finding the area under a curve given by the graph of a continuous, non-negative
function y = f (x) between two points x = a and x = b; see Figure 57.
y=f(x)
Figure 57: We need to define the area of the region S = {(x, y)|a x b, 0 y f (x)}.
250
Although we might have an intuitive notion of what we mean by area, how do we define the area under
such a curve?
We start by subdividing S into n strips S1 , S2 , . . . , Sn of equal width. The width of the interval [a, b] is
b a so the width of each strip is
These strips divide the interval [a, b] into n subintervals

[x0 , x1 ], [x1 , x2 ], . . . , [xn1 , xn ]
where x0 = a and xn = b. If ci is any point in the ith subinterval [xi1 , xi ], then we can approximate
the ith strip Si by a rectangle of width x and height f (ci ), which is the value of f at the point ci (see
Figure 58).
y=f(x)
a c1
c2
c3
c4
c5
c6
Figure 58: A method of approximating the area of the region S.

The area of such a rectangle is f (ci )x. Therefore we can approximate the area of S by taking the sum
of the area of these rectangles. We call this sum Rn :
Now what happens if we increase n? Consider the graph in Figure 59. It seems that as x becomes
smaller (when n becomes larger), the approximation gets better.
Therefore we define the area A of the region S as follows.
251
y=f(x)
Figure 59: The smaller the width of the rectangle, the better the approximation, in general. Notice that
in this diagram we chose ci = a + ix.
18.3.1
Definition: Area of a region, Riemann sum
The area A of the region S that lies under the graph of the continuous, non-negative function f is
It can be shown that for continuous f , this limit always exists and is independent of the choice of ci .
This can be proved using the fact that f is continuous on [a, b]. This sum Rn is called a Riemann sum.
Note that lim [f (c1 )x + + f (cn )x] is called the Riemann integral of f on [a, b] and is denoted
n
Rb
f (x)dx.
IMPORTANT: this definition can be carried over to functions which are not necessarily non-negative.
The intuitive interpretation of A as an area is then no longer valid. The quantity A should then be
viewed as a signed area, i.e.
A = (area above x-axis, below graph) (area below x-axis, above graph).
Stewart, 7ed. p. 297; 6ed. p. 301.
18.4 Fundamental Theorem of Calculus
18.4
18.4.1
252
Fundamental Theorem of Calculus

Theorem: Fundamental Theorem of Calculus
(a) If f is continuous on [a, b], a x b, then

Zx
A(x) =
f (t) dt
a
gives an antiderivative of f (x) such that A(a) = 0. Since A(x) is an antiderivative of f , A0 = f . Thus
Z x
d
f (t) dt = f (x)
dx a
(b) If F (x) is any antiderivative of f (x), then
Zb
F (b) F (a) =
f (t) dt.
a
18.4.2
Proof (outline)
18.4.3
253
Notation
We will write
b h
ib

F (x) = F (x) = F (b) F (a).
a
18.4.4
Properties of the definite integral
For some c R, f (x) continuous on [a, b] there holds:

Z
Z
f (x) dx =
f (x) dx
a
c dx = c(b a)
Z
[f (x) g(x)] dx =
Z
f (x) dx +
f (x) dx =
c
f (x) dx
a
f (x) dx = 0
a
18.4.5
g(x) dx
a
f (x) dx
f (x) dx
cf (x) dx = c
Examples
Z
(1) Evaluate
1
1
dt.
t
/2
(2) Evaluate
sin xdx.
0
Zu
(3) If F (u) =
2
dt
, find F 0 (u).
1 + t2
sin
Z x
(4) If G(x) =
2
dt
, find G0 (x).
1 + t2
254
18.4.6
255
Example
Find the area between the x-axis and the curve y = sin x on [0, ].
18.4.7
Example
Find the area bounded by the curves y = x2 and y = x3 on the interval [0, 1].
18.4.8
Corollary
If f is non-negative and continuous on [a, b] then

Z
f (x)dx 0.
a
If f is non-positive and continuous on [a, b] then

Z
f (x)dx = (area above graph below interval [a, b]) 0.

a
18.5 Approximate integration (trapezoidal rule)
18.4.9
256
Example
Find the area enclosed by the curves y = x and y = x3 .
18.5
Approximate integration (trapezoidal rule)
We can approximate the area under a curve by taking the sum of a finite number of rectangles under the
graph as outlined in the previous section. If we choose the ci to be the left endpoints of each rectangle,
we obtain the approximation
n
X
`n =
f (xi1 )x.
i=1
If we choose the ci to be the right endpoints of each rectangle we obtain the approximation
rn =
n
X
i=1
f (xi )x.
18.6 Improper integrals - Stewart, 7ed. pp. 543-551; 6ed. pp. 544-551
257
The trapezoidal rule is the approximation of the area under a curve which takes the average value of `n
and rn .
The formula for this approximation is therefore given by
18.6
Improper integrals - Stewart, 7ed. pp. 543-551; 6ed. pp. 544-551
Definition
18.6 Improper integrals - Stewart, 7ed. pp. 543-551; 6ed. pp. 544-551
18.6.1
Z
Find
f (x)dx are called convergent if the limit exists and divergent otherwise.
f (x)dx and
The integrals
258
Example
ex dx.
18.6.2
Note
It is still possible to define the integral for some classes of functions which are not necessarily continuous:
Stewart, 7ed. p. 548; 6ed. p. 548. Note also that there are other types of improper integral: Stewart,
7ed. p. 547; 6ed. p. 547.
18.7 Techniques of integration - Stewart, 7ed. ch. 7; 6ed. ch. 8
18.7
Techniques of integration - Stewart, 7ed. ch. 7; 6ed. ch. 8
18.7.1
Substitution
If u = g(x) is a differentiable function whose range is an interval I and f is continuous on I, then
If g 0 is continuous on [a, b] and f is continuous on the range of u = g(x), then
18.7.2
Z
Find

x3 cos(x4 + 2)dx.
259
18.7.3

Z
Evaluate
1
18.7.4
dx
.
(5x 3)2
Example: Trig substitution

Z
Evaluate
dx
.
x2
a2
260
18.7.5
Example
Z
Evaluate
18.7.6
cos x sin x dx.
Example
Z
Evaluate
(3x + 1)(3x2 + 2x)3 dx.
261
18.7.7
262
Integration by parts
Given two differentiable functions u(x) and v(x) we have a product rule for differentiation:
Therefore uv is an antiderivative of uv 0 + u0 v. Hence
The aim is to simplify the function that is integrated, i.e.

