LESSON

5
LINEAR PROGRAMMING: SIMPLEX METHOD

CONTENTS
5.0 Aims and Objectives
5.1 Introduction
5.2 Additional Variables used in Solving LPP
5.3 Maximization Case
5.4 Solving LP Problems Using Computer with TORA
5.5 Minimization LP Problems
5.6 Big M Method
5.7 Degeneracy in LP Problems
5.8 Unbounded Solutions in LPP
5.9 Multiple Solutions in LPP
5.10 Duality in LP Problems
5.11 Sensitivity Analysis
5.12 Let us Sum Up
5.13 Lesson-end Activities
5.14 Keywords
5.15 Questions for Discussion
5.16 Terminal Questions
5.17 Model Answers to Questions for Discussion
5.18 Suggested Readings

5.0 AIMS AND OBJECTIVES

In the previous lesson we have learnt linear programming with the help of graphical now
we will learn the linear programming with the help of Simplex Method using minimization
and maximization problems and the degeneracy in LP problems and also the Duality and
Sensitivity Analysis.

5.1 INTRODUCTION
In practice, most problems contain more than two variables and are consequently too
large to be tackled by conventional means. Therefore, an algebraic technique is used to
solve large problems using Simplex Method. This method is carried out through iterative
process systematically step by step, and finally the maximum or minimum values of the
objective function are attained.
Quantitative Techniques The basic concepts of simplex method are explained using the Example 1.8 of the
for Management
packaging product mix problem illustrated in the previous chapter. The simplex method
solves the linear programming problem in iterations to improve the value of the objective
function. The simplex approach not only yields the optimal solution but also other valuable
information to perform economic and 'what if' analysis.

5.2 ADDITIONAL VARIABLES USED IN SOLVING LPP

Three types of additional variables are used in simplex method such as,
(a) Slack variables (S1, S2, S3..…Sn): Slack variables refer to the amount of unused
resources like raw materials, labour and money.
(b) Surplus variables (-S1, -S2, -S3..…-Sn): Surplus variable is the amount of resources
by which the left hand side of the equation exceeds the minimum limit.
(c) Artificial Variables (a1, a2, a3.. …an): Artificial variables are temporary slack
variables which are used for purposes of calculation, and are removed later.
The above variables are used to convert the inequalities into equality equations, as given
in the Table 5.1 below.
Table 5.1: Types of Additional Variables
Constraint Type Variable added Format
a) Less than or equal to < Add Slack Variable +S
b) Greater than or equal to > Subtract surplus variable and add -S+a
artificial variable
c) Equal to = Add artificial variable +a

5.3 MAXIMIZATION CASE

The packaging product mix problem is solved using simplex method.
Maximize Z = 6x1 + 4x2
Subject to constraints,
2x1+3x2 £ 120 (Cutting machine) .....................(i)
2x1+ x2 £ 60 (Pinning machine) ......................(ii)
where x1, x2 ³ 0
Considering the constraint for cutting machine,
2x1+ 3x2 ³ 120
The inequality indicates that the left-hand side of the constraints equation has some
amount of unused resources on cutting machine. To convert this inequality constraint
into an equation, introduce a slack variable, S3 which represents the unused resources.
Introducing the slack variable, we have the equation
2x1+ 3x2 + S3 = 120
Similarly for pinning machine, the equation is
2x1+ x2 + S4 = 60
The variables S3 and S4 are known as slack variables corresponding to the three constraints.
Now we have in all four variables (which includes slack variable) and two equations. If
any two variables are equated to zero, we can solve the three equations of the system in
144 two unknowns.
If variables x1 and x2 are equated to zero, Linear Programming :
Simplex Method
i.e., x1 = 0 and x2 = 0, then
S3 = 120
S4 = 60
This is the basic solution of the system, and variables S3 and S4 are known as Basic
Variables, SB while x1 and x2 known as Non-Basic Variables. If all the variables are
non-negative, a basic feasible solution of a linear programming problem is called a Basic
Feasible Solution.
Rewriting the constraints with slack variables gives us,
Zmax = 6x1 + 4x2 + 0S3 + 0S4
Subject to constraints,
2x1 + 3x2 + S3 = 120 ....................(i)
2x1 + x2 + S4 = 60 ....................(ii)
where x1, x2 ³ 0
Though there are many forms of presenting Simplex Table for calculation, we represent
the coefficients of variables in a tabular form as shown in Table 5.2.
Table 5.2: Co-efficients of Variables

Iteration Basic Solution X1 Minimum Equation

X2 S3 S4
Number Variables Value KC Ratio

0 S3 120 2 3 1 0 60

S4 60 2 1 0 1 30

– Zj 0 –6 –4 0 0

If the objective of the given problem is a maximization one, enter the co-efficient of the
objective function Zj with opposite sign as shown in Table 5.3. Take the most negative
coefficient of the objective function and that is the key column Kc. In this case, it is -6.
Find the ratio between the solution value and the key column coefficient and enter it in
the minimum ratio column. The intersecting coefficients of the key column and key row
are called the pivotal element i.e. 2. The variable corresponding to the key column is the
entering element of the next iteration table and the corresponding variable of the key
row is the leaving element of the next iteration table. In other words, x1 replaces S4 in the
next iteration table. Table 5.3 indicates the key column, key row and the pivotal element.
Table 5.3

