MATH 4 – APPLIED MATHEMATICS

MODULE 1

OPTIMIZATION OF DISCRETE
MODEL:

Operation Research

By: MARY JANE ALIVIO-BUNDAC

 Module 1 – Optimization of Discrete Model

Lesson 1.3 Linear Programming - Simplex Method

The simplex method is also known as simplex algorithm. This is one of the
mathematical tools used for linear programming. As per the journal Computing in
Science and Engineering, this method is considered one of the top 10 algorithms
that originated during the twentieth century as developed by George Dantzig.

The simplex method presents an organized strategy for evaluating a feasible

region’s vertices. This helps to figure out the optimal value of the objective function.
Unlike in graphical method that can handle two to three variables, the simplex
method can handle infinite number of variables involved in linear programming
problems.

Terminologies:

Slack Variables (Sn) - variables that fill the gap between the left and the right
members of the inequalities. It represents the unused
quantities or idle resources in the constraints and has
zero contribution to profit or cost.

Artificial Variables (An) - variables used as a computational device in linear

programming that prevent a slack variable from
becoming negative as well as to prevent the equality
constraint from equating a constant to zero. These
variables contribute a large amount to cost in
minimization problem. It contributes to the objective an
amount greater than any of the coefficients of solution
variables which is in a power ten.

Cj Column - a column in a simplex table which represent a contribution to

profit or cost per unit for the solution variable. It
represents the objective function column.

Cj – Zj row - a row contains the net profit or loss.

Solution Variable Column - a column of entries of the given decision

variables in the simplex table.

Solution Values Column - a column in the simplex table indicating the

quantities of the values of the variables that are in the
solution.

Optimum Column - a column in a simplex table which has a greatest positive

value in the Cj-Zj row in the case of maximization
problem or which has the greatest negative value in the
case of minimization problem.

Pivotal Row - a row which contains the smallest quotient between the
solution value and the non-zero or non-negative entry in
the optimum column.
Pivot - an intersection between the optimum column and the pivotal row. A
number that serves as a divisor of the solution value and
which gives the smallest quotient.

Replacing row - a row which is also a pivotal row on the previous table. The
entries are the quotients after dividing the pivotal row by
the pivot.
 Module 1 – Optimization of Discrete Model

Steps in Solving Simplex Maximization Problems:

1. Formulate the constraints from the conditions of the problem.

2. Write the new program with slack variables on the objective function and
on the explicit constraints. (Convert the inequality explicit constraints to
equation).

3. Enter the numerical coefficients and variables found in the objective

function and equality constraints in the initial simplex table.

4. Compute the Cj and Zj entry values.

5. Determine the optimum column or entering variable by choosing the

greatest positive value in the difference of Cj and Zj row (Cj – Zj).

6. Get the quotient between the solution values column by the non-zero and
non-negative entries in the optimum column. The least quotient becomes
the pivotal row.

7. Divide all entries in the pivotal row by the pivot to determine the values for
the replacing row. Then enter the computed result in the next table.

8. Calculate the new entries for the remaining rows by adding the product of
the entries in the replacing row and the additive inverse of a number in the
optimum column and the old entries in the previous table.

9. Obtain the Cj and Zj values. Calculate the difference of Cj and Zj (Cj – Zj)
rows.

10. The decision or final solution in the optimum table has been obtained if all
of the entries in the Cj – Zj row are all zeros and negative. Otherwise,
return to step no. 5.

Summary of Converting Constraints to Equations in a Maximization Problem

➢ Add a slack variable on the left member of the constraint containing less than
or equal to symbol (≤).
➢ Multiply the constraint contains greater than or equal (≥) by (-1) in order to
reverse the symbol into less than or equal to (≤) then add a slack variable.
➢ Add a slack variable on the left member of the constraint containing equal to
(=) symbol.

Simplex Minimization Method

In simplex minimization method subtraction of slack variable is allowed, but

not in maximization method, for the reason that if we aim to minimize, it is logically
mean to subtract, but if we aim to maximize, it is otherwise.
 Module 1 – Optimization of Discrete Model

Steps in Solving Simplex Minimization Problems:

“The steps are almost the same to maximization problems except for the
following three procedures”.

1. Determine the Cj column in the initial table by entering all the coefficients of
artificial variables and slack variables in the objective function with positive
coefficients only in the constraints.

2. Look for the greatest negative entry in the Cj – Zj row.

3. The decision or final solution in the optimum table has entries in the Cj – Zj
row with zeros or positive.

Summary of Converting Constraints to Equations in a Minimization Problem

➢ Add a slack variable on the left member of the constraint containing less than or
equal to symbol (≤).
➢ Subtract a slack variable on the left member of the constraint but add an artificial
variable at the same time if the symbol is greater then or equal to (≥).
➢ Add an artificial variable on the left member of the constraint containing equal to
(=) symbol.

Examples:
Solve the following maximization problems using simplex method.