Also note that the antiderivative of v 0 needs to be found!
18.7.8
Example
Z
Evaluate
x3 ln xdx.
u0 vdx should be easier to find than
uv 0 dx.
18.7.9
Example
Z
Evaluate
xex dx.
18.7.10
General rule
R
For (polynomial) eax+b dx, choose u to be the polynomial and v 0 = eax+b .
R
For (polynomial) ln (ax + b)dx, choose u = ln (ax + b) and v 0 to be the polynomial.
Note that sometimes integration by parts needs to be applied more than once.
18.7.11
Example
Z
Evaluate
x sin xdx.
263
18.9 Integration of rational functions by partial fractions
264
18.8
Integrals involving ln function
18.9
Integration of rational functions by partial fractions
A ratio of two polynomials,

f (x) =
P (x)
Q(x)
is called a rational function. Integration of a rational function is made possible by expressing it as the
sum of simpler fractions, called partial fractions.
(In the three examples which follow, we will only consider rational functions for which the degree of P (x)
is less than the degree of Q(x).)
The general process for obtaining the sum of partial fractions is as follows:
Factor the denominator Q(x) as much as possible;
A
Ax + B
or
,
i
2
(ax + b)
(ax + bx + c)j
where the denominators of these fractions come from the factorisation of Q(x).
Express f (x) as a sum of fractions each of which take either the form
Consider the following examples.

18.9.1
Example
Z
Evaluate
x+2
dx.
x2 + x
265
18.9.2
Example
Z
Evaluate
x2
dx
, a 6= 0.
a2
266
18.9.3
Example
Z
Evaluate
2x2 x + 4
dx.
x3 + 4x
267
268
18.10 Volumes of revolution
18.10
269
Volumes of revolution
Suppose f (x) is positive and continuous on [a, b]. By rotating the graph of f (x) above [a, b] about the
x-axis, we obtain a cylindrical solid. What is the volume of such a solid? This is what we call a volume
of revolution.
y=f(x)
y=f(x)
(a)
(b)
Figure 60: Rotating the curve in (a) about the x-axis gives the solid in (b).
18.10.1
Formula for volume
Suppose f (x) 0 and continuous on [a, b]. The volume of revolution of the solid obtained by rotating
the graph y = f (x) above [a, b] about the x-axis is
18.10.2
Outline of proof
For x [a, b], let V (x) be the volume obtained by rotating the graph of f above the interval [a, x] about
the x-axis. So V (a) = 0, V (b) = V . Then for h 0 small
V (x + h) V (x) =
volume obtained by rotating graph

of f above [x, x + h] about x-axis
' volume of cylinder with radius f (x) and height h

= [f (x)]2 h (becomes exact as h 0)
V (x + h) V (x)
' [f (x)]2
h
270
Taking the limit h 0 we obtain

V 0 (x) = [f (x)]2
so the Fundamental Theorem yields
Z
a
18.10.3
[f (x)]2 dx = [V (x)]ba = V (b) V (a) = V (b) = V.
Example
Let a > 0. Find the volume of the solid obtained by rotating y = f (x) =
18.10.4
x about the x-axis over [0, a].
Example
Determine the volume of the solid obtained by rotating y = x about the x-axis over [0, 1].
18.10.5
271
Example
Determine the volume of the solid obtained by rotating y = x2 about the x-axis over [0, 1].
18.10.6
Example
The region enclosed by the curves y = x and y = x2 is rotated about the x-axis. Find the volume of the
resultant solid.
Notes
272
19.2 Functions
19
273
Appendix - Practice Problems
You should attempt the practice problems in this section while learning the material in the course. Try to
solve all questions. Do at least questions marked with a star (*) before attempting assignment questions.
Have a look at the questions before your tutorial. Try to solve them yourself first, and ask your tutor at
the tutorial if you need help. Solutions to questions on this sheet are not to be handed in.
19.1
Numbers
1. Solve for x.
(a) |x + 1| < 1
(b) |x + 1| > 1
(c) |3x + 5| 2
2. Simplify.
(a)
(b)
3+i
3i
(2+3i)(3+2i)
1+i
3. Write z = 1
3i in polar form.
4. Find the roots of x2 4x + 5 = 0.
19.2
Functions
1. Graph the following functions and give their domain and range.
(a) f (x) = x2 + 2
(b)
f (x) =
(c)
f (x) =
(d) f (x) =
(e)
f (x) =
(f)
f (x) =
1
x2
1
x2
1
x2 +2
x+2
x2 4
x2 9
(g) f (x) = x2 + 9
(h) f (x) = ex+1

(i) f (x) = 1 + e2x+3 .
2. Determine whether the graphs below represent graphs of functions. If it is state the range and
domain of the function.
19.3 Limits
274
3. Find the domain and sketch the graph of the function

g(x) =
4. Given
|x|
.
x2
(
2x + 3,
f (x) =
3 x,
x < 1
x > 1
State the domain and range of f .