Iteration Basic Solution X1 Minimum

X2 S3 S4 Equation
Number Variables Value KC Ratio

0 S3 120 2 3 1 0 60

Kr S4 60 2 1 0 1 30

-Zj 0 -6 -4 0 0
145
Quantitative Techniques In the next iteration, enter the basic variables by eliminating the leaving variable (i.e., key
for Management
row) and introducing the entering variable (i.e., key column). Make the pivotal element
as 1 and enter the values of other elements in that row accordingly. In this case, convert
the pivotal element value 2 as 1 in the next interation table. For this, divide the pivotal
element by 2. Similarly divide the other elements in that row by 2. The equation is S4 /2.
This row is called as Pivotal Equation Row Pe. The other co-efficients of the key column
in iteration Table 5.4 must be made as zero in the iteration Table 5.5. For this, a solver, Q,
is formed for easy calculation. Change the sign of the key column coefficient, multiply
with pivotal equation element and add with the corresponding variable to get the equation,
Solver, Q = SB + (–Kc ´ Pe)
The equations for the variables in the iteration number 1 of table 8 are,
For S3 Q = SB + (– Kc ´ Pe)
= S3 + (–2x Pe)
= S3 – 2Pe …………………………(i)
For – Z, Q = SB + (– Kc ´ Pe)
= – Z + ((– 6) ´ Pe)
= – Z + 6Pe …………………………(ii)
Using the equations (i) and (ii) the values of S3 and –Z for the values of Table 1 are
found as shown in Table 5.4
Table 5.4: S3 and –Z Values Calculated

Iteration Basic Solution X1 X2 Minimum

S3 S4 Equation
Number Variables Value KC KC Ratio

0 S3 120 2 3 1 0 60

Kr S4 60 2 1 0 1 30

– Zj 0 –6 –4 0 0

1
S3 60 0 2 1 –1 30 S3 – 2Pe
Kr

Pe x1 30 1 ½ 0 ½ 60 S4 / 2

– Zj 100 0 –1 0 3 – Z + 6Pe

Using these equations, enter the values of basic variables SB and objective function Z. If
all the values in the objective function are non-negative, the solution is optimal. Here, we
have one negative value – 1. Repeat the steps to find the key row and pivotal equation
values for the iteration 2 and check for optimality.
In the iteration 2 number of Table 5.5, all the values of Zj are non-negative, Zj ³ 0, hence
optimality is reached. The corresponding values of x1 and x2 for the final iteration table
gives the optimal values of the decision variables i.e., x1 = 15, x2 = 30. Substituting these
values in the objectives function equation, we get
Zmax = 6x1 + 4x2
= 6(15) + 4(30)
= 90 + 120

146 = Rs. 210.00

 Table 5.5: Iteration Table Linear Programming :
Simplex Method
Iteration Basic Solution Minimum
X1 X2 S3 S4 Equation
Number Variables Value Ratio

0 S3 120 2 3 1 0 60
S4 60 2 1 0 1 30
Kr – Zj 0 –6 –4 0 0

1 S3 60 0 2 1 –1 30 S3 – 2pe
Kr x1 30 1 ½ 0 ½ 60 S4/2
Pe – Zj 100 0 –1 0 3 – Z + 6Pe

2 Pe X2 30 0 1 ½ – S3/2
x1 15 1 0 – 1/2 S3 – Pe/2
– Zj 210 0 0 1/4 ¾ – Z + Pe
½ 5/2

The solution is,

x1 = 15 corrugated boxes are to be produced and
x2 = 30 carton boxes are to be produced to yield a
Profit, Zmax = Rs. 210.00

Summary of LPP Procedure

Step 1: Formulate the LP problem.
Step 2: Introduce slack /auxiliary variables.
if constraint type is £ introduce + S
if constraint type is ³ introduce – S + a and
if constraint type is = introduce a
Step 3: Find the initial basic solution.
Step 4: Establish a simplex table and enter all variable coefficients. If the objective
function is maximization, enter the opposite sign co-efficient and if minimization,
enter without changing the sign.
Step 5: Take the most negative coefficient in the objective function, Zj to identify the
key column (the corresponding variable is the entering variable of the next
iteration table).
Step 6: Find the ratio between the solution value and the coefficient of the key column.
Enter the values in the minimum ratio column.
Step 7: Take the minimum positive value available in the minimum ratio column to
identify the key row. (The corresponding variable is the leaving variable of the
table).
Step 8: The intersection element of the key column and key row is the pivotal element.
Step 9: Construct the next iteration table by eliminating the leaving variable and
introducing the entering variable.
Step 10: Convert the pivotal element as 1 in the next iteration table and compute the
other elements in that row accordingly. This is the pivotal equation row (not
key row).
Step 11: Other elements in the key column must be made zero. For simplicity, form the
equations as follows: Change the sign of the key column element, multiply with
pivotal equation element and add the corresponding variable. 147
Quantitative Techniques Step 12: Check the values of objective function. If there are negative values, the solution
for Management
is not an optimal one; go to step 5. Else, if all the values are positive, optimality
is reached. Non-negativity for objective function value is not considered. Write
down the values of x1, x2,……..xi and calculate the objective function for
maximization or minimization.
Note:
(i) If there are no x1, x2 variables in the final iteration table, the values of x1 and x2 are
zero.
(ii) Neglect the sign for objective function value in the final iteration table.

