Scribd
Upload
50 views21 pages

LP - V Lecture (Sensitivity Analysis)

This document discusses sensitivity analysis in linear programming. It explains how changes to the objective function coefficients, right-hand side values of constraints, or constraint coefficient matrix can impact the optimal solution. Specifically: [1] Changes to objective function coefficients of non-basic variables does not change the optimal solution if it remains within certain bounds. Changes to basic variables' coefficients may change the optimal solution completely or change only the optimal objective value. [2] Increasing the right-hand side of a constraint keeps the optimal solution if it remains within bounds determined by the constraint coefficient matrix. Decreasing the right-hand side may change the optimal solution. [3] Changing a constraint coefficient related to a non-basic

Uploaded by

manendra chopra
Copyright
© Attribution Non-Commercial (BY-NC)
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPT, PDF, TXT or read online on Scribd
50 views21 pages

LP - V Lecture (Sensitivity Analysis)

This document discusses sensitivity analysis in linear programming. It explains how changes to the objective function coefficients, right-hand side values of constraints, or constraint coefficient matrix can impact the optimal solution. Specifically: [1] Changes to objective function coefficients of non-basic variables does not change the optimal solution if it remains within certain bounds. Changes to basic variables' coefficients may change the optimal solution completely or change only the optimal objective value. [2] Increasing the right-hand side of a constraint keeps the optimal solution if it remains within bounds determined by the constraint coefficient matrix. Decreasing the right-hand side may change the optimal solution. [3] Changing a constraint coefficient related to a non-basic

Uploaded by

manendra chopra
Copyright
© Attribution Non-Commercial (BY-NC)
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPT, PDF, TXT or read online on Scribd
You are on page 1/ 21

Linear Programming

Sensitivity Analysis
or
Post Optimality Analysis
Max Z = 25x
1
+ 40x
2
+ 50x
3
S/t. 5x
1
+ 20x
2
+ 25x
3
400
10x
1
+ 15x
2
+ 20x
3
500
x
1
, x
2
, x
3
0
25 40 50 0 0
RHS
x
1
x
2
x
3
w
1
w
2

40 x
2
0 1 6/5 2/25 1/25 12
25 x
1
1 0 1/5 3/25 4/25 32
0 0 3 1/5 12/5 1280
Final Table :

[ 1 ] Change in C
j

x
j
non-basic
x
j
basic
[ 2 ] Change in RHS
[ 3 ] Change in a
ij

x
j
non-basic
x
j
basic
[ 4 ] Addition of a new variable
[ 5 ] Addition of a new constraint
Effect of Change :
[ 1 ] Present Optimal Solution unchanged.
[ 2 ] Basic variables same values change Z &/or R.
[ 3 ] Optimal Solution changes completely.
Max Z = 25x
1
+ 40x
2
+ 50x
3
S/t. 5x
1
+ 20x
2
+ 25x
3
400
10x
1
+ 15x
2
+ 20x
3
500
x
1
, x
2
, x
3
0
25 40 50 0 0
RHS
x
1
x
2
x
3
w
1
w
2

40 x
2
0 1 6/5 2/25 1/25 12
25 x
1
1 0 1/5 3/25 4/25 32
0 0 3 1/5 12/5 1280
Final Table :
Therefore in general :
o o (I
j
)
xj non-basic

25 40 50 0 0
RHS
x
1
x
2
x
3
w
1
w
2

40 x
2
0 1 6/5 2/25 1/25 12
25 x
1
1 0 1/5 3/25 4/25 32
0 0 3 1/5 12/5 1280
[ 1 ] ( a ) C
j
w.r.t. x
j
non-basic
To retain optimality : 3 o 0
o 3
+o
o
25 40 50 0 0
RHS
x
1
x
2
x
3
w
1
w
2

40 x
2
0 1 6/5 2/25 1/25 12
25 x
1
1 0 1/5 3/25 4/25 32
0 0 3 1/5 12/5 1280
[ 1 ] ( b ) C
j
w.r.t. x
j
basic
To retain optimality :
3+6o/5 0
1/5+2o/25 0
12/51o/25 0
5/2 o 60
+o
+o
+6o/5 +2o/25 1o/25
o 5/2
o 5/2
o 60
+12o
Therefore in general : for Cj w.r.t. xj basic
( ) ( )
j j
j j
ij x ij x
I I
a a
o


s s



Max
m

of
a
ij
> 0
Min
m

of
a
ij
< 0
25 40 50 0 0
RHS
x
1
x
2
x
3
w
1
w
2

40 x
2
0 1 6/5 2/25 1/25 12
25 x
1
1 0 1/5 3/25 4/25 32
0 0 3 1/5 12/5 1280
To get o for x
2
:
3/(6/5),
(1/5)/(2/25), (12/5)/(-1/25)
5/2,
5/2 ,
60
5/2 o 60
25 40 50 0 0
RHS
x
1
x
2
x
3
w
1
w
2