5. Two species of kangaroos inhabit a national park. There are currently 4,500 grey kangaroos, whose
population increases at a rate of 4% a year. There are only 3,000 red kangaroos, but they increase
at a rate of 6% a year. In how many years will the red kangaroos outnumber the grey kangaroos?
Use logarithms to find the answer and draw a graph of the two populations.
6. Let P (t) be the amount of a radioactive material at time t. Experiments indicate P (t) decays
exponentially with time, so
P (t) = P0 ekt
where P0 is the amount of material at time t = 0 (initial amount) and k is a positive constant
(called the decay constant). The half-life T is the time taken for the material to decay to one half
of its initial amount. Obtain T in terms of k.
7. If y = tan(x), 0 x < /2, find cos(x) as a function of y.
8. Find the inverse of f (x) =
1
3
cos(2x) and state its domain and range.
9. For f (x) = 3x + 2, g(x) = ln(x2 + 1) and h(x) = 3e2x+1 , find

(a) f (g(x))
(b) g(f (x))
(c) f (h(x))
(d) h(f (x))
(e) g(h(x))
(f) f (g(h(x)))
(g) h(g(f (x))).
10. Show from the definition that the following two functions are 1-1 on their domains (a graph is not
sufficient here). Find the inverse function, and state the domain and range of both f and f 1 .
(a) f (x) = e3x+1
19.3
(b) f (x) =
x+1
x+2 .
Limits
1. Let
f (x) =
x3
x2 9
Use the limit laws to find L = limx3 f (x).

(a) Find a small interval around 3 where |f (x) L| .01
2x+1
2. Prove that lim x4 cos 3x
2 +5 = 0.
x0
3. Evaluate the following limits
19.4 Continuity
275
x2 +2x8
2
x2 x 5x+6
(a) lim
x2 x12
x+3
x1
(b) lim
x+1
2
x1 x x2
(c) lim
t3 t
2
t
t1 1
2
lim x+2
x
x0
(d) lim
(e)
(Hint: multiply top and bottom by the conjugate surd
4. Show
lim
x3
x+2+
2).
x
3
=
5
5
using the , definition of the limit.

5. Prove that
1
lim x sin( ) = 0.
x0
x
6. Evaluate the following limits.

3x2 +2x+1
2
x x +x+3
(a) lim
(b) lim ( x + 1 x)
x
(Hint: multiply top and bottom by the conjugate surd ( x + 1 + x) )
x2 1
3
x
x +2
(c) lim
xx3
2
x x +1
lim x+1 .
x 1x
(d) lim
(e)
19.4
Continuity
1. Determine whether or not f (x) is continuous at x = 1 and sketch the graph of f (x) for the following
choices of f (x):
2
x 2x + 2 , if x < 1
(a) f (x) =
3x
, if x 1

x3
, if x < 1
(b) f (x) =
(x 2)2 , if x 1
x1
, if x 6= 1
x2 1
(c) f (x) =
6
, if x = 1.
2. For what value of a is f (x) continuous for all x, where

( 2
x 3x+2
x1 , x 6= 1
f (x) =
a,
x=1
Give clear explanations.
3. Is the following function continuous at x = 5?
x+5
, x 6= 5
x2 25
f (x) =
1
10 , x = 5
19.5 Derivatives
19.5
276
Derivatives
1. (a) Sketch the graph of f (x) = x|x|.

(b) For which values of x is f differentiable?
(c) Obtain a formula for f 0 (x).
2. The hyperbolic sine and cosine functions are defined by
1
1
sinh x = (ex ex ), cosh x = (ex + ex )
2
2
respectively.
(a) Prove cosh2 x sinh2 x = 1.
(b) Prove
(c) Find
d
dx
sinh x = cosh x,
d
dx
cosh x = sinh x.
lim sinh x .
x0 x
3. The Tasmanian Devil often preys on animals larger than itself. Researchers have identified the
mathematical relationship
L(w) = 2.265w2.543 ,
where w (in kg) is the weight and L(w) (in mm) is the length of the Tasmanian Devil. Suppose
that the weight of a particular Tasmanian Devil can be estimated by the function
w(t) = 0.125 + 0.18t,
where w(t) is the weight (in kg) and t is the age, in weeks, of a Tasmanian Devil that is less than
one year old. How fast is the length of a 30-week-old Tasmanian Devil changing?
4. Show that the equation of the tangent line to a point (x0 , y0 ) on the ellipse
x2
a2
+ yb2 = 1 is given by
x0 x y0 y
+ 2 = 1.
a2
b
5. Consider the function f (x) = xx . Evaluate f 0 (x) for x > 0.

6. A new car costing P dollars depreciates at an annual rate of r%, so that after t years the car will
be worth C dollars where
r t
C(t) = P (1
).
100
(a) Find
dC
dt
(b) How much will the car be worth in the long run?
7. Find
d
.
d cos(sin )
8. A spherical snowball is melting in such a way that its volume is decreasing at a rate of 1 cm3 /min.
At what rate is the diameter decreasing when the diameter is 10 cm?
9. Using LH
opitals rule evaluate the following limits
(a)
lim
ex 1
x0 sin x
lim sinxxx
3
x0
1
lim ( 1 x1
)
x1 ln x
(b) lim
(c)
(d)
1cos
2
19.6 Integration
277
(e) lim ex ln x
x
(sin )n
n ,
0 sin( )
(f) lim
1 n N.
10. Use the mean value theorem to prove the following (this is the second derivative test).
Suppose f : [a, b] R is continuous, and has continuous first and second derivatives on [a, b]. Then
if c (a, b) is a point with f 0 (c) = 0 and f 00 (c) > 0 then f has a local minimum at c.
11. Determine where the following functions are increasing/decreasing.
(a) f (x) = (x 1)(x + 2) = x2 + x 2
(b) f (x) = x|x|.
(c) f (x) = sin x, x (0, )
(d) f (x) =
ex +1
ex
12. Prove the following:

ex x + 1, for all x 0.
Hint: Consider f (x) = ex x 1.
13. Prove the following:
sin x x, for all x 0.
14. Find the absolute maximum and minimum of the function f (x) = x3 3x + 2 on the closed interval
[0, 2].
15. Find all relative maxima and minima and thus sketch the graph of the function
f (x) = x +
16. Show that the function

f (x) =
1
.
x
x ln x , x > 0
0
, x=0
is continuous on [0, ). What are the absolute maximum and minimum values of f (x) on [0, 2]?
17. To increase the velocity of the air flowing through the trachea when a human coughs, the body
contracts the windpipe, producing a more effective cough. Tuchinsky formulated that the velocity
of air that is flowing through the trachea during a cough is
V = C(R0 R)R2 ,
where C > 0 is a constant based on individual body characteristics, R0 is the radius of the windpipe
before the cough, and R is the radius of the windpipe during the cough. Find the value of R that
maximizes the velocity. (Interestingly, Tuchinsky also states that X-rays indicate that the body
naturally contracts the windpipe to this radius during a cough.)
19.6
Integration
1. Evaluate
R
(a) (x3 2x + 1)dx
R
(b) xa dx, a > 0.
R
(c) sin(x + 3 + 2x)dx
R
(d) (x + x1 )2 dx
19.6 Integration
278
2. Let f be continuous on an interval I. Let F1 and F2 be two different anti-derivatives of f . Show

that F1 and F2 differ by a constant only. (Hint: Recall that if f 0 (x) = 0 on an interval I then f is
constant on this interval.)
3. Estimate the area under the below function using:
(a) the right rectangle rule (Riemann sums) with 2 subintervals, then 4 subintervals then 8 subintervals.
(b) the trapezoidal rule with 4 subintervals.
y = f(x)
1
1
4. What is wrong with the following argument.

F (x) = 1/(x 1) is an anti derivative of the function f (x) = 1/(x 1)2 . Thus by the
fundamental theorem, the area under f on the interval [0, 3] is given by
Z
0

3
1
1 3 1 1
2 dx = (x 1) = 2 1 = 2 .
(x 1)
0
Does this show a positive function can have a negative area?
5. Find the area enclosed by the graphs of the functions f (x) = x2 and g(x) = x.
6. Derive a formula for the area of the ellipse
x2 y 2
+ 2 = 1.
a2
b
7. Find the area under the graph of
(a) y = 4 x2 above [0, 1]

(b) y = sec x tan x above [0, 3 ].
8. Evaluate each of the following integrals by interpreting it in terms of areas:
R3
(a) (1 + 2x)dx
(b)
9. Determine
R0
(1 +
9 x2 )dx
(c)
R2
2
|x|dx
(d)
R2
4 x2 dx.
xex dx.
10. By finding a suitable substitution, evaluate the following indefinite integrals:

R ex
(a)
dx
1+e2x
19.7 Sequences
(b)
(c)
279
dx
dx
x2 +4x+9
3
cos x sin xdx.
11. Use integration by parts to find:

R
(a)
x cos xdx
R
(b) cos2 xdx
R
(c) ex cos xdx
R
R
(d) ln xdx (e) x sec2 xdx
12. Use partial fractions to find
R
x
(a)
(x+1)(x+2) dx
R x+3
dx
(b) x(x+2)
R 4x+8
(c) x2 +2x3 dx.
13. Evaluate the following integrals:
R
R
x
(a) x2x1 dx
(b) 1+e
1ex dx.
14. Evaluate the following indefinite integrals
R
2)
(a) sin(1/t
dt
t3
R
(b) sin(x) cos(cos(x)) dx
Check by explicitly differentating your answer to confirm it is indeed an anti-derivative.
15. Find the volume of the ellipsoid obtained by rotating the ellipse
x2
a2
y2
b2
= 1 about the x-axis.
16. Let a > 0 be constant. Find the volume V of the solid of revolution obtained by rotating y = f (x)
above the interval [0, a] about the x-axis, for the following choices of f (x):
(a) y = x2
p
(b) y = ln(x + 1).
17. Use Eulers formula to express cos(2) and sin(2) in terms of cos() and sin().
19.7
Sequences
1. Use LH
opitals rule to find
lim
sin(1/n)
.
log(1 + 1/n)
2. Evaluate the limit of the sequence {an }

n=1 for the following choices of an :
(a) an = ncn , 0 c < 1
(b) an =
1+2n
1+3n .
3. Prove that
lim n1/n = 1.
19.8 Series
19.8
280
Series
1. It is known that a repeated decimal can always be expressed as a fraction. Consider, for example,
the decimal 0.727272 . . . . Use the fact that
0.727272 = 0.72 + 0.0072 + 0.000072 + . . .
to express the decimal as a geometric series. Hence show that
0.727272 =
8
.
11
2. Determine whether or not the following series converge:

(a)
(b)
n=1
1
1
12 + 23
nn
n!
3. (a) Prove that
1
34
+ ...
nxn =
n=1
x
,
(1x)2
for |x| < 1.
1
d
( 1x
))
(Hint: Consider x dx
(b) Hence, or otherwise, evaluate
P
n=1
n
2n .
4. Consider the p = 1 case of the p-series (called the harmonic series):
X
1
1
1 1
1 1 1 1
1
1
1
=1+( )+( + )+( + + + )+( +
+ + ) + ...
n
2
3 4
5 6 7 8
9 10
16
n=1
and group together its terms as shown. Show that the sum of each group of fractions is more than
1
2 . Hence deduce that the harmonic series diverges.
5. Evaluate the following series:

(a)
(b)
n=1
( 1n
1 )
n+1
(1)n xn .
n=0
6. (a) For p > 0, show that f (x) = xp is strictly increasing on [0, ).