5.4 SOLVING LP PROBLEMS USING COMPUTER WITH

TORA
From the MAIN MENU, select LINEAR PROGRAMMING option, and enter the
input values of the previously discussed problem as shown in the Figure 5.1.

Figure 5.1: Solving LPP using Computer with TORA (Input Screen )
Click Solve Menu, and select Solve Problem → Algebraic → Iterations → All-Slack
Starting Solution. Now, click Go To Output screen, then the first iteration table will be
displayed. To select the entering variable, click a non-basic variable (if correct, the column
turns green). Similarly, select the leaving variable (if correct, the row turns red),
Figure 5.2.

Figure 5.2: Selecting the Leaving Variable (TORA, Output Screen)

148
Then click Next Iteration button to display the next iteration table as shown in Figure 5.3. Linear Programming :
Simplex Method

Figure 5.3: Next Iteration Table (TORA, Output Screen)

Again click next iteration button to get the third and final iteration table. A pop-up menu
also indicates that the solution has reached the optimal level. Now we can notice that all
the values in the objective function Zmax row are non-negative which indicates that the
solution is optimal. The final Iteration Table is shown in Figure 5.4.

Figure 5.4: Final Iteration Table (TORA, Output Screen)

From the final Iteration Table, the values of X1, X2 and Zmax are taken to the corresponding
values in the solution column (last column) of the simplex table.
i.e., Zmax = 210.00
X1 = 30.00
X2 = 15.00
Example 1: Solve the LP problem using Simplex method. Determine the following :
(a) What is the optimal solution?
(b) What is the value of the objective function? 149
Quantitative Techniques (c) Which constraint has excess resources and how much?
for Management
Zmax = 5x1 + 6x2
Subject to constraints,
2x1 + x2 £ 2000 ....................(i)
x1 £ 800 ....................(ii)
x2 £ 200 ....................(iii)
where x1, x2 ³ 0
Solution: Converting the inequality constraints by introducing the slack variables,
Zmax = 5x1 + 6x2 + 0S3 + 0S4 + 0S5
2x1 + x2 + S3 = 2000
x1 + S4 = 800
x2 + S5 = 200
Equate x1 and x2 to zero , to find the initial basic solution
2(0) + 0 + S3 =2000
0 + S4 = 800
0 + S5 = 200
The initial basic solution is,
S3 = 2000
S4 = 800
S5 = 200
Establish a simplex table to represent the co-efficient of variables for optimal computation
as shown in Table 5.6.
Table 5.6: Simplex Table
Iteration Basic Solution Min
X1 X2 S3 S4 S5 Equation
Number Variable Value Ratio
0 S3 2000 2 1 1 0 0 2000
S4 800 1 0 0 1 0 ∝
Kr S3 200 0 1 0 0 1 200
-Z 1200 -5 -6 0 0 0
1 S3 1800 2 0 1 0 -1 900 S3 – P e
Kr S4 800 1 0 0 1 0 800 S4
Pe X2 200 0 1 0 0 1 ∝ S5
-Z 1200 -5 0 0 0 6 – Z + 6Pe
2 S3 200 0 -2 1 -2 -1 S3 – 2Pe
Pe X1 800 1 0 0 1 0 S4
X2 200 0 1 0 0 1 X2
-Zj 5200 0 0 0 5 6 – Z + 5Pe

In the final table, all the values of –Zj are ³ 0, hence optimality is reached. The optimum
solution is,
(a) The value of x1 = 800 units

150 x2 = 200 units

(b) Objective function Zmax = 5x1 + 6x2 Linear Programming :
Simplex Method
= 5(800) + 6(200)
= Rs. 5200.00
(c) In the final iteration Table 5.2, slack variable S3 represents the first constraint,
therefore this constraint has excess unused resources of 200 units.

