COM 3372

Management Science
Shanika Kumarasinghe

Lecturer (Probationary)

Department of Commerce

Faculty of management studies and commerce

University of Sri Jayewardenepura

 Session Three:
Linear Programming Solving
Simplex Method (SM)
Recommended Reading………………..

• Read Chapter 3 & 4,

• Hamdy A. Taha. (2017). Operations

Research: An Introduction, Tenth
Edition, Pearson Education Limited,
England.
Learning Outcome..
After Studying This Lesson, You Should Be Able To;

✓ Illustrate simplex method and optimal findings

 Solution method of Linear Programme
1) Graphical solution method (GSM)
2) Simplex Solution Method (SM)

Simplex Solution Method (SM)

We can use this method for more than two variables.

 Start

Develop LP Model

Develop simplex Model

Develop initial solution

Optimum? Yes Interpret the optimum

Cj-Z< 0 solution
No
Identifying the entering
variable End
Determining the leaving
variable

Develop a new solution

Step 1: Develop LP Model
Example:
Maximize Z = 40X + 50X 1 2 (Total Profit)

Subject to;

3X + 3X < 90
1 2 (Raw materials-Kgs)
2X + 4X < 80
1 2 (Labour-Hrs)

X1,x2 > 0 (Non-negativity)

Step 2: Develop the simplex Model
• In order to convert the constraint into equality, we must add a slack
variable to the left hand side of the less than or equal type constraints.
It represents unused resources.
Maximize Z = 40X1 + 50X2 + OS1 + OS2
Subject to;

3X1 + 3X2 + S1 = 90
2X1 + 4X2 + S2 = 80
X1,x2, S1,S2 > 0
Step 3: Develop the initial solution
Cb Cj Basis Solution X1 X2 S1 S2

ZJ
Cj-Zj

Cj = Unit contribution associated with each of the variables in the objective function.
Cb=Unit contribution of each basic variables (variables in the solution basis).
Basis= Variables of the current solution. These variables usually have non zero values.
Solution = the current solution values of the basic variables.
Zj (Variable column) = amount of profit lost for each unit of variable that is brought into the
solution at the current iteration.
Zj (Solution column)= Total profit
Cj-Zj=Net increase in profit (or cost in a minimization problem) associated with one unit of
each product that is brought into the solution at the current iteration.
 Initial solution (Table 1-Point 0)
Cb Cj Basis Solution 40X1 50X2 0S1 0S2
0 S1 90 3 3 1 0
0 S2 80 2 4 0 1
ZJ 0 0 0 0 0
Cj-Zj 40 50 0 0

Step 4: Test the optimality

Since Cj-Zj > 0, further can be improved. If answer is 0 and negative, answer is optimal.
Step 5: Identify the entering variable

Select the variable with largest positive Cj-Zj value (in maximization problem)
Step 6: Determining the leaving variable

Select the minimum non negative value when the solution values are divided by the
coefficient in the pivot column.
Step 6: Develop a new solution
i. Dividing all row values by the pivot element, yield new row values

New value (Pivot row) = Old value / Pivot element

II. Calculate new values in the other constraints rows

New value (Other rows) = Old value – (row element*new value in pivot row)

2nd Tableau
Cb Cj Basis Solution 40X1 50X2 0S1 0S2
0 S1 30 3/2 0 1 -3/4
0 X2 20 1/2 1 0 1/4
ZJ 1,000 25 50 0 50/4
Cj-Zj 15 0 0 -50/4
Step 6: Develop a new solution
i. Dividing all row values by the pivot element, yield new row values

New value (Pivot row) = Old value / Pivot element

II. Calculate new values in the other constraints rows

New value (Other rows) = Old value – (row element*new value in pivot row)

3rd Tableau (Point B)

Cb Cj Basis Solution 40X1 50X2 0S1 0S2

40 X1 20 1 0 2/3 -1/2
50 X2 10 0 1 -1/3 1/2
ZJ 1,300 40 50 10 5
Cj-Zj 0 0 -10 -5

Test the optimality

Since all cj-zj values are < 0, (zero or minus), the solution is optimal.
• Test the optimality

• Since all cj-zj values are < 0, (zero or minus), the solution is
optimal.

Step 8: Interpret the optimum solution

Basic variables Non basic variables Total profit

X1 =20, X2=10 S1,S2 Z=1300

Shadow prices
S1=Rs.10; S2=Rs.5
Can you calculate the total profit using shadow prices???
CJ - ZJ Maximization
ZJ – CJ Minimization
 Initial solution (Table 1-Point 0)

Cb Cj Basis Solution 40X1 50X2 0S1 0S2

0 S1 90 3 3 1 0
0 S2 80 2 4 0 1
ZJ 0 0 0 0 0
Cj-Zj 40 50 0 0

Substitution Rate: Numbers in the body of the table.

Example – Coefficients under X1 are (3/2).

For every unit of the X1 product introduced into the current solution, 3 units of S1 and 2
units of S2 must be removed from the solution.
Example 2
• Minimization problem
Minimize Z = 15X + 10X (Cost ingredients A & B)
1 2

Subject to;

X + X > 300
1 2 (Demand)
3X > 150
1 (Demand)

X1,X2 > 0 (Non-negativity)

Constraints with < Add a slack variable (s).

Constraints with > We have to use two variables.

1. Subtract a surplus variable (S) – It shows how much the

solution exceeds the constraint resource.

2. Add an Artificial variable (A) - To avoid negative values of the

variables in the initial solution.

Coefficient of objective function of the artificial variable should be a big

M. As the artificial variable does not have any meaning (physical), it must
be kept out of any final solution.
Mixed Constraints problem
Minimize Z = 5X + 6X
1 2

Subject to;

X + X = 1000
1 2 (Demand)
X < 300
1 (Demand)
X > 150 2

X1,X2 > 0 (Non-negativity)

www.facebook.com/mgt.sjp.ac.lk https://mgt.sjp.ac.lk/