Chapter 4:

Duality and Post-Optimal Analysis

Operations Research: An Introduction, 10th Ed.

By Taha, H.A.

ISE 303 - Term 201 Dr. Yasser Almoghathawi, KFUPM 1

Introduction
 For every LP formulation, there exists another unique linear
programming formulation called the 'Dual' (the original
formulation is called the 'Primal')
 The dual formulation can be derived from the same data
from which the primal was formulated
 The dual can be considered as the 'inverse' of the primal in
every respect
 The dual formulation can be solved in the same manner in
which the primal is solved since the dual is also a LP
formulation
ISE 303 - Term 201 Dr. Yasser Almoghathawi, KFUPM 2
Dual Problem Construction
 If the primal is a maximization problem, then its dual is a
minimization problem (and vise versa)
 Use the variable type of one problem to find the constraint
type of the other problem
 The RHS elements of one problem become the objective
function coefficients of the other problem (and vice versa)
 The matrix coefficients of the constraints of one problem is
the transpose of the matrix coefficients of the constraints
for the other problem
ISE 303 - Term 201 Dr. Yasser Almoghathawi, KFUPM 3
Duality Formulation
 The following is a summary of how the dual is constructed
from the primal:
1. Convert the primal problem into Standard form
2. A dual variable is assigned to each primal (equation) constraint
3. A dual constraint is assigned to each primal variable
4. The right-hand sides of the primal constraints provide the
coefficients of the dual objective function
5. The coefficients of the primal objective function provide the
right-hand sides of the dual constraints

ISE 303 - Term 201 Dr. Yasser Almoghathawi, KFUPM 4

Duality Formulation
 The following is a summary of how the dual is constructed
from the primal: (Cont.)
6. The dual constraint corresponding to a primal variable is
constructed by transposing the primal variable column into a
row with
 That is, the columns of primal problem are changed to rows in the
dual problem (i.e., the matrix A of the primal problem will be changed
to its transpose (A) for the dual problem)

ISE 303 - Term 201 Dr. Yasser Almoghathawi, KFUPM 5

Duality Formulation
 The following is a summary of how the dual is constructed
from the primal: (Cont.)
7. The sense of optimization, direction of inequalities, and the
signs of the variables in the dual are shown below:

Dual
Primal
Objective Variables
Objective Constraint type
sign
Max Min ≥ Unrestricted
Min Max ≤ Unrestricted

Note: the slack/surplus variables of the primal constraints identify the range of
each associated dual variable.
ISE 303 - Term 201 Dr. Yasser Almoghathawi, KFUPM 6
 The Dual of a Minimization LP
Primal LP Dual LP
Maximize Minimize
+ + … + + + … +
Subject to Subject to
+ +… + ≤ + +… + ≥
+ +… + ≤ + +… + ≥

+ +…+ ≤ + +…+ ≥
, , …. , ≥0 , , …. , ≥0
The dual of the dual takes us back to the primal
ISE 303 - Term 201 Dr. Yasser Almoghathawi, KFUPM 7
 Duality Formulation: Example
 Example 4.1-1:
Primal Standard Form Dual
Max = 5 + 12 +4 Max = 5 + 12 +4 +0 Min = 10 +8

subject to subject to subject to

+2 + ≤ 10 +2 + + = 10 +2 ≥5

2 − +3 =8 2 − +3 +0 =8 2 − ≥ 12

, , ≥ 0 , , , ≥ 0 +3 ≥4
+0 ≥0

, unrestricted
≥ 0, unrestricted

ISE 303 - Term 201 Dr. Yasser Almoghathawi, KFUPM 8

Duality Formulation: Summary

Max problem Min problem

Constraints Variables
≥ ⇔ ≤0
≤ ⇔ ≥0
= ⇔ Unrestricted
Variables Constraints
≥0 ⇔ ≥
≤0 ⇔ ≤
Unrestricted ⇔ =

ISE 303 - Term 201 Dr. Yasser Almoghathawi, KFUPM 9

 Primal-Dual Relationships
 If one problem (either primal or dual) is feasible the other will be
also feasible
• Objective function of a max LP ≤ Objective function of a min LP

 If one problem has an optimal feasible solution, other problem

also has an optimal feasible solution
• The optimal objective function value is same for both primal and dual
• Objective function of a max LP = Objective function of a min LP

 If one problem has no solution (infeasible), the other problem is

either infeasible or unbounded
 If one problem is unbounded, the other problem is infeasible
ISE 303 - Term 201 Dr. Yasser Almoghathawi, KFUPM 10
Primal-Dual Relationships
 Standard LP Starting Tableau
− − =0 Basic Var.s Sol.
− − 0
=

: original variables
: additional variables (e.g., slacks)

 Example:
Max = 5 +6 Basic Var.s Sol.

subject to -5 -6 0 0 0

+2 ≤ 20 1 2 1 0 20

2 + ≤ 16 2 1 0 1 16
, ≥ 0
ISE 303 - Term 201 Dr. Yasser Almoghathawi, KFUPM 11
Primal-Dual Relationships
 Standard LP Starting Tableau

− − =0 Basic Var.s Sol.