5.5 MINIMIZATION LP PROBLEMS

In real life we need to minimize cost or time in certain situations. The objective now is
minimization. Procedure for minimization problems is similar to maximization problems.
The only difference is, enter the coefficients of the objective function in the simplex
table without changing the sign.
Another way to solve minimization problems is by converting the objective function as a
maximization problem by multiplying the equation by (– 1).
For example, if the objective function is,
Minimize Z =10x1+ 5x2
Convert the objective function into maximization and solve
Maximize Z = – 10x1 – 5x2

5.6 BIG M METHOD

So far, we have seen the linear programming constraints with less than type. We come
across problems with ‘greater than’ and ‘equal to’ type also. Each of these types must
be converted as equations. In case of ‘greater than’ type, the constraints are rewritten
with a negative surplus variable S1 and by adding an artificial variable a. Artificial variables
are simply used for finding the initial basic solutions and are thereafter eliminated. In
case of an ‘equal to’ constraint, just add the artificial variable to the constraint.
The co-efficient of artificial variables a1, a2,….. are represented by a very high value M,
and hence the method is known as BIG-M Method.
Example 2: Solve the following LPP using Big M Method.
Minimize Z = 3x1 + x2
Subject to constraints
4x1 + x2 = 4 ....................(i)
5x1 + 3x2 ³ 7 ....................(ii)
3x1 + 2x2 £ 6 ....................(iii)
where x1 , x2 ³ 0
Solution: Introduce slack and auxiliary variables to represent in the standard form.
Constraint 4x1 + x2 = 4 is introduced by adding an artificial variable a1, i.e.,
4x1 + x2 + a1 = 4
Constraint, 5x1 + 3x2 ³ 7 is converted by subtracting a slack S1 and adding an auxiliary
variable a2.
5x1+ 3x2 – S1 + a2 = 7
Constraint 3x2 + 2x2 £ 6 is included with a slack variable S2
3x2 + 2x2 + S2 = 6
151
Quantitative Techniques The objective must also be altered if auxiliary variables exist. If the objective function is
for Management
minimization, the co-efficient of auxiliary variable is +M (and -M, in case of maximization)
The objective function is minimization,
Minimize Z = 3x1+ x2 + 0S1+ 0S2+ Ma1+ Ma2
Zmin = 3x1 + x2+ Ma1+ Ma2
The initial feasible solution is (Put x1, x2, S1 = 0)
a1 = 4
a2 = 7
s2 = 6
Establish a table as shown below and solve:
Table 5.7: Simplex Table

Iteration B asic S olution a a M in

X1 X2 S1 S2 E quation
N um ber V ariables V alue 1 2 R atio

0 Z 0 3 1 0 0 M M

Kr a1 4 4 1 0 0 1 0 0.75

a2 7 5 3 –1 0 0 1 1.6

S2 6 3 2 0 1 0 0 2 Z + (– M a 1 )
– M a2
Z1 – 11M 9M + 3 – 4M + 1 M 0 0 0

1 Pe X1 1 1 ¼ 0 0 4 a 1 /4

Kr a2 2 0 7/4 –1 0 1.14 a 2 – 5P e

S2 3 0 5/4 0 1 2.4 S 2 – 3P e

Z1 – 2M -3 0 7M /4 M 0 Z 1 + (9M –
+1/4 3) P e
2 x1 5/7 1 0 – 0 X 1 – P e /4
1/7

x2 8/7 0 1 – 0 a2
7/4
4/7

S2 22/14 0 0 10/ 1 Z 1 + (7M /4

14 – ¼ ) Pe

Z1 – 23/7 0 0 1/7 0

The solution is,

x1 = 5/7 or 0.71
x2 = 8/7 or 1.14
Zmin = 3 x 5 / 7 + 8/7
= 23/7 or 3.29

Check Your Progress 5.1

1. What are the different types of additional variables used in simplex method?
2. How will you introduce/auxiliary variables in solving LPT problem?
152 Contd....
 Notes: (a) Write your answer in the space given below. Linear Programming :
Simplex Method
(b) Please go through the lesson sub-head thoroughly you will get your
answers in it.
(c) This Chek Your Progress will help you to understand the lesson better.
Try to write answers for them, but do not submit your answers to the
university for assessment. These are for your practice only.
_____________________________________________________________________
____________________________________________________________________________________________________________________
_____________________________________________________________________
__________________________________________________________________
__________________________________________________________________

5.7 DEGENERACY IN LP PROBLEMS

In solving linear programming problem, while improving the basic solution, it may so
transpire that there is no scope to generate an optimal solution. This is known as
'degeneracy' in linear programming. This occurs when there is a tie in the minimum ratio
column. In other words, two or more values in the minimum ratio column are the same.
To resolve degeneracy, the following method is used. Divide the key column values (of
the tied rows) by the corresponding values of columns on the right side. This makes the
values unequal and the row with minimum ratio is the key row.
Example 3: Consider the following LPP,
Maximize Z = 2x1+ x2
Subject to constraints,
4x1 + 3x2 £ 12 ...................(i)
4x1+x2 £8 ...................(ii)
4x1 – x2 £8 ...................(iii)
Solution: Converting the inequality constraints by introducing the slack variables,
Maximize Z=2x1+ x2
Subject to constraints,
4x1+ 3x2+ S3 = 12 ...................(iv)
4x1+ x2 + S4 = 8 ...................(v)
4x1– x2+ S5 = 8 ...................(vi)
Equating x1, x2 = 0, we get
S3 = 12
S4 = 8
S5 = 8
Table 5.8: Illustrating Degeneracy
Iteration Basic Solution Minimum
X1 X2 S3 S4 S5 Equation
Number Variables Value Ratio
0 S3 12 4 3 1 0 0 3
S4 8 4 1 0 1 0 2
S5 8 4 -1 0 0 1 2
-Z 0 -2 -1 0 0 0 153
Quantitative Techniques For S4; 4/2 = 2
for Management tie
For S5; 4/2 = 2
After entering all the values in the first iteration table, the key column is -2, variable
corresponding is x1. To identify the key row there is tie between row S4 and row S5 with
same values of 2, which means degeneracy in solution. To break or to resolve this,
consider the column right side and divide the values of the key column values. We shall
consider column x2, the values corresponding to the tie values 1, –1. Divide the key
column values with these values, i.e., 1/4, –1/4 which is 0.25 and – 0.25. Now in selecting
the key row, always the minimum positive value is chosen i.e., row S4. Now, S4 is the
leaving variable and x1 is an entering variable of the next iteration table. The problem is
solved. Using computer and the solution is given in the Figure 5.5.

