Sensitivity/Post Optimal Analysis

 Sensitivity analysis (or post-optimality

analysis) is used to determine how the optimal
solution is affected by changes, within specified
ranges, in:
 the objective function coefficients
 the right-hand side (RHS) values
 introduction of a constraint
 deletion of a variable
 Sensitivity analysis is important to a manager
who must operate in a dynamic environment
with imprecise estimates of the coefficients.

12/07/21 1
Discrete Variation in the Cost
Vector (C)-Example
Discuss the effect of discrete changes
in the cost vector for the following
L.P.P.
Maximize Z = 3X1+4X2+X3+7X4
S.t. 8X1+3X2+4X3+X4≤7
2X1+6X2+X3+5X4≤3
X1+4X2+5X3+2X4≤8
X1, X2, X3, X4≥0
12/07/21 2
Example Contd…
Let us first introduced three slack
variables S1≥0, S2≥0 and S3≥0.
Therefore the initial basic feasible
solution is S1 =7, S2 =3 and S3 =8.

12/07/21 3
 Example Contd…
Initial Table:
3 4 1 7 0 0 0 Sol
B.V. CB X1 X2 X3 X4 S1 S2 S3 XB XBi/yir
S1 0 8 3 4 1 1 0 0 7 7/1
S2 0 2 6 1 5 0 1 0 3 3/5
S3 0 1 4 5 2 0 0 1 8 8/2
Cj-Zj 3 4 1 7 0 0 0

12/07/21 4
 Example Contd…
First Iteration: X4 will enter and S2 will depart
3 4 1 7 0 0 0 Sol
B.V. CB XB XBi/yir
X1 X2 X3 X4 S1 S2 S3
S1 0 38/5 9/5 19/5 0 1 -1/5 0 32/5 32/38
X4 7 2/5 6/5 1/5 1 0 1/5 0 3/5 3/2
S3 0 1/5 8/5 23/5 0 0 -2/5 1 34/5 34/1

Cj-Zj 1/5 -22/5 -2/5 0 0 -7/5 0

12/07/21 5
 Example Contd…
Second Iteration: X1 will enter and S1 will depart
3 4 1 7 0 0 0 Sol
B.V. CB XB
X1 X2 X3 X4 S1 S2 S3
X1 3 1 9/38 19/38 0 5/38 -1/38 0 16/19
X4 7 0 21/19 0 1 -1/19 8/38 0 5/19
S3 0 0 59/38 9/2 0 -1/38 -15/38 1 126/19

Cj-Zj 0 -169/38 -1/2 0 -1/38 -53/38 0

Since the optimality condition is satisfied. Therefore, the

optimal solution has been obtained.

12/07/21 6
Example Contd…
Here the non- basic variables are X2 and X3.
Therefore
ΔC2≤169/38 and ΔC3≤1/2.
Now considering the basic variables
For X1,
  169 / 38  1 / 2  1 / 38 
Max  , ,   C1
 9 / 38 1 / 2 5 / 38 
  53 / 38 
 Min  
  1 / 38 
1
i.e.  C1  53
5
12/07/21 7
 Example Contd…
1
i.e. 3   C1  3  53
5
14
or  C1  56
5
Under this range the optimality of the solution is not changed.
Similarly for X4, -169/42 ≤ ΔC4 ≤ ½ or 125/42 ≤ C4 ≤ 15/2

12/07/21 8
Introduction of a Constraint-
Example
Consider the following table which present
an optimal solution to some L.P.P.
2 4 1 3 2 0 0 0 Sol
B.V. CB X1 X2 X3 X4 X5 X6 X7 X8 XB

X1 2 1 0 0 -1 0 0.5 0.2 -1 3
X2 4 0 1 0 2 1 -1 0 0.5 1
X3 1 0 0 1 -1 -2 5 -0.3 2 7

Cj –Zj 0 0 0 -2 0 -2 -0.1 -2

12/07/21 9
Example Contd….
If the additional Constraint
2X1+3X2-X3+2X4+4X5≤5
were annexed to the system, would
there be any change in the optimal
solution?
An optimal solution from the above
table is X1=3, X2=1, X3=7 and
X4=X5=X6=X7=X8=0
12/07/21 10
Example Contd….
We see that the above optimal
solution satisfies the new constraint.
Thus the optimum basic feasible
solution of the original problem will
be the optimum for the new problem
if we introduce a constraint.
Hence the additional constraint is
redundant (it does not effect the
optimal solution).