(b) For 0 < p < 1, show that
1
1
, n = 1, 2, 3 . . .
np
n
(c) Show that the p-series
X
1
np
n=1
diverges for 0 p 1.
7. Determine whether or not the following series converge.
(a)
(b)
n=1
(c)
n=0
P
n=1
(1)n
n3n
22n
(2n)!
ln(n)
n! .
19.9 Power series and Taylor series
281
8. (a) Write
P down from the examples in lectures or the text book two series of positive terms
&
bn satisfying the following conditions:

X
an+1
= 1 and the
lim
an converges.

n
an
m

X
bn+1

= 1 and
lim
bn diverges.
n
bn
an
This shows the ratio test gives no information if the limit is 1. You can accept any proofs of
convergence in lectures or the text book but prove both the limits above.
9. Find an example of a sequence an with an = f (n)/g(n) for differentiable functions f and g, where
limn an and limn f 0 (n)/g 0 (n) both exist, and
f 0 (n)
.
n g 0 (n)
lim an 6= lim
This shows before using LH

opitals rule we must remember to check that either limn f (n) = 0
= limn g(n) or limn f (n) = = limn g(n)!
19.9
Power series and Taylor series
1. This question is about the MacLaurin series for

(1 + x)3/2 .
Find the first three terms of this series. If this series is written as
Find a formula for cn so an+1 = cn an .
n=0
find the formula for an .
(1)n
.
(2n + 1)3n
3. Obtain the MacLaurin series expansions and radius of convergence for the following functions:
1
cosh x = (ex + ex )
2
n=0 an ,
2. Obtain the MacLaurin series expansion for arctan x and its radius of convergence.
1
d
(arctan x) = 1+x
(Hint: Use dx
2 ).
Using the fact that tan(/6) = 1/ 3, show that
X
=2 3
1
sinh x = (ex ex ).
2
4. Show that for |x| < 1,

(a)
1+x=1+
X
(1)n1 (2n 2)!
n=1
22n1 n!(n 1)!
xn .
(b)
X (1)n (2n)!
1
=
xn .
1 + x n=1 22n (n!)2
5. The theory of relativity predicts that the mass m of an object when it is moving at a speed v is
given by the formula
m0
m= p
1 v 2 /c2
where c is the speed of light and m0 is the mass of the object when it is at rest. Express m as
a MacLaurin series in vc and determine for which values of v this series converges. (Hint: Use a
result from the previous question.)
19.10 Vectors
282
6. Show that the functions

(a) f (x) = (1 + x2 )ex
(b) f (x) = x2 ln(1 + x2 )

both satisfy f 0 (x) = f 00 (x) = 0 at x = 0 (so the second derivative test fails here). Show, using
Taylor series, that (a) has a relative maximum and (b) a relative minimum at x = 0.
7. Find the sum of each of the following convergent series by recognizing it as a Taylor series evaluated
at a particular value of x.
(a) 1 +
(b) 1
(c) 1 +
(d) 1
19.10
2
4
8
2n
1! + 2! + 3! + + n! + . . .
(1)n
1
1
1
3! + 5! 7! + + (2n+1)! + . . .
1
1 2
1 3
1 n
4 + (4) + (4) + + (4) + . . .
(1)n 102n
100
10000
+ ...
2! + 4! + +
(2n)!
Vectors
1. A canoe travels at a speed 3km/h in a direction due west across a river flowing due north at 4km/h.
Find the actual speed and direction of the canoe.
2. Draw a clock with vectors v 1 , v 2 , . . . , v 12 pointing from the centre to each of 1, 2, . . . , 12. Describe
each of the following vectors.
(a) v 1 + + v 12 .
(b) v 1 + v 2 + 3v 3 + v 4 + v 5 + v 6 + v 7 + v 8 + 3v 9 + v 10 + v 11 + v 12 .
(c) v 1 + v 2 + v 3 + v 4 + v 6 + v 7 + v 9 + v 10 + v 12 .
3. Find the lengths of the sides of the triangle ABC and determine whether the triangle is a right
triangle.
(a) A = (4, 2, 0), B = (6, 6, 8), C = (10, 8, 6)
(b) A = (5, 5, 1), B = (6, 3, 2), C = (1, 4, 4).
4. (a) Suppose u and v are orthogonal vectors. Prove that
kuk2 + kvk2 = ku + vk2 . What geometric fact have you proved?
(b) Prove that ku + vk2 + ku vk2 = 2(kuk2 + kvk2 ). What geometric fact have you proved?
(c) Consider a triangle with side lengths a, b, c. Bisect side c, and draw a line connecting the
midpoint to the opposite vertex. Let the length of this line be m. Using vectors prove that
a2 + b2 = 2(m2 + (c/2)2 ). This is called Apollonius Theorem.

1
4
5. Find a non-zero vector in R3 orthogonal to 2 and 5 .
3
6
6. Given the point P = (2, 1, 4), find a point Q where P Q has the same direction as v = 7i+6j 3k.
7. For what values of c is the angle between the vectors (1, 2, 1) and (1, 1, c) equal to 60 ? For what
values of c are they orthogonal?
8. Determine whether or not each of the following sets of vectors are linearly independent.

1
1
1
(a) 2 , 1 , 2 .
0
1
1
19.11 Matrix Algebra
283
(b) i + j, i + j + k, i j.