Figure 5.5: LPP Solved Using Computer with TORA (Output Screen)

5.8 UNBOUNDED SOLUTIONS IN LPP

In a linear programming problem, when a situation exists that the value objective function
can be increased infinitely, the problem is said to have an 'unbounded' solution. This can
be identified when all the values of key column are negative and hence minimum ratio
values cannot be found.
Table 5.9: Illustrating Unbounded Solution
Iteration Basic Solution Minimum
X1 X2 S3 S4 S5 Equation
Value Variable Value Ratio
1 S3 12 1 -2 1 0 0
X1 8 3 -1 0 1 0
S4 4 2 -4 0 0 1
-Z 0 -4 -8 0 0 0

For S3; 12/-2

For X1; 8 /-1 all values are negative
For S4; 4/-4
154
 Linear Programming :
5.9 MULTIPLE SOLUTIONS IN LPP Simplex Method

In the optimal iteration table if (Pj - Zj) value of one or more non-basic variable is equal
to 0, then the problem is said to have multiple or alternative solutions.
Table 5.10: Illustrating Multiple Solutions

Pj 4 1 0 0 0

Iteration Basic Solution Minimum

X1 X2 S3 S4 S5 Equation
Number Number Value Ratio
2 X2 6 5 2 0 1 0
S2 3 4 1 2 1 0
Zj 4 4 1 1 2 0
Pj – Zj 0 0 -1 -2 0

5.10 DUALITY IN LP PROBLEMS

All linear programming problems have another problem associated with them, which is
known as its dual. In other words, every minimization problem is associated with a
maximization problem and vice-versa. The original linear programming problem is known
as primal problem, and the derived problem is known as its dual problem. The optimal
solutions for the primal and dual problems are equivalent.
Conversion of primal to dual is done because of many reasons. The dual form of the
problem, in many cases, is simple and can be solved with ease. Moreover, the variables
of the dual problem contain information useful to management for analysis.

Procedure
Step 1: Convert the objective function if maximization in the primal into minimization in
the dual and vice versa. Write the equation considering the transpose of RHS of
the constraints
Step 2: The number of variables in the primal will be the number of constraints in the
dual and vice versa.
Step 3: The co-efficient in the objective function of the primal will be the RHS constraints
in the dual and vice versa.
Step 4: In forming the constraints for the dual, consider the transpose of the body matrix
of the primal problems.
Note: Constraint inequality signs are reversed
Example 4: Construct the dual to the primal problem
Maximize Z = 6x1 + 10x2
Subject to constraints,
2x1 + 8x2 £ 60 .......................(i)
3x1 + 5x2 £ 45 .......................(ii)
5x1 - 6x2 £ 10 .......................(iii)
x2 £ 40 .......................(iv)
where x1, x2 ³ 0
Solution:
Minimize W = 60y1 + 45y2 + 10y3 + 40y4 155
Quantitative Techniques Subject to constraints,
for Management
2y1+3y2+5y3+ 0y4 ³ 6
8y1+ 5y2 + 6y3+ y4 ³ 10
where y1, y2, y3, y4 ³ 0
Example 5: Construct a dual for the following primal
Minimize Z = 6x1 – 4x2+ 4x3
Subject to constraints,
6x1 – 10x2 + 4x3 ³ 14 ..................(i)
6x1+ 2x2 + 6x3 ³ 10 ..................(ii)
7x1 – 2x2 + 5x3 £ 20 ..................(iii)
x1 – 4x2 + 5x3 ³ 3 ..................(iv)
4x1+ 7x2 – 4x3 ³ 20 ..................(v)
where x1, x2, x3 ³ 0
Solution: Convert 'less than' constraints into 'greater than' type by multiplying by
(–1) on both sides (i.e., for e.g. iii).
6x1 – 10x2 + 4x3 ³ 14
6x1+ 2x2 + 6x3 ³ 10
– 7x1 + 2x2 – 5x3 ³ 20
x1 – 4x2 + 5x3 ³ 3
4x1 + 7x2 – 4x3 ³ 20
The dual for the primal problem is,
Maximize W = 14y1+10y2+20y3+3y4+20y5
Subject to constraints,
6y1+ 6y2 – 7y3+ y4 + 4y5 £ 6
10y1+ 2y2 + 2y3 – 4y4+7y5 £ – 4
4y1+ 6y2 – 5y3+ 5y4 – 4y5 £ 4
where y1, y2, y3, y4 and y5 ³ 0