− − 0
=

: original variables
: additional variables (e.g., slacks)

Subsequent Tableau
Basic Var.s Sol.
− −

ISE 303 - Term 201 Dr. Yasser Almoghathawi, KFUPM 12

Primal-Dual Relationships
Subsequent Tableau
Basic Var.s Sol.
−

: basic = m x m matrix of original columns of current basic variables (in order)

: current basic variables (in order)
: original coefficients of z-row of current basic variables (in order)
: original coefficient column of variable j, (jth column)
: original coefficient of z-row of variable j,
: right-hand side

ISE 303 - Term 201 Dr. Yasser Almoghathawi, KFUPM 13

Primal-Dual Relationships
Primal Standard Form
 Example: Max = 5 +6 Max = 5 +6 +0 +0
subject to subject to
+2 ≤ 20 +2 + = 20
2 + ≤ 16 2 + + = 16
, ≥ 0 , , , ≥ 0

Dual Standard Form

Min = 20 + 16 Min = 20 + 16 +0 +0
subject to subject to
+2 ≥5 +2 − =5
2 + ≥6 2 + − =6
, ≥ 0 , , , ≥ 0
ISE 303 - Term 201 Dr. Yasser Almoghathawi, KFUPM 14
Primal-Dual Relationships
 Example: Construct the simplex tableau given that the basic
variables are and .
 Solution:
2 1 20
= , = 6 5 , =
1 2 16
2 11 0 1 0 2/3 −1/3
by row operations
1 20 1 0 1 −1/3 2/3
2/3 −1/3
= , = 7/3 4/3
−1/3 2/3
8
− = = = 68, = = =
4

ISE 303 - Term 201 Dr. Yasser Almoghathawi, KFUPM 15

Primal-Dual Relationships
 Example: Construct the simplex tableau given that the basic
variables are and .
 Solution:

Basic Var.s Sol.

0 0 7/3 4/3 68

0 1 2/3 -1/3 8

1 0 -1/3 2/3 4

ISE 303 - Term 201 Dr. Yasser Almoghathawi, KFUPM 16

 Optimal Dual Solution
Method I:
optimal primal z−row coef. of basic variable
= +
original objective coef. of basic variable

Example: starting slacks in primal correspond to dual variables

= +0= , = +0=

Method II:
*= Row vector of original objective optimal primal
x
coeff.s of optimal primal basic variables inverse
=
Example:
= = 7/3 4/3

ISE 303 - Term 201 Dr. Yasser Almoghathawi, KFUPM 17

 Simplex Tableau Computations
A. Constraint column computation
Constraint column Inverse in original constraint
= x
in iteration i iteration i column
=
Example:
2/3 −1/3 1 0
= = =
−1/3 2/3 2 1

B. Objective z-row computation

Primal Z−row coeff. LHS of the jth RHS of the jth
= −
of variable dual constraint dual constraint
Example:
Primal Z-row coeff. of 1 = LHS of 1st dual constraint – RHS of 1st dual constraint
= +2 −5= +2 −5= 5−5=0

ISE 303 - Term 201 Dr. Yasser Almoghathawi, KFUPM 18

 Economic Interpretations of Duality
Primal Dual

= =

subject to subject to

≤ , = 1, … , ≥ , = 1, … ,

≥ 0, = 1, … , ≥ 0, = 1, … ,

 Viewed as resource allocation model:

• The primal problem has economic activities and resources
• The coefficient in the primal represents the revenue per unit of activity and
resource with availability of is consumed at the rate units per unit of
activity
ISE 303 - Term 201 Dr. Yasser Almoghathawi, KFUPM 19
 Economic Interpretations of Duality
 Dual Variables
For any two primal and dual feasible solutions, the values of the
objective functions, when finite, must satisfy the following:

=∑ ≤∑ =

Obj. value of Obj. value of

≤
max problem min problem

At the optimum, the two objective values are equal (i.e., = w)