1
1
1
0 2 2

(c)
0 , 0 , 1 ,
0
0
0
9. Prove that the vectors
4
4
.
1
0

1
1
1
0 , 0 , 1
1
1
0
2
form a basis for R3 . What are the components of the vector 1 in this basis?
0
10. Given
1
2
2
v 1 = 5 , v 2 = 1 , v 3 = 9 ,
3
6
h
for what values of h is v 3 in span{v 1 , v 2 }, and for what values of h is {v 1 , v 2 , v 3 } linearly independent? Justify your answer.
11. Which of the following subsets of R3 are vector spaces? Find a basis in each case.
(a) {(a, b, a + 2b) | a, b R}.
(b) {(a, b, c)|a b = 3c + 1}.
12. Suppose S = {v 1 , v 2 , . . . , v m } Rn and let V be the set of all linear combinations of the elements
of S, so
V = {1 v 1 + 2 v 2 + + m v m | 1 , 2 , . . . , m R}.
Prove that V is a vector space.
13. If {a, b, c} forms a basis for a 3-dimensional vector space, could {a + b, b + c, c + a} also be a basis?
Give your reasons.
14. Let u1 = (0, 0, 3), u2 = (2, 1, 0), u3 = (1, 0, 0). Which of u1 , u2 , u3 are linear combinations of
v 1 = (1, 2, 3), v 2 = (4, 5, 6), v 3 = (7, 8, 9)? Do {v 1 , v 2 , v 3 } form a basis of R3 ?
15. Determine if the following vectors are a basis for R3 . Justify your answer.

0
5
6
1 , 7 , 3
2
4
5
16. Let v 1 = (1, 2, 3), v 2 = (4, 5, 6), v 3 = (2, 1, 0). Do {v 1 , v 2 , v 3 } form a basis of R3 ? Explain.
17. Let v 1 = (1, 4, 6), v 2 = (2, 0, 1), v 3 = (1, 12, 16). Let V = span{v 1 , v 2 , v 3 }. Find dim V .
19.11
Matrix Algebra

cos sin
1. Let A =
sin cos
. Find simple expressions for A2 and A3 . What do you notice?
2. Find 2 2 matrices A, B such that (A + B)2 6= A2 + 2AB + B 2 .
3. Prove that (A + B)T = AT + B T for any m n matrices A and B.

4. Let A =
284
1 2
3 4

. Find all 2 2 matrices B such that AB = BA.
5. Find a 2 2 matrix A with no entry of A equal to 0 such that A2 = 0.

6. A matrix A is called idempotent if A2 = A.
(a) Find x so that A is idempotent:
2/3 0 x
A = 0 1 0
2
0 1/3
3
(b) What is the smallest (integer) index N so that B N = 0? A matrix with such a property is
called nilpotent.
0 1 0 0
0 0 2 1
B=
0 0 0 1 .
0 0 0 0
1 2
1
2 . Show that C is invertible, ie show linear independence of a certain set of
7. Let C = 1 1
1 3 1
vectors.
4 -1 2
8. Let A = 3 0 x . For which value(s) of x is A not invertible?
8 -3 2
9. Find the inverse matrix for:

1 0
(a)
0 1

1 2
(b)
3 4
1 2 3
(c) 0 1 2
0 0 1
10. Let
A=
1
5
3
5
2
5
1
5
3
5
2
5
1
5
1 2 x
and B = 3 1 y .
0
0 z
Find real numbers x, y and z such that B = A1 .

11. A swimming pool can be filled by 3 pipes A,B and C. Pipe A can fill the pool by itself in 8 hours.
Pipes A and C used together can fill the pool in 6 hours and pipes B and C used together can fill
the pool in 10 hours. How long does it take to fill the pool if all 3 pipes are used?
12. Use Gaussian elimination to solve the following system of equations, and check your answer by
substitution.
x1 2x2 + 3x3 = 6
4x1 + 2x3 = 2
3x1 + x2 + 2x3 = 3
285
13. Let xp be a particular solution of Ax = b. Show that every solution of this equation is of the form
x = xp + h where h is a solution of the homogeneous equation Ax = 0.
14. Show that the set of solutions of
Ax = b, with b = 0
form a vector space for A.
If b 6= 0, does the set of solutions still form a vector space?

2 1
15. Show that A =
satisfies A2 4A + 3I = 0. Hence deduce that A is invertible with inverse
1 2
A1 = 31 (4I A).
16. Let a R. Use Gauss-Jordan elimination to find the inverses of the following matrices. Check by
multiplication if your inverse is correct.
1 1
0
(a) 1 2

2 1 1
0 0 1
(b) 1 2 1 .
1 0
17. Find the inverses of the following matrices.
1 1 1
(a) 1 2 1 .
2 1 1
1 2 3 1
1 3 3 2
(b)
2 4 3 3 .
1 1 1 1

1 1
.
(c)
1 2
1 0 5
18. Let A = 2 -1 2 . Calculate the determinant of A.
3 1 2
19. Determine for which values of x the following determinants vanish:

1x
2
(a)
3
2x

1 2

(b) 2
x x

1 1 4

(c) 2
x 1 .

2x x2 x
20. Evaluate the following determinants

cos sin

(a)
sin cos

3 1 0

(b) 0 6 4
0 0 1

(c)

(d)

1 3
4
2 4 1
3 1
0
1 1 5

1 1 2
2 1 1
1 1 1
1
3
0
4
21. Use determinants to establish the linear dependence or independence of the following row vectors:
(a) (4, 1, 4), (2, 0, 4), (1, 2, 2)
(b) (3, 3, 1), (6, 6, 0), (3, 3, 2)
(c) (1, 1, 1), (0, 1, 0), (0, 1, 1)
286
22. For an invertible matrix A, prove that An is also invertible, for all natural numbers n.
20
20
APPENDIX - MATHEMATICAL NOTATION
287
Appendix - Mathematical notation
>
<
,
{x|...}
...
n!
is strictly greater than

is strictly less than
is greater/less than or equal to
is an element of
is not an element of
is a subset of
set of all x for which a condition is fulfilled
for all
continue the pattern
n factorial, n! = n (n 1) ... 3 2 1
implies, if ... then
is equivalent to, if and only if
e.g.
e.g.
e.g.
e.g.
e.g.
e.g.
e.g.
e.g.
e.g.
e.g.
e.g.
e.g.
5>2
2<5
2 2, 4 3
3.5 R
3.5
/N
NR
{x R|x > 0}
x R
1, 2, 3, ..., 9, 10
3! = 3 2 1 = 6
x = 2 x2 = 4
x2 = 4 x = 2 or x = 2
21.1 Powers
21
288
Appendix - Basics
This appendix contains some very important basic mathematical rules. Since this is highschool material,
you must know and be able to apply all of them to successfully complete MATH1051. Go through
the sections and check your knowledge - do some additional exercises from textbooks if you encounter
problems, or check the MATH1040/MATH1050 material on the course websites.
21.1
21.1.1
Powers
Notation (product of two real numbers)
The product of two real numbers a and b can be written in either of the following ways:
a b = a b = ab.
21.1.2
Notation (powers)
The product after multiplying the same real number a n-times is