5.11 SENSITIVITY ANALYSIS

Sensitivity analysis involves 'what if?' questions. In the real world, the situation is constantly
changing like change in raw material prices, decrease in machinery availability, increase
in profit on one product, and so on. It is important to decision makers for find out how
these changes affect the optimal solution. Sensitivity analysis can be used to provide
information and to determine solution with these changes.
Sensitivity analysis deals with making individual changes in the co-efficient of the objective
function and the right hand sides of the constraints. It is the study of how changes in the
co-efficient of a linear programming problem affect the optimal solution.
We can answer questions such as,
i. How will a change in an objective function co-efficient affect the optimal solution?
ii. How will a change in a right-hand side value for a constraint affect the optimal
156 solution?
For example, a company produces two products x1 and x2 with the use of three different Linear Programming :
Simplex Method
materials 1, 2 and 3. The availability of materials 1, 2 and 3 are 175, 50 and 150 respectively.
The profit for selling per unit of product x1 is Rs. 40 and that of x2 is Rs. 30. The raw
material requirements for the products are shown by equations, as given below.
Zmax = 40x1 + 30x2
Subject to constraints
4x1 + 5x2 £ 175 ....................(i)
2x2 £ 50 ....................(ii)
6x1 + 3x2 £ 150 ....................(iii)
where x1, x2 ³ 0
The optimal solution is
x1 = Rs. 12.50
x2 = Rs. 25.00
Zmax = 40 ´ 12.50 + 30 ´ 25.00
= Rs. 1250.00
The problem is solved using TORA software and the output screen showing sensitivity
analysis is given in Table 5.11.

Change in objective function co-efficients and effect on optimal solution

Table 5.11: Sensitivity Analysis Table Output

Referring to the current objective co-efficient, if the values of the objective function co-
efficient decrease by 16 (Min. obj. co-efficient) and increase by 20 (Max. obj. co-
efficient) there will not be any change in the optimal values of x 1 = 12.50 and
x2 = 25.00. But there will be a change in the optimal solution, i.e. Zmax. 157
Quantitative Techniques Note: This applies only when there is a change in any one of the co-efficients of variables
for Management
i.e., x1 or x2. Simultaneous changes in values of the co-efficients need to apply for 100
Percent Rule for objective function co-efficients.
For x1, Allowable decrease = Current value - Min. Obj. co-efficient
= 40 – 24
= Rs. 16 ------------------ (i)
Allowable increase = Max. Obj. co-efficient – Current value
= 60 – 40
= Rs. 20.00 ---------------- (ii)
Similarly, For x2, Allowable decrease = Rs. 10.00 ---------------- (iii)
Allowable increase = Rs. 20.00 --------------- (iv)
For example, if co-efficient of x 1 is increased to 48, then the increase is
48 – 40 = Rs. 8.00. From (ii), the allowable increase is 20, thus the increase in x1 co-
efficient is 8/20 = 0.40 or 40%.
Similarly,
If co-efficient of x2 is decreased to 27, then the decrease is 30 - 27 = Rs. 3.00.
From (iii), the allowable decrease is 10, thus the decrease in x2 co-efficient is 3/10 = 0.30
or 30%.
Therefore, the percentage of increase in x1 and the percentage of decrease in x2 is 40
and 30 respectively.
i.e. 40% + 30% = 70%
For all the objective function co-efficients that are changed, sum the percentage of the
allowable increase and allowable decrease. If the sum of the percentages is less than or
equal to 100%, the optimal solution does not change, i.e., x1 and x2 values will not change.
But Zmax will change, i.e.,
= 48(12.50) + 27(25)
= Rs. 1275.00
If the sum of the percentages of increase and decrease is greater than 100%, a different
optimal solution exists. A revised problem must be solved in order to determine the new
optimal values.
Change in the right-hand side constraints values and effect on optimal solution
Suppose an additional 40 kgs of material 3 is available, the right-hand side constraint
increases from 150 to 190 kgs.
Now, if the problem is solved, we get the optimal values as
x1 = 23.61, x2 = 16.11 and Zmax = 1427.78
From this, we can infer that an additional resources of 40 kgs increases the profit by
= 1427.78 – 1250 = Rs. 177.78
Therefore, for one kg or one unit increase, the profit will increase by
= 177.78 / 40
= Rs. 4.44
Dual price is the improvement in the value of the optimal solution per unit increase in the
158 right-hand side of a constraint. Hence, the dual price of material 3 is Rs 4.44 per kg.
Increase in material 2 will simply increase the unused material 2 rather than increase in Linear Programming :
Simplex Method
objective function. We cannot increase the RHS constraint values or the resources. If
the limit increases, there will be a change in the optimal values.
The limit values are given in Table 2.10, i.e., Min RHS and Max RHS values.
For example, for material 3, the dual price Rs. 4.44 applies only to the limit range 150 kgs
to 262.50 kgs.
Where there are simultaneous changes in more than one constraint RHS values, the 100
per cent Rule must be applied.
Reduced Cost

