Linear Programming Problem

(Simplex Method)
Course Code: MAT2003

Dr. Sukanta Nayak

Department of Mathematics
School of Advanced
Sciences
VIT-AP University, Andhra Pradesh
Email: sukantgacr@gmail.com
 Outline
 General Form of LPP

 Canonical and Standard Forms of LPP

 Special Examples

 Practice Set
 General form of LPP
Find the unknown (variables) 𝑥1, 𝑥2, … , 𝑥𝑛 which maximize (or
minimize) the objective function.
𝑍 = 𝑐1𝑥1 + 𝑐2𝑥2 + ⋯ + 𝑐𝑛𝑥𝑛
Subject to the constraints
𝑎11𝑥1 + 𝑎12𝑥2 + ⋯ + 𝑎1𝑛𝑥𝑛 ≤ 𝑏1
𝑎21𝑥1 + 𝑎22𝑥2 + ⋯ + 𝑎2𝑛𝑥𝑛 ≤ 𝑏2
⋮
𝑎𝑚1𝑥1 + 𝑎𝑚2𝑥2 + ⋯ + 𝑎𝑚𝑛𝑥𝑛 ≤ 𝑏𝑚
and the non-negativity restrictions
𝑥1, 𝑥2, … , 𝑥𝑛 ≥ 0,
where, 𝑐𝑗, 𝑏𝑖, 𝑎𝑖𝑗; 𝑖 = 1,2, … , 𝑚; 𝑗 = 1, 2, … , 𝑛 are constants.
3
 General form of LPP
Case 1: If the constraints of a general LPP are of the following type
σ𝑛𝑗=1 𝑎𝑖𝑗𝑥𝑗 ≤ 𝑏𝑖, 𝑖 = 1, 2, 3, … , 𝑚
then the non-negative variables 𝑠𝑖 which satisfy
σ𝑛𝑗=1 𝑎𝑖𝑗𝑥𝑗 + 𝑠𝑖 = 𝑏𝑖, 𝑖 = 1, 2, 3, … , 𝑚
are called the slack variables.
Case 2: If the constraints of a general LPP are of the following type
σ𝑛𝑗=1 𝑎𝑖𝑗𝑥𝑗 ≥ 𝑏𝑖, 𝑖 = 1, 2, 3, … , 𝑚
then the non-negative variables 𝑠𝑖 which satisfy
σ𝑛𝑗=1 𝑎𝑖𝑗𝑥𝑗 − 𝑠𝑖 = 𝑏𝑖, 𝑖 = 1, 2, 3, … , 𝑚
are called the surplus variables.

4
 Canonical and standard forms of LPP
Canonical form
The general LPP is very often expressed in the following
form Maximize 𝑍 = 𝑐1𝑥1 + 𝑐2𝑥2 + ⋯ + 𝑐𝑛𝑥𝑛
subject to the constraints
𝑎𝑖1𝑥1 + 𝑎𝑖2𝑥2 + ⋯ + 𝑎𝑖𝑛𝑥𝑛 ≤ 𝑏𝑖; 𝑖 = 1, 2, … , 𝑚
𝑥1, 𝑥2, … , 𝑥𝑛 ≥ 0,
by doing some elementary transformations. The current form of the
LPP is known as canonical form and it has the following
characteristics.
 The objective function is of a maximization type,
 All constraints are of (≤) type,
 All variables 𝑥𝑖 are non-negative.
5
 Canonical and standard forms of LPP
Standard form
The general LPP can also be written in the following
form Maximize 𝑍 = 𝑐1𝑥1 + 𝑐2𝑥2 + ⋯ + 𝑐𝑛𝑥𝑛
subject to the constraints
𝑎𝑖1𝑥1 + 𝑎𝑖2𝑥2 + ⋯ + 𝑎𝑖𝑛𝑥𝑛 = 𝑏𝑖; 𝑖 = 1, 2, … , 𝑚
𝑥1, 𝑥2, … , 𝑥𝑛 ≥ 0.
The current form of the LPP is known as standard form and it has the
following characteristics.
 The objective function is of a maximization type,
 All the constraints are expressed as equations (equality),
 The right-hand side of each constraint is non-negative,
 All variables are non-negative.
6
 Special Examples

1. Convert the following LPP to the standard form

Maximize 𝑍 = 2𝑥1 + 3𝑥2 + 5𝑥3,
subject to
7𝑥1 − 4𝑥2 ≤ 5,
5𝑥1 + 2𝑥2 + 3𝑥3 ≥ 13,
6𝑥1 + 4𝑥3 ≤ 3,
𝑥1, 𝑥2 ≥ 0.