aa
a} = an .
| {z
n factors
Note that an = aan1 = an1 a.
Then a0 = 1 (for a 6= 0) and a1 = a, aa = a2 = aa1 , aaa = a3 = aa2 .
We also have 0n = 0 for n 6= 0.
21.1.3
Power rules
Let a, b, c R. We set
ab =
1
.
ab
Then
ab ac = ab+c , a > 0 and
c
ab = abc .
We also have
aq =
ap ,
with p, q Z and q > 0.

These formulae also apply for some other values of a, p and q but care needs to be taken.
1
For example, a = 1, p = 1. If q = 3, then (1) 3 = 1.
1
But if a = 2, then (1) 2 = i, which is not a real but a complex number!
21.1.4
Examples
1. (a a) (a a) (a a) = a2
2. a2 a3 = a2+3 = a5
1
3. a 2 = a
1
4. a 3 = 3 a
3
= a23 = a6
21.3 Fractions
21.2
289
Multiplication/Addition of real numbers
21.2.1
Laws (multiplication and addition)
1. a + b = b + a and ab = ba
(Commutative Law)
2. (a + b) + c = a + (b + c) and (ab)c = a(bc)
(Associative Law)
3. a(c + d) = ac + ad = (c + d)a = ca + da
21.2.2
(Distributive Law)
Examples (multiplications of sums)
1. (a + b)(c + d) = ac + ad + bc + bd,
as (a + b)(c + d) = a(c + d) + b(c + d) = ac + ad + bc + bd.
2. (a + b)2 = a2 + 2ab + b2
as (a + b)2 = (a + b)(a + b) = a(a + b) + b(a + b)
= aa + ab + ba + bb = a2 + 2ab + b2 .
3. (a + b)3 = a3 + 3a2 b + 3ab2 + b3
as (a + b)3 = (a + b)(a + b)2 = (a + b)(a2 + 2ab + b2 )
= a(a2 + 2ab + b2 ) + b(a2 + 2ab + b2 )
= a3 + 3a2 b + 3ab2 + b3 .
4. (a b)2 = a2 2ab + b2
5. (a + b)(a b) = a2 b2
21.3
21.3.1
Fractions
Rule (multiplying and dividing two fractions)
Let a, b, c, d Z. Then
and
21.3.2
a c
ac
=
b d
bd
a
b
c
d
a d
ad
=
.
b c
bc
Rule (adding two fractions)
Let a, b, c, d Z. Then
a b
a+b
+ =
.
c c
c
If the two fractions have different denominator, a common denominator needs to be found:
a c
ad + bc
+ =
.
b d
bd
21.4 Solving quadratic equations
21.3.3
290
Examples
1.
x2 + 1
x2
1
1
=
+ =x+ .
x
x
x
x
2.
x
(x + 1) 1
x+1
1
1
=
=
=1
x+1
x+1
x+1 x+1
x+1
3. Simplify
1
x
1
2
x+1 x+2
1
2
1(x + 2) 2(x + 1)
x
Solution: First
=
=
x
+
1
x
+
2
(x
+
1)(x
+
2)
(x
+
1)(x
+ 2)

1
1
x
1
1
2
=
=
x x+1 x+2
x (x + 1)(x + 2)
(x + 1)(x + 2)
21.4
Solving quadratic equations
A quadratic equation is an equation of the form ax2 + bx + c = 0, or x2 + bx + c = 0 if all coefficients

are divided by a. Here a, b and c are given. Quite often you will be required to solve equations of this
type, which means we try to determine x. There are many different ways of finding the solution. A few
are listed here. They are completing the square, using a formula, or factorizing.
21.4.1
Completing the square
We may solve x2 + bx + c = 0 by first writing the quadratic polynomial x2 + bx + c in the form (x + d)2 + e,
and then solving (x + d)2 + e = 0.
Applying the formula from example 21.2.2.(2),
x2 + bx = (x + 2b )2 ( 2b )2 (check!)
x2 + bx + c = (x + 2b )2 ( 2b )2 + c = (x + d)2 + e, with d =
b
2
and e = c ( 2b )2 = c d2 .
To solve: (x + d)2 + e = 0 (x + d)2 = e x + d = e x = d e or x = d + e.

21.4.2
Examples
1. Express x2 + 6x + 5 = (x + d)2 + e. Then solve the equation x2 + 6x + 5 = 0.

x2 + 6x + 5 = (x + 3)2 32 + 5 = (x + 3)2 4.
We solve x2 + 6x + 5 = (x + 3)2 4 = 0
(x + 3)2 = 4 x + 3 = 2 x = 5 or x = 1.
2. The method of completing the square can also be used to find the turning point of a parabola.
Express f (x) = x2 + 4x + 7 = (x + d)2 + e. Find the turning point of this parabola.
Now b = 4 so d = 2.
Also c = 7 so e = 7 4 = 3. Thus
x2 + 4x + 7 = (x + 2)2 + 3.
The turning point of a parabola f (x) = (x + d)2 + e is at (d, e).
In this case the turning point is a minimum located at (2, 3).
21.4 Solving quadratic equations
21.4.3
291
Formula (Solution to quadratic equation)
There is also a formula you can use to solve a quadratic equation.