 Cost of consumed  Profit per unit 

   
Reduced cost / unit of activity =  resources per unit  −  of activity 
 of activity   
   
If the activity's reduced cost per unit is positive, then its unit cost of consumed resources
is higher than its unit profit, and the activity should be discarded. This means that the
value of its associated variable in the optimum solution should be zero.
Alternatively, an activity that is economically attractive will have a zero reduced cost in
the optimum solution signifying equilibrium between the output (unit profit) and the input
(unit cost of consumed resources).
In the problem, both x1 and x2 assume positive values in the optimum solution and hence
have zero reduced cost.
Considering one more variable x3 with profit Rs. 50
Zmax = 40x1 + 30x2 + 50x3
Subject to constraints,
4x1 + 5x2 + 6x3 £ 175 ....................(i)
2x2 + 1x3 £ 50 ....................(ii)
6x1 + 3x2 + 3x3 £ 150 ....................(iii)
where x1, x2, x3 ³ 0
The sensitivity analysis of the problem is shown in the computer output below in
Table 5.12.
Table 5.12: Sensitivity Analysis

159
Quantitative Techniques The reduced cost indicates how much the objective function co-efficient for a particular
for Management
variable would have to improve before that decision function assumes a positive value in
the optimal solution.
The reduced cost of Rs.12.50 for decision variable x2 tells us that the profit contribution
would have to increase to at least 30 + 12.50 = 42.50 before x3 could assume a positive
value in the optimal solution.

Check Your Progress 5.2

1 What is Duality concept?

2. What is meant by degeneracy in Linear Programming?
Notes: (a) Write your answer in the space given below.
(b) Please go through the lesson sub-head thoroughly you will get your
answers in it.
(c) This Chek Your Progress will help you to understand the lesson better.
Try to write answers for them, but do not submit your answers to the
university for assessment. These are for your practice only.
_____________________________________________________________________
____________________________________________________________________________________________________________________
_____________________________________________________________________
__________________________________________________________________
__________________________________________________________________

5.12 LET US SUM UP

Thus LP is a planning technique of selecting the best possible (optimal) strategy among
number of alternatives. The chosen strategy is said to be the best because it involves
minimization/maximization of source desired action e.g. maximization of profits,
minimization of costs, smoothening running of the business.

5.13 LESSON-END ACTIVITIES

1. Linear Programming is a general method usable for a wide range of problems.
Go to any nutrition center which sells health-food. Bring into play the applications
of LP in formation and building.
2. LP is no doubt an vital problem. Not in this counters of petty problems with only a
couple of variables, but is much bigger problem.
Exaggerate this logic with the help of illstrations which can be matched and linked
with you real-life-situations.

5.14 KEYWORDS
Slack
Simplex method
Surplus
160
Variable Linear Programming :
Simplex Method
Solution

5.15 QUESTIONS FOR DISCUSSION

1. Write True or False against each statement:
(a) Artificial variable are imaginary and do not have any physical meaning.
(b) Simplex method solve the LPP in iteration to enhance the value of the objective
function.
(c) Sensitivity analysis can not be used to provide information and to determine
solution with these changes.
2. Briefly comment on the following statement:
(a) Two or more entries in the ratio column.
(b) LP is a planning techniques.
(c) LP techniques are used to optimise the resources for best result.
(d) LP in a part of management science.
(e) Algebraic techniques is used to solve large problems using simplex method.

5.16 TERMINAL QUESTIONS

1. Explain the procedure involved in the simplex method to determine the optimum
solution.

2. What are slack, surplus and artificial variables ?

3. What is degeneracy in LP problems ? When does it occur ? How can degeneracy

problem be resolved ?

4. What is a basic variable and a non-basic variable ?

5. Explain what is an unbounded solution in LPP.

6. Differentiate between primal and dual problems.

7. Why is the simplex method more advantageous than the graphical method?

8. What are the rules in selecting key column, key row and pivotal element?

9. Discuss the role of sensitivity analysis in linear programming.

10. In sensitivity analysis, explain

i. The effect of change of objective function coefficients

ii. The effect of change in the right hand side of constraints

Exercise Problems
1. A company manufactures three products A, B and C, which require three raw
materials I, II and III. The table given below shows the amount of raw materials
required per kg of each product. The resource availability per day and the profit
contribution for each product is also given. 161
Quantitative Techniques
Product A B C Availability (kg)
for Management
Raw Material
I 4 1 6 800
II 5 6 8 1500
III 2 4 1 1200
Profit per unit (Rs) 9 10 6

i. Formulate the problem as a linear programming problem.

ii. Solve the problem and determine the optimal product mix.
2. A metal fabricator manufactures three types of windows. Each of the windows
needs four processes. The time taken on various machines differ due to the size of
windows. The time taken and available hours are given in the table below:
Window Type Cutting Heat Treating Forging Grinding