40 x
2
0 1 6/5 2/25 1/25 12
25 x
1
1 0 1/5 3/25 4/25 32
0 0 3 1/5 12/5 1280
To get o for x
1
:
3/(1/5),
(1/5)/(-3/25), (12/5)/(4/25)
5/3,
15 ,
-15
15 o 5/3
Q.1 What will happen if C
3
is changed to :
( i ) 45 ( ii ) 52 ( iii ) 55
Q.2 What will happen if C
1
is changed to :
( i ) 15 ( ii ) 5 ( iii ) 30
Q.3 What will happen if C
2
is changed to :
( i ) 35 ( ii ) 50 ( iii ) 110
Answer the following questions.
Change in RHS.
Max Z = 25x
1
+ 40x
2
+ 50x
3
S/t. 5x
1
+ 20x
2
+ 25x
3
400+ o
10x
1
+ 15x
2
+ 20x
3
500
x
1
, x
2
, x
3
0
25 40 50 0 0
RHS
x
1
x
2
x
3
w
1
w
2

40 x
2
0 1 6/5 2/25 1/25 12+ 2o /25
25 x
1
1 0 1/5 3/25 4/25 32- 3o /25
0 0 3 1/5 12/5 1280+1o/5
To retain optimality :
32- 3o /25 > 0
150
800/ 3
o
o
>
s +
150 o 800/3
12+ 2o /25 > 0
Change in RHS.
Max
m

of
a
ij
> 0
Min
m

of
a
ij
< 0
( ) ( )
j
i i
ij ij x
R R
a a
o



s s





Where a
ij
Column corresponding to slack variable w.r.t. i
th

constraint
For AR
1
12/(2/25), 32/(3/25) = 150, 800/3
150, s AR
1
s 800/3
For AR
2
12/( 1/25), 32/(4/25) = 300, 200
200, s AR
1
s 300
To get range of AR for which basic variables are retained :
New value of RHS in final table = old value + (a
ij
) . AR
Effect of AR
1
:
New value of b
1
= b
1
+ (a
ij
) AR
1

New value of b
2
= b
2
+ (a
ij
) AR
1

Effect of AR
2
:
New value of b
1
= b
1
+ (a
ij
) AR
2

New value of b
2
= b
2
+ (a
ij
) AR
2

Effect of AR
1
and AR
2
simultaneously :
New value of b
1
= b
1
+ (a
ij
) AR
1
+ (a
ij
) AR
2

New value of b
2
= b
2
+ (a
ij
) AR
1
+ (a
ij
) AR
2

Answer the following questions.
Q.1 What will happen if R
1
is changed to :
( i ) 300 ( ii ) 600 ( iii ) 200 ( iv ) 700
Q.2 What will happen if R
2
is changed to :
( i ) 400 ( ii ) 600 ( iii ) 250 ( iv ) 900
Q.3 What will happen if R
1
is changed to 300 and R
2
is
changed to 600 simultaneously ?
Q.4 What will happen if R
1
is changed to 200 and R
2
is
changed to 300 simultaneously ?
Change in a
ij
w.r.t. non-basic variable
&
Addition of a new variable
2 1
25
25 25
3 4 20
25 25




6
5
1
5






What will happen if is changed to ?
25
20



10
10



2 1
10
25 25
3 4 10
25 25




0
1



25 40 50 0 0
RHS
x
1
x
2
x
3
w
1
w
2

40 x
2
0 1 0 2/25 1/25 12
25 x
1
1 0 1 3/25 4/25 32
0 0 25 1/5 12/5 1280
25 40 50 0 0
RHS
x
1
x
2
x
3
w
1
w
2

40 x
2
0 1 0 2/25 1/25 12
50 x
3
1 0 1 3/25 4/25 32
25 0 0 14/5 32/5 2080
x
1
x
2
x
3
w
1
w
2
RHS
0 w
1
150
50 x
3
50
25 45 0 0 5 2500
Thank you

For any Query or suggestion :
Contact :
Dr. D. B. Naik
Professor, Training & Placement Section
Sardar Vallabhbhai National Institute of
Technology, Surat
Ichchhanath, Surat 395007, Gujarat.

Email: [email protected]
Phone no. 0261-2255225 (Direct)

You might also like