ISE 303 - Term 201 Dr. Yasser Almoghathawi, KFUPM 20

Economic Interpretations of Duality
 Dual Variables
Complimentary slackness implies that at optimal point, the following
holds:
∀ ∈{ ,…, } = 0 and ∀ ∈{ ,…, } =0

X =0 (vice versa)

• If a primal variable is basic: ≥ 0

 Its z-row coef. = LHS – RHS of dual constraint = 0
= slack of the dual constraint = 0
• If a primal variable is nonbasic: = 0
 Its z-row coef. = slack of the dual constraint ≠ 0
ISE 303 - Term 201 Dr. Yasser Almoghathawi, KFUPM 21
 Economic Interpretations of Duality
 Dual Variables
In terms of the resource allocation model,
• represents $ revenue
• represents available units of resource
• represents worth per unit of resource
• = implies
$ revenue = ∑
=∑ X($ )

• < implies
< ℎ
• As long as the revenue from all activities is less than the worth of resources,
the solution is not optimal
• Optimality reached when all resources are exploited completely
 i.e., input (worth of resources) = output (revenue)
ISE 303 - Term 201 Dr. Yasser Almoghathawi, KFUPM Example 22
 Economic Interpretations of Duality
 Dual Constraints
Objective coef. LHS of the jth RHS of the jth
= −
of ( ) dual constraint dual constraint
=∑ −
Reduced Represents profit
Represents cost
cost

$
$ cost = = X

Maximization optimality condition says that an increase in the level of

unused (nonbasic) activity can improve revenue only if its reduced cost is
negative
∑ < or < 0
ISE 303 - Term 201 Dr. Yasser Almoghathawi, KFUPM Example 23
 Dual Simplex Algorithm
 Consider the following example:
Min =2 1 + 15 2 + 18 3

subject to: • Some of the components of the

− 1+ 2 2 − 6 3 ≤ −10 constant vector b are negative, but
2 + 2 3 ≤ 6 • All the components of the
2 1 + 11 3 ≤ 19
objective vector c are nonnegative
− 1 − 2 ≤ −2
1, 2, 3 ≥ 0

 To solve this problem without the use of artificial variables we use a

method known as the Dual Simplex Algorithm
 This method is very similar to the Simplex method but uses some of the
negative components of the constant vector to obtain a pivot entry
ISE 303 - Term 201 Dr. Yasser Almoghathawi, KFUPM 24
Dual Simplex Algorithm
 Simplex starts from a feasible solution and try to get the
objective function satisfies optimality condition
• If max all coefficients non-negative
• If min all coefficients non-positive
• Feasible but nonoptimal

 Dual Simplex starts with an objective function that satisfies

the optimality conditions, but with an infeasible solution
• Hence, the aim being, iteration after iteration, is to get the solution
feasible
• Optimal but infeasible
ISE 303 - Term 201 Dr. Yasser Almoghathawi, KFUPM 25
 Dual Simplex Algorithm
A. Dual Feasibility Condition
• The leaving variable, , is the basic variable with the most
negative value
• If all the basic variables are nonnegative, the algorithm ends

B. Dual Optimality Condition

• is the reduced cost of nonbasic variable
• being the constraint coef. of the row of the leaving variable
• The entering variable is a non-basic variable that satisfies

min , <0

• If ≥ 0 for all nonbasic , the problem has no feasible solution

ISE 303 - Term 201 Dr. Yasser Almoghathawi, KFUPM 26
Dual Simplex Algorithm

 To start the LP optimal and infeasible, two requirements

must be met:

1. The objective function must satisfy the optimality

condition of the regular simplex method

2. All the constraints must be of type (≤)

ISE 303 - Term 201 Dr. Yasser Almoghathawi, KFUPM 27

Dual Simplex Algorithm
 Example: solve the following LP using dual simplex method
Standard Form
Min = 2 +3 Min = 2 +3 Min = 2 +3
subject to subject to subject to
2 +2 ≤ 30 2 +2 ≤ 30 2 +2 + = 30
+ 2 ≥ 10 − −2 ≤ −10 − −2 + = −10
, ≥ 0 , ≥ 0 , , , ≥ 0

 Since all the objective coefficients are  0 and the problem is

minimization, optimality is met
 We start the solution by writing the Simplex tableau

ISE 303 - Term 201 Dr. Yasser Almoghathawi, KFUPM 28

Dual Simplex Algorithm
 Example: solve the following LP using dual simplex method

Basic Var.s Sol.