A 5 7 1 4

B 7 4 4 8

C 4 8 6 2

Available time (Hrs) 20 24 28 22

The profit contribution for windows A, B and C are Rs. 3.00, Rs. 4.00 and Rs. 5.00
respectively.
a. Formulate the problem.
b. Solve the problem using simplex method to maximize the profit.
c. Determine the excess time available in each processes and by how much.
3. Solve the following LPP using simplex method.
Maximize , Z = 2x1 + x2
Subject to constraints,
4x1+ 3x2 £12 .....................(i)
4x1+ x2 £ 8 .....................(ii)
4x1 – x2 £ 8 .....................(iii)
where x1, x2 ³ 0
4. Solve the following LPP:
Zmax = 20x1 + 28x2 + 23x3
Subject to constraints,
4x1+ 4x2 £ 75 ....................(i)
2x1+ x2 + 2x3 £ 100 ....................(ii)
3x1+ 2x2 + x3 £ 50 ....................(iii)
where x1, x2, x3 ³ 0
5. Three high precision products are manufactured by a Hi-Tech Machine Tools
Company. All the products must undergo process through three machining centers
A, B and C. The machine hours required per unit are,
162
 Linear Programming :
Machining Center Product
Simplex Method
I II III
A 2 4 6
B 3 6 2
C 3 2 1

The available time in machine hours per week is

Machining Center Machine Hours Per Week
A 150
B 100
C 120

It is estimated that the unit profits of the product are

Product Unit Profits (Rs)
I 3
II 4
III 6

a. Formulate the problem as a LPP.

b. Solve the problem to determine the optimal solution. What is the number of
units to be made on each product.
c. Does machining center C has any extra time to spare? If so, how much spare
time is available ?
d. If additional 10 machine hours are available with machining center A, then
what is the optimal product mix ? What is the change in the value of profit ?
6. Raghu Constructions is considering four projects over the next 3 years. The expected
returns of each project and cash outlays for these projects are listed in the tables
given. All values are in Lacs of Rupees.
Cash outlay (lakh Rs.)
Project Return
Year 1 Year 2 Year 3
1 12.32 11.10 9.50 42.25
2 11.15 9.75 8.11 31.20
3 7.65 5.50 4.75 15.10
4 10.71 10.31 7.77 12.05
Available funds 110.00 40.00 35.00
(lakh Rs.)

Raghu has to decide to undertake construction projects. Ignore the time value of
money. As a consultant, what suggestion you would like to give Raghu in deciding
about the projects to select. Determine the solution using TORA.
7. Solve the following LP Problem using Big M Method.
Minimize, Z = 2x1+ 9x2 + x3
Subject to constraints,
x1+ 4x2+ 2x3 ³ 5 .......................(i)
3x1 + x2 + 2x3 ³ 4 .......................(ii)
where x1, x2, x3 ³ 0
8. Solve the following LPP
Zmin = 4x1 + x2 163
Quantitative Techniques Subject to constraints,
for Management
3x + x2 = 3 .................(i)
4x +3x2 ³ 6 .................(ii)
x1 + 2x2 £ 3 ................(iii)
where x1, x2 ³ 0
9. Solve the following LPP. Find whether multiple or alternate solution exists
Maximize Z = 2x1+ 4x2 + 6x3
Subject to constraints,
10x1 + 4x2 + 6x3 £ 150 ..................(i)
8x1+ 6x2 + 2x3 £ 100 ..................(ii)
x1 + 2x2 + x3 £ 120 ....................(iii)
where x1, x2, x3 ³ 0
10. Write the dual of the following LP problem
Minimize Z = 6x1 – 4x2 + 4x3
Subject to constraints,
3x1 + 10x2 + 4x3 ³ 15 .......................(i)
12x1 + 2x2 + 5x3 ³ 4 ...................(ii)
5x1 – 4x2 – 2x3 £ 10 ...................(iii)
x1 – 3x2 + 6x3 ³ 3 ...................(iv)
4x1 + 9x2 – 4x3 ³ 2 ...................(v)
where x1, x2, x3 ³ 0
11. Obtain the dual of the following linear programming problem
Maximize Z , = 4x1 + 9x2 + 6x3
Subject to constraints,
10x1 + 10x2 – 2x3 £ 6 .......................(i)
-5x1 + 5x3 + 6x3 ³ 8 .......................(ii)
14x1 – 2x2 – 5x3 £ 20 ........................(iii)
5x1 – 4x2 +7x3 ³ 3 ........................(iv)
8x1+ 10x2 – 5x3 = 2 .......................(v)
where x1, x2, x3 ³ 0

5.17 MODEL ANSWERS TO QUESTIONS FOR

DISCUSSION
1. (a) True (b) True (c) False

5.18 SUGGESTED READINGS

Dantzig, G and M. Thapa, Linear Programming 1: Introduction, Springer,
New York 1997.
Simonnard M., Linear Programming. Englewood Cliffs, N.J. Prentice Hall, 1966.

164
Bersitman, D, and J Tsitsiklin, Introduction to Linear Optimization, Belmont. Mass:
Athena Publishing 1997.