You are on page 1of 23

Submitted to:Dr V.K.

Pathak
Objective of the
Presentation:

Using the dynamic programming


to solve a problem of reliability
maximization of a system with a
multiplicative optimization
function.
N devices Di ,
1<i<N ,

Let ri be the reliability of a decive Di ,which


means that ri is the probability that the device i
will function properly.
Then the reliability of the entire system is ∏ri.

Reliability of the system= (r1*r2*r3*r4…..*rn)


The Problem :

even if the individual reliabilities


are very close to 1, the system
reliability is low.

Say for example : if there


are 10 devices with
reliability equal to
0.99( which is very good) .
THE RELIABILITY OF THE
DEVICE
DUPLICATION
Multiple copies of the same device type are connected in
parallel!

if stage i contains mi copies of device Di,


then the probability that all mi have a
malfunction is :

Hence the reliability of stage i becomes:


•Thus reliability of the system of stages is :

Stage Stage Stage Stage


1 2 3 n
Our problem is to use device duplication to
maximize reliability. This maximization is to be
carried out under a cost constraint!.
Let Ci be the cost of each unit of device I and let c
be the maximum allowable cost of the system
being designed.
we wish to solve the following

MAXIMIZE :

Subject
to :

Each mi must be in the range 1≤mi≤ui

Ui is the maximum value


where., ui= of mi.
i.e. The maximum no. of
times the device i can be
duplicated.
How do we get ui=
• ui is the maximum value of mi….
• All the devices have to be taken
atleast once…

• We can calculate it this way…


• 1*c1+1*c2+……+ui*ci+….+1*cn=C
• ui*ci=C+ci-(c1+c2+c3+…..+cn)
SO……In short
• There are n stages
• The number of copies of a device i
at stage i is mi.
• The reliability of device i is ri
• Reliability of system is

• Cost of device i is Ci & total


permissible cost -C
N devices Di ,
1<i<N ,

Here , mi=3,
i=1 (stage
1)
3
Reliability of
stage1= 1-
(1-r1)

Stage Stage Stage Stage


1 2 3 n
HOW TO SOLVE A PROBLEM
USING RELIABILITY DESIGN
APPROACH
• From the given data of ri’s and no. of device
types, calculate ui’s .
• We ,then define a set
• a= number of stages traversed…
• b=number of devices taken at stage…
• Then define which is a set ( , …. ).
• The set contains the tuples (f,x).

a starts with 1 and goes till last stage


• Where, f=reliability of the system
considered so far .
• x= cost calculated.
• The final set is formed after
eliminating tuples by
DOMINANCE RULE…
DOMINANCE RULE
• The tuple (f1,x1) is dominated by the tuple (f2,x2) if and only if
f2≥f1 and x2≤x1.

 The set contains ONLY


UNDOMINATED TUPLES…………..
Example Problem:
Q. Design a 3 stage system with device types
D1,D2,D3.the costs are 30,15 and 20 units
respectively .the cost of the system can be no
more than 105 units. The reliability of each
device type is 0.9,0.8,0.5 respectively.
SOLUTION
We first need to find out the
value of u1,u2,u3…we already
have r1,r2 ,r3 and c1,c2,c3 and
C.

Use the formula: ui=


The values of u1=2 , u2=3 and
u3=3
That means : value of m1 can be 1 or 2
value of m2 can be 1 or 2 or 3
value of m3 can be 1 or 2 or
3

SO……how many systems are


possible????

=18
The total no. of possible systems :
5.
1.

2. 6
.

3. 7
.

4. 8
.
9 13
. .

14
10 .
.

11 15
. .

12 16
. .
17.

18.
But with the help of our
reliability design approach we
do not need to bother about
these
• We have 18 cases!!!!!!
already calculated the
respective ui’s which are 2,3,3.
We now start with finding =( , ) )

f is the reliability of the


(f,x)=(0.9,3
device D1 ,x is the cost
0)
of it !
2
(f,x)=( 1- (1-0.9) ,30+30 )
=(0.99,60)

( , ) = ( (0.9,30),
(0.99,60))
• Now ,we need to compute
= ( , , )

={ (0.72,45),
(0.792,75) }
= { (0.864,60)}

Note : the tuple (0.9504,90) which comes


from (0.99,60) has been eliminated from
as this leaves the cost of only 10
units for 3rd device .
={(0.8928,75)}

Combining, we get: ={ (0.72,45),(0.864,60),


(0.8928,75)}
(NOTE: (0.792,75) is dominated by (0.864,60))

Finally : ={(0.36,65),(0.432,80),
(0.4464,95)}

={(0.54,85),(0.648,100)}

={(0.63,105)}

BEST
={(0.36,65),(0.432,80),
(0.54,85),(0.648,100)} SOLUTION
• Back tracking through S’s we get

M1=1
M2=
2
M3=2

You might also like