-2 -3 0 0 0
2 2 1 0 30
-1 -2 0 1 -10

ISE 303 - Term 201 Dr. Yasser Almoghathawi, KFUPM 29

Dual Simplex Algorithm
 Example: solve the following LP using dual simplex method

Enter
Basic Var.s Sol.
-2 -3 0 0 0
2 2 1 0 30
Leave -1 -2 0 1 -10 most negative
Ratio 2 3/2 ignore

Basic Var.s Sol.

-1/2 0 0 -3/2 15
Optimal solution:
1 0 1 1 20
=0, =5,
1/2 1 0 -1/2 5 = 15
ISE 303 - Term 201 Dr. Yasser Almoghathawi, KFUPM 30
Generalized Simplex Algorithm
 Recall that:
• Primal simplex algorithm starts feasible but nonoptimal
• Dual simplex algorithm start optimal but infeasible
• What if an LP model starts nonoptimal and infeasible?
• How can it be solved without using artificial variables/constraints?

 Method:
• FIRST: use dual simplex to get rid of infeasibility (without worrying
about optimality)
• SECOND: use primal simplex to find the optimum

• Can we do alternatively do the other way around? YES!

• FIRST: use primal simplex to secure optimality (without worrying
about feasibility)
• SECOND: use dual simplex to seek feasibility.

ISE 303 - Term 201 Dr. Yasser Almoghathawi, KFUPM 31

Generalized Simplex Algorithm
 Example:
Max = 2 3
subject to:
− 1+2 2 −2 3 ≥ 8
− 1+ 2+ 3 ≤ 4
2 1− 2+ 4 3 ≤ 10
1, 2, 3 ≥ 0

Basic Var.s RHS (Sol.)

0 0 -2 0 0 0 0
1 -2 2 1 0 0 -8
-1 1 1 0 1 0 4
2 -1 4 0 0 1 10

ISE 303 - Term 201 Dr. Yasser Almoghathawi, KFUPM 32

Generalized Simplex Algorithm
 Example:
First: Use dual simplex to secure feasibility

Enter
Basic Var.s RHS (Sol.)
0 0 -2 0 0 0 0
Leave 1 -2 2 1 0 0 -8 most negative
-1 1 1 0 1 0 4
2 -1 4 0 0 1 10
Ratio X 0 X

ISE 303 - Term 201 Dr. Yasser Almoghathawi, KFUPM 33

Generalized Simplex Algorithm
 Example:
Basic Var.s RHS (Sol.)
0 0 -2 0 0 0 0
-1/2 1 -1 -1/2 0 0 4
-1/2 0 2 1/2 1 0 0
3/2 0 3 -1/2 0 1 14

The solution is now feasible but nonoptimal!

Second: Use primal simplex to find the optimal solution

ISE 303 - Term 201 Dr. Yasser Almoghathawi, KFUPM 34

Generalized Simplex Algorithm
 Example:
Basic Var.s RHS (Sol.) Ratio
0 0 -2 0 0 0 0
-1/2 1 -1 -1/2 0 0 4 Ignore
-1/2 0 2 1/2 1 0 0 0/2=0
3/2 0 3 -1/2 0 1 14 14/3

Basic Var.s RHS (Sol.) Ratio

-1/2 0 0 1/2 1 0 0
-3/4 1 0 -1/4 1/2 0 4 Ignore
-1/4 0 1 1/4 1/2 0 0 Ignore
9/4 0 0 -5/4 -3/2 1 14 56/9
ISE 303 - Term 201 Dr. Yasser Almoghathawi, KFUPM 35
 Sensitivity and Post Optimal Analysis
 It may be desirable to study how the current optimal solution
changes when the parameters of an LP problem are changed

 The discrete changes in the parameters include

• changes in the values of or
• addition/deletion of some constraints/variables

 These parameter changes result in one of the following three

cases:
• the optimal solution remains unchanged
 i.e., the basic variables and their values remain unchanged
• the basic variables remain unchanged but their values change
• the basic variables as well as their values are changed

ISE 303 - Term 201 Dr. Yasser Almoghathawi, KFUPM 36

Sensitivity and Post Optimal Analysis
 Four cases to be considered:
• Changes in the right-hand side of the constraints
• Changes in the cost coefficients
• Addition of a new constraint
• Addition of new activity/variable

 Difference between Sensitivity and Post Optimal Analysis

• In Sensitivity Analysis, we are generally trying to compute the range
of change for which the current solution remains optimal
• In Post Optimal Analysis, given the change (known), how does the
solution change