Let a, b, c R.
The solution of the equation
ax2 + bx + c = 0
is
x=
b2 4ac
.
2a
The solution of the equation

x2 + bx + c = 0
is
b
x=
2
21.4.4
b2
c.
4
Factorizing
Another way of solving quadratic (and higher order polynomial) equations is by factorizing. We try
to write the polynomial as a product of sums. You can sometimes use the formulas from Examples
21.2.2.(2),(4) or (5) for this. Factorizing is also used to simplify expressions.
If you cannot use one of the formulas from Examples 21.2.2.(2),(4) or (5), you can sometimes factorize
the following way. We would like to express x2 + bx + c = (x + p)(x + q). Multiplying out we have
x2 + bx + c = (x + p)(x + q) = x2 + (p + q) + pq. We now need to find p and q such that b = p + q and
c = pq. Have a look at these examples.
21.4.5
Examples
1. Factorize f (x) = x2 + 6x + 5 using the formula from Example 21.2.2.(2). Then find the zeros.
x2 + 6x + 5 = (x + 3)2 22 = ([x + 3] 2)([x + 3] + 2) = (x + 1)(x + 5).
The zeros can be read directly from this expression:
(x + 1)(x + 5) = 0 x = 1 or x = 5.
2. Factorize x2 + 3x + 2 without using the formula from Example 21.2.2.(2).
x2 + 3x + 2 = x2 + bx + c. We must find p and q such that b = 3 = p + q and c = 2 = pq.
A solution is p = 1 and q = 2, so x2 + 3x + 2 = (x + 1)(x + 2).
3. Factorize x2 7x + 6.
Now 7 = p + q and 6 = pq. A solution is p = 6 and q = 1, so x2 7x + 6 = (x 6)(x 1).
21.4.6
Factorizing using common factors
The following formula uses common factors to factorize :

a(x + d) + b(x + d) = (a + b)(x + d).
21.5 Surds
21.4.7
292
Examples
1. Factorize x(x + 3) + 2(x + 3).

x(x + 3) + 2(x + 3) = (x + 2)(x + 3)
setting a = x, b = 2, and d = 3 in the above formula.
2. Factorize (x + 4)(x + 3) + 2(x + 3).
(x + 4)(x + 3) + 2(x + 3) = ([x + 4] + 2)(x + 3)
setting a = [x + 4], b = 2, and d = 3.
21.5
Surds
The following formula is often used to simplify surds:
b=
ab
.
a+ b
To see this
b =
=
=
a + b
a b
a+ b

( a b)( a + b)
a+ b
( a)2 ( b)2
ab
.
=
a+ b
a+ b
Similarly,
21.5.1
a+
ab
b=
a b
Example
Simplify
2x + 1
x
x+1
, where x 6= 0
First
2x + 1
x+1 =
=
2x + 1
x
x+1
=
=
2
2
2x + 1
x+1
2x + 1 + x + 1
x
2x + 1 + x + 1

1
x
, where x 6= 0
x
2x + 1 + x + 1
1
, where x 6= 0
2x + 1 + x + 1

1051 Workbook

Uploaded by

Document Information

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

1051 Workbook

Uploaded by

Copyright:

Available Formats

MATH1051

How to use this workbook

Vectors & Matrices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Systems of Linear Equations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Eigenvalues & Eigenvectors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

0.10 Useful Commands for MATH1051

Row and column vectors

2.10 The Projection Formula . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

2.12 Properties of Dot Product . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

3 Matrices and Linear Transformations

The effect of matrix multiplication on vectors . . . . . . . . . . . . . . . . . . . . . . . . .

Composing Transformations & Matrix Multiplication . . . . . . . . . . . . . . . . . . . . .

The general solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

The Identity Matrix . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Inverses via GaussJordan. Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Connection with inverses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

7 Vector Products in 3-Space

Definition: Cross product . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Application: area of a triangle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Scalar Triple Product . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

8 Eigenvalues and eigenvectors

Eigenvalues & Vectors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

How to find eigenvalues . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Linear combinations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104

Linear independence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106

How to test for linear independence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106

Invertible matrices and linear independence . . . . . . . . . . . . . . . . . . . . . . . . . . 109

Vector spaces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111

The span of a set of vectors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117

9.10 Components . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 122

10.1 Number systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126

11.1 Notation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 131

11.13Natural logarithm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 146

12.1 Definition: Limit - Stewart, 7ed. p. 50; 6ed. p. 66 . . . . . . . . . . . . . . . . . . . . . . . 156

13.1 Formal Definition: Sequence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 170

14.1 Definition: Continuity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 179

15.1 Tangents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 187

15.13Increasing/Decreasing test - Stewart, 7ed. p. 214; 6ed. p. 221 . . . . . . . . . . . . . . . . 201

16.1 Infinite sums (notation) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 217

16.3 Definition: Convergence - Stewart, 7ed. p. 729; 6ed. p. 724 . . . . . . . . . . . . . . . . . 218

17.1 Definition: Power series . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 234

18.1 Definition: antiderivative - Stewart, 7ed. p. 269; 6ed. p. 274 . . . . . . . . . . . . . . . . . 248

19.1 Numbers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 273

21.1 Powers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 288

0.1 Introduction to MATLAB

The MATLAB Program Window

0.1 Introduction to MATLAB

Variables & Basic Commands

0.1 Introduction to MATLAB

0.1 Introduction to MATLAB

0.2 Plotting Functions

Labelling & Editing Figures

0.3 Vectors & Matrices

Vectors & Matrices

0.4 Systems of Linear Equations

Systems of Linear Equations

You know from lectures that finding

the solution to this is equivalent to solving the matrix equation

Remember that the commands

0.4 Systems of Linear Equations

One method for solving a matrix equation Ax = b is by noting that x = A1 Ax = A1 b, as long as A1

Solution by Gaussian Elimination

0.6 Plotting points

Eigenvalues & Eigenvectors