12/07/21 11
Example Contd….
If the additional Constraint be
2X1+3X2-X3+2X4+4X5≤1
were annexed to the system, would there
be any change in the optimal solution?
Clearly the optimal solution X1=3, X2=1,
X3=7 does not satisfy the additional
constraint.
Hence the optimal solution of the L.P.P. will
be changed if we introduce the above
constraint. The modified optimum simplex
table corresponding to the new constraint is

12/07/21 12
Example Contd…
2 4 1 3 2 0 0 0 0 Sol
B.V. CB X1 X2 X3 X4 X5 X6 X7 X8 X9 XB

X1 2 1 0 0 -1 0 0.5 0.2 -1 0 3
X2 4 0 1 0 2 1 -1 0 0.5 0 1
X3 1 0 0 1 -1 -2 5 -0.3 2 0 7
X9 0 2 3 -1 2 4 0 0 0 1 1

Cj -Zj

12/07/21 13
Example Contd…
Since X1, X2 and X3 are in the basis,
the corresponding coefficients in the
additional constraints must vanish.
Thus, we get (after applying
elementary row operation)

12/07/21 14
Example Contd…
2 4 1 3 2 0 0 0 0 Sol
B.V. CB X1 X2 X3 X4 X5 X6 X7 X8 X9 XB

X1 2 1 0 0 -1 0 0.5 0.2 -1 0 3
X2 4 0 1 0 2 1 -1 0 0.5 0 1
X3 1 0 0 1 -1 -2 5 -0.3 2 0 7
X9 0 0 3 -1 4 4 -1 -0.4 2 1 -5

Cj -Zj

12/07/21 15
Example Contd…
2 4 1 3 2 0 0 0 0 Sol
B.V. CB X1 X2 X3 X4 X5 X6 X7 X8 X9 XB

X1 2 1 0 0 -1 0 0.5 0.2 -1 0 3
X2 4 0 1 0 2 1 -1 0 0.5 0 1
X3 1 0 0 1 -1 -2 5 -0.3 2 0 7
X9 0 0 0 -1 -2 1 2 -0.4 0.5 1 -8

Cj -Zj

12/07/21 16
Example Contd…
2 4 1 3 2 0 0 0 0 Sol
B.V. CB X1 X2 X3 X4 X5 X6 X7 X8 X9 XB

X1 2 1 0 0 -1 0 0.5 0.2 -1 0 3
X2 4 0 1 0 2 1 -1 0 0.5 0 1
X3 1 0 0 1 -1 -2 5 -0.3 2 0 7
X9 0 0 0 0 -3 -1 7 -0.7 2.5 1 -1

Cj -Zj 0 0 0 -2 0 -2 -0.1 -2 0
2/3 0/-1 1/7

Here apply dual simplex method so X9 will depart and

X5 will enter.
12/07/21 17
Example Contd…
2 4 1 3 2 0 0 0 0 Sol
B.V. CB X1 X2 X3 X4 X5 X6 X7 X8 X9 XB

X1 2 1 0 0 -1 0 0.5 0.2 -1 0 3
X2 4 0 1 0 -1 0 6 -0.7 3 1 0
X3 1 0 0 1 5 0 -9 1.1 -3 -1 9
X5 2 0 0 0 3 1 -7 0.7 -2.5 -1 1

Cj -Zj 0 0 0 -2 0 -2 -0.1 -2 -1

Therefore optimal solution is X1*=3, X2*=0, X3*=9

and X5*=1
12/07/21 18
Deletion of a Variable
 In case deletion of a variable, check
whether the variable is a basic variable or
not.
 If it is not a basic variable then delete the
corresponding column from the optimum table.
 If it is a basic variable then assign a cost –M (in
case of maximization) and +M (in case of
minimization).
 Take the optimum table as the initial simplex
table for the new L.P.P.