ISE 303 - Term 201 Dr. Yasser Almoghathawi, KFUPM 37

Sensitivity and Post Optimal Analysis

Condition after parameters

Recommended action
change

Current solution remains

No further action is necessary
optimal and feasible

Current solution becomes Use dual simplex to recover

infeasible feasibility

Current solution becomes Use primal simplex to recover

nonoptimal optimality

Current solution becomes Use the generalized simplex to

nonoptimal and infeasible recover optimality and feasibility

ISE 303 - Term 201 Dr. Yasser Almoghathawi, KFUPM 38

Sensitivity and Post Optimal Analysis
 Example (TOYCO):
Max = 3 1 +2 2 + 5 3
subject to:
1+2 2 + 3 ≤ 430 (Operation 1)
3 1 + 2 3 ≤ 460 (Operation 2)
1 +4 2 ≤ 420 (Operation 3)
1, 2, 3 ≥ 0

Basic Var.s Sol.

Optimal Tableau

4 0 0 1 2 0 1350
-1/4 1 0 1/2 -1/4 0 100
3/2 0 1 0 1/2 0 230
2 0 0 -2 1 1 20
ISE 303 - Term 201 Dr. Yasser Almoghathawi, KFUPM 39
Sensitivity and Post Optimal Analysis
 Example (TOYCO):
Primal Dual
Max = 3 +2 +5 Min = 430 + 460 +420
subject to subject to
+2 + ≤ 430 +3 + ≥3
3 +2 ≤ 460 2 +4 ≥2
+4 ≤ 420 +2 ≥5
, , ≥0 , , ≥ 0
Optimal solution Optimal solution
= 0, = 100, = 230 = 1, = 2, = 0,
= $1350 = $1350
ISE 303 - Term 201 Dr. Yasser Almoghathawi, KFUPM 40
Sensitivity and Post Optimal Analysis
A. Changes Affecting Feasibility
1. Changes to the right-hand side
2. Addition of a new constraint

B. Changes Affecting Optimality

1. Changes in the objective function coefficients
2. Addition of a new activity/variable

ISE 303 - Term 201 Dr. Yasser Almoghathawi, KFUPM 41

 Sensitivity and Post Optimal Analysis
A. Changes Affecting Feasibility
1. Changes to the right-hand side

New RHS of tableau Inverse in New RHS of

= x
in iteration i iteration i constraints
Example A.1-1:
Suppose that TOYCO in increasing the daily capacity of operations 1, 2 and 3 to
600, 640, and 590 minutes, respectively. How would this change affect the
total revenue?
1/2 −1/4 0 600 140
= 0 1/2 0 640 = 320
−2 1 1 590 30
The current basic variables remain feasible at new values. So, the associated
optimum revenue is = 3 0 + 2 140 + 5 320 = $1880
ISE 303 - Term 201 Dr. Yasser Almoghathawi, KFUPM 42
 Sensitivity and Post Optimal Analysis
A. Changes Affecting Feasibility
1. Changes to the right-hand side

New RHS of tableau Inverse in New RHS of

= x
in iteration i iteration i constraints
Example A.1-2:
TOYCO would like to shift the slack capacity of operation 3 ( =20 min) to the
capacity of operation 1. How would this change impact the optimal solution?
1/2 −1/4 0 450 110
= 0 1/2 0 460 = 230
−2 1 1 400 −40
The resulting solution is infeasible ( = −40), which requires applying the
dual simplex to recover feasibility. First, modify the RHS of the tableau with
the new infeasible RHS. Then, apply the dual simplex.
ISE 303 - Term 201 Dr. Yasser Almoghathawi, KFUPM 43
Sensitivity and Post Optimal Analysis
A. Changes Affecting Feasibility
2. Addition of a new constraint
The addition of a new constraint can never improve the
current optimum objective value:
a. The new constraint is redundant
• it’ll have no effect on the current solution, otherwise

b. The current solution doesn’t satisfy the new constraint

• A new solution is determined by the dual simplex method

ISE 303 - Term 201 Dr. Yasser Almoghathawi, KFUPM 44

 Sensitivity and Post Optimal Analysis
A. Changes Affecting Feasibility
2. Addition of a new constraint

Example A.2-1:
Suppose TOYCO is changing the design of its toys and that change requires an
addition of a fourth assembly operation. The daily capacity of the new
operation is 500 minutes and the times per unit for the three products are 3,
1, and 1 minutes, respectively.
The new constraint for operation 4 is 3 + + ≤ 500
The constraint is redundant because its satisfied by the current optimum
solution = 0, = 100, and = 230. Hence, the current optimum
solution remains unchanged.