7
 Special Examples

Solution:
Given LPP, the variable 𝑥3 is unrestricted. So we may take
𝑥3 = 𝑥 ′ − 𝑥 ′′, where 𝑥 ′ , 𝑥′′ ≥ 0.
3 3 3 3
Now, we may write the given constraints in the following form
7𝑥1 − 4𝑥2 ≤ 5,
5𝑥1 + 2𝑥2 + 3𝑥′ − 3𝑥′′ ≥ 13,
3 3
′ ′′
6𝑥1 + 4𝑥 − 4𝑥 ≤ 3,
3 3
′ ′′
𝑥1, 𝑥2, 𝑥 , 𝑥 ≥ 0.
3 3

8
 Special Examples

It can be converted into standard form by introducing the

slack/surplus variables which results
Maximize 𝑍 = 2𝑥1 + 3𝑥2 + 5𝑥′ − 5𝑥′′,
3 3
subject to
7𝑥1 − 4𝑥2 + 𝑠1 = 5,
5𝑥1 + 2𝑥2 + 3𝑥′ − 3𝑥′′ − 𝑠2 = 13,
3 3
′ ′′
6𝑥1 + 4𝑥 − 4𝑥 + 𝑠3 = 3,
3 3
′ ′′
𝑥1, 𝑥2, 𝑥 , 𝑥 , 𝑠1, 𝑠2, 𝑠3 ≥ 0.
3 3

9
 Special Examples

2. Express the following problem in standard form.

Minimize 𝑍 = 4𝑥1 + 5𝑥2
subject to
2𝑥1 − 𝑥2 − 3𝑥3 = −4,
3𝑥1 + 4𝑥2 + 𝑥4 = 10,
𝑥1 − 5𝑥2 = 15,
𝑥1, 𝑥3, 𝑥4 ≥ 0.

10
 Special Examples
Solution:
The current problem, as 𝑥3 and 𝑥4 are the slack/surplus variables.
Whereas, 𝑥1 and 𝑥2 are the decision variables. Also, the variable 𝑥2
is unrestricted.
So, we may consider 𝑥2 = 𝑥 ′ − 𝑥′′ where 𝑥 ′ , 𝑥′′ ≥ 0.
2 2 2 2
We can write the problem in standard form as given below.
Maximize 𝑍′ = −𝑍 = −4𝑥1 − 5𝑥′ + 5𝑥′′
2 2
subject to
−2𝑥1 + 𝑥 ′ − 𝑥′′ + 3𝑥3 = 4
2 2
3𝑥1 + 4𝑥 − 4𝑥′′ + 𝑥4 = 10
′

2 2
𝑥1 − 5𝑥 + 5𝑥′′ = 15
′

2 2
′ ′′
𝑥1, 𝑥 , 𝑥 , 𝑥3, 𝑥4 ≥ 0.
2 2
11
Practice Set

12
Practice Set

13
 References
1. S. S. Bhavikatti (2010), Fundamentals of Optimum Design in
Engineering. New Delhi: New Age International (P) Limited
Publishers.
2. K. Deb (2004), Optimization for Engineering Design
Algorithms and Examples. New Delhi: Prentice-Hall of India.
3. S. Nayak (2020), Fundamentals of Optimization Techniques
with Algorithms. Academic Press, San Diego, California, USA.
4. S. S. Rao (1995), Optimization Theory and Applications. New
Delhi: New Age International (P) Limited Publishers.

14
15