12/07/21 19
 Example- Deletion of a variable
Consider the following table which present
an optimal solution to some L.P.P.
2 4 1 3 2 0 0 0 Sol
B.V. CB X1 X2 X3 X4 x5 x6 X7 X8 XB

X1 2 1 0 0 -1 0 0.5 0.2 -1 3
X2 4 0 1 0 2 1 -1 0 0.5 1
X3 1 0 0 1 -1 -2 5 -0.3 2 7

Cj -Zj 0 0 0 -2 0 -2 -0.1 -2

Let the variable X2 be deleted from the L.P.P. of the same

Problem. Obtain the optimal solution to the new L.P.P.
12/07/21 20
Example- Deletion of a variable
Let the optimum table be the initial table of the
new problem
2 -M 1 3 2 0 0 0 Sol
B.V. CB X1 X2 X3 X4 x5 x6 X7 X8 XB

X1 2 1 0 0 -1 0 0.5 0.2 -1 3
X2 -M 0 1 0 2 1 -1 0 0.5 1
X3 1 0 0 1 -1 -2 5 -0.3 2 7

Cj -Zj 0 0 0 2M+6 M+4 -M-6 -0.1 -0.5M-1

12/07/21 21
Example- Deletion of a variable
First iteration: X4 will enter and X2 will depart

2 -M 1 3 2 0 0 0 Sol
B.V. CB X1 X2 X3 X4 x5 x6 X7 X8 XB

X1 2 1 1/2 0 0 1/2 0 0.2 -3/4 7/2

X4 3 0 1/2 0 1 1/2 -1/2 0 1/4 1/2
X3 1 0 1/2 1 0 -3/2 9/2 -0.4 9/4 15/2

Cj -Zj 0 -M-3 0 0 1 -3 -0.1 -3/2

12/07/21 22
 Example- Deletion of a variable
Second iteration: X5 will enter and X4 will depart

2 -M 1 3 2 0 0 0 Sol
B.V. CB X1 X2 X3 X4 x5 x6 X7 X8 XB

X1 2 1 0 0 -1 0 0.5 0.2 -1 3
X5 2 0 1 0 2 1 -1 0 0.5 1
X3 1 0 2 1 3 0 3 -0.3 3 9

Cj -Zj 0 -M-4 0 -2 0 -2 -0.1 -2

Hence the optimal solution is X1*=3, X3*=9, X5*=1

and Max Z* = 17
12/07/21 23
Discrete Variation in the
requirement Vector (b)
 The required range for discrete change in
Δbk for which the optimum basic feasible
solution remains optimum is

  X bi    X bi 
Max    bk  Min 
bik  0 b
 ik  bik  0
 bik 

12/07/21 24
 Example
 Referring to the same example we had
before
3 4 1 7 0 0 0 Sol
B.V. CB X1 X2 X3 X4 S1 S2 S3 XB

X1 3 1 9/38 19/38 0 5/38 -1/38 0 16/19

X4 7 0 21/19 0 1 -1/19 8/38 0 5/19
S3 0 0 59/38 9/2 0 -1/38 -15/38 1 126/19

Cj -Zj 0 -169/38 -1/2 0 -1/38 -53/38 0

12/07/21 25
Example Contd…
Basic inverse matrix corresponding to
slack variable

 5 / 38  1 / 38 0
 
1
 
B    1 / 19 8 / 38 0
 
  1 / 38  15 / 38 1
 

12/07/21 26
Example Contd…
  16 19    5 19  126 19 
Max    b1  Min  , 
 5 38    1 / 19  1 / 38 
i.e.  32 / 5  b1  5 or 7  32 / 5  b1  7  5
i.e. 3 / 5  b1  12
  5 19    16 19  126 19 
Max    b2  Min  , 
 8 / 38    1 / 38  15 / 38 
i.e.  5 / 4  b2  84 / 5 or 3  5 / 4  b2  3  84 / 5
i.e. 7 / 4  b2  99 / 5
Also, 126 / 19  b3    or 26 / 19  b3   

12/07/21 27