ISE 303 - Term 201 Dr. Yasser Almoghathawi, KFUPM 45

 Sensitivity and Post Optimal Analysis
A. Changes Affecting Feasibility
2. Addition of a new constraint

Example A.2-2:
Suppose TOYCO units times on the fourth operation are changed for the three
products to 3, 3, and 1 minutes, respectively.
The new constraint for operation 4 is 3 + 3 + ≤ 500
This constraint is NOT satisfied by the current optimum solution = 0, =
100, and = 230.
= 500 − 3 0 + 3 100 + 230 = −30
This means that the new constraint is not redundant. Add the new constraint
to the tableau. Make sure the tableau is consistent. Apply the dual simplex to
recover feasibility. The new optimum solution: = 0, = 90, = 230,
and = $1330 (Verify!)
ISE 303 - Term 201 Dr. Yasser Almoghathawi, KFUPM 46
Sensitivity and Post Optimal Analysis
B. Changes Affecting Optimality
1. Changes in the objective function coefficients
a. Compute the dual values
*= Row vector of original objective optimal primal
x
coeff.s of optimal primal basic variables inverse
=
b. Substitute the new dual values in the following formula to
determine the new reduced costs (z-row coefficients)
Primal z−row coeff. LHS of the jth RHS of the jth
= −
of variable dual constraint dual constraint
If the new z-row satisfies the optimality condition, the solution
remains unchanged. Otherwise, apply primal simplex to recover
optimality
ISE 303 - Term 201 Dr. Yasser Almoghathawi, KFUPM 47
 Sensitivity and Post Optimal Analysis
B. Changes Affecting Optimality
1. Changes in the objective function coefficients

Example B.1-1:
Suppose TOYCO is considering a revised pricing policy to stay competitor. The
new revenues are $2, $3, and $4 for the three products, respectively.
The new objective function is Max z = 2 + 3 + 4
1/2 −1/4 0
= 3 4 0 0 1/2 0 = 3/2 5/4 0
−2 1 1
Reduced cost of = + 3 + − 2 = 13/4
Reduced cost of = − 0 = 3/2
Reduced cost of = − 0 = 5/4
The computations show that the current solution remains optimal, = $1220
ISE 303 - Term 201 Dr. Yasser Almoghathawi, KFUPM 48
 Sensitivity and Post Optimal Analysis
B. Changes Affecting Optimality
1. Changes in the objective function coefficients

Example B.1-2:
Suppose TOYCO objective function is changed to
Max z = 6 + 3 + 4

Reduced cost of = + 3 + − 6 = −3/4

Reduced cost of = − 0 = 3/2
Reduced cost of = − 0 = 5/4
The reduced cost of shows that the current solution is not optimal. Update
the z-row in the tableau with the new values of reduced costs and apply the
primal simplex to recover optimality
The new optimal solution is: = 10, = 102.5, = 215, and
= $1227.50 (Verify!)
ISE 303 - Term 201 Dr. Yasser Almoghathawi, KFUPM 49
Sensitivity and Post Optimal Analysis
B. Changes Affecting Optimality
2. Addition of a new activity/variable
Compute the reduced cost of the new variable:
a. If it satisfies the optimality condition, its not profitable
• The current solution remains optimal
b. If it doesn’t satisfy the optimality condition, its profitable
• The current solution is not optimal. Apply primal simplex to
recover optimality

ISE 303 - Term 201 Dr. Yasser Almoghathawi, KFUPM 50

 Sensitivity and Post Optimal Analysis
B. Changes Affecting Optimality
2. Addition of a new activity/variable

Example:
TOYCO recognizes that toy trains are not are not profitable. Hence, they
replace them with a new product with estimated revenue of $4 per toy and
the assembly times of 1, 1, & 2 minutes on operations 1, 2, & 3, respectively.
Let represents the new product, given the optimal dual values are
= 1 2 0 , we get:
Reduced cost of = + + 2 − 4 = −1
Hence, the result show that the new product is profitable. Add a new column
for the new variable in the optimal tableau using . Apply the primal
simplex to recover optimality. The new optimal solution is: = 0, = 0,
= 125, = 210 and = $1465 (Verify!)
ISE 303 - Term 201 Dr. Yasser Almoghathawi, KFUPM 51