1. The Adlex Furniture Shop produces cabinets, chairs, and tables. Each cabinet
requires 6 minutes in machine center M, 5 minutes in machine center A, and 8
minutes in machine center B. Each chair produced requires 5, 4, and 5 minutes
in machine centers M, A, and B respectively. Likewise, each table produced
requires 6, 4, and 6 minutes in machine centers M, A, and B respectively. The
Shop has a daily availability of 2,400 minutes in center M, 2,000 minutes in center
A, and 3,000 minutes in machine center B. Each cabinet gives PhP600 profit, a
chair gives PhP200, and a table gives PhP500. Find the number of each type of
products to produce in order to maximize the Furniture Shop’s profit.

Solution:

Product Machine Machine Machine Profit

Center M Center A Center B
(in min.) (in min.) (in min.) (in peso)
Let x – be no. of cabinet 6 5 8 600
y – be no. of chairs 5 4 5 200
Z – be no. of tables 6 4 6 500
2,400 2,000 3,000

Objective Function and Constraints:

Maximize P: 600𝑥 + 200𝑦 + 500𝑧
6𝑥 + 5𝑦 + 6𝑧 ≤ 2,400
Subject to: {5𝑥 + 4𝑦 + 4𝑦 ≤ 2,000
8𝑥 + 5𝑦 + 6𝑧 ≤ 3,000
𝑥; 𝑦; 𝑧 ≥ 0
 Module 1 – Optimization of Discrete Model

New Program with Slack Variables (Sn):

Maximize P: 600𝑥 + 200𝑦 + 500𝑧 + 0𝑆1 + 0𝑆2 + 0𝑆3

6𝑥 + 5𝑦 + 6𝑧 + 𝑆1 = 2,400
Subject to: {5𝑥 + 4𝑦 + 4𝑦 + 𝑆2 = 2,000
8𝑥 + 5𝑦 + 6𝑧 + 𝑆3 = 3,000
𝑥; 𝑦; 𝑧 ≥ 0
𝑆1 ; 𝑆2 ; 𝑆3 ≥ 0

Table I. Initial Feasible Solution

Contribution to profit column
Solution variable column
Solution values column

600 200 500 0 0 0

Cj SVar SVal X Y Z S1 S2 S3
0 S1 2,400 6 5 6 1 0 0
0 S2 2,000 5 4 4 0 1 0

0 S3 3,000 8 5 6 0 0 1

Zj 0 0 0 0 0 0 0
Cj - Zj 600 200 500 0 0 0

Pivotal row Optimum column

Divide all entries in Row 3 by 8 to get the replacing row.

“A row with smallest quotient after dividing the solution value (3,000) by a
divisor (8) becomes a pivotal row”.

(3,000 8 5 6 0 0 1) ÷ 8

(375 1 5/8 ¾ 0 0 1/8) entries on replacing ro

Table II. Second Feasible Solution

Cj 600 200 500 0 0 0

SVar SVal X Y Z S1 S2 S3
0 S1 150 0 5/4 3/2 1 0 -3/4
0 S2 125 0 7/8 1/4 0 1 -5/8
600 X 375 1 5/8 3/4 0 0 1/8
Zj 225,000 600 375 450 0 0 75
Cj - Zj 0 -175 50 0 0 -75
 Module 1 – Optimization of Discrete Model

Row 1 of table II:

Multiply the entries in the replacing row by the additive inverse of a
number in the optimum column (-6), and add the entries of Row 1 in
table I.

375(−6) + 2,400 = 150

1(−6) + 6 = 0
5 5
(−6) + 5 =
8 4
3 3
(−6) + 6 =
4 2
0(−6) + 1 = 1
0(−6) + 0 = 0
1 −3
(−6) + 0 =
8 4

Row 1 of table II:

Multiply the entries in the replacing row by the additive inverse of a
number in the optimum column (-6), and add the entries of Row 1 in
table I.

375(−6) + 2,400 = 150

1(−6) + 6 = 0
5 5
(−6) + 5 =
8 4
3 3
(−6) + 6 =
4 2
0(−6) + 1 = 1
0(−6) + 0 = 0
1 −3
(−6) + 0 =
8 4

Row 2 of table II:

Multiply the entries in the replacing row by the additive inverse of a

number in
the optimum column (-5), and add the entries of Row 2 in table I.

375(−5) + 2,000 = 125

1(−5) + 5 = 0
5 7
(−5) + 4 =
8 8
3 1
(−5) + 4 =
4 4
0(−5) + 0 = 0
0(−5) + 1 = 1
1 −5
(−5) + 0 =
8 8

3
Divide Row 1 in table II by 2 to get the entries in the replacing row.
“A row with smallest quotient after dividing the solution value (150) by a
3
divisor (2) belongs to a pivotal row.”

5 3 −3 3
(150 0 1 0 )÷2
4 2 4
5 2 −1
(100 0 1 0 ) entries on replacing row
6 3 2
 Module 1 – Optimization of Discrete Model

Table III. Third Feasible Solution

Cj 600 200 500 0 0 0

SVar SVal X Y Z S1 S2 S3
500 Z 100 0 5/6 1 2/3 0 -1/2
0 S2 100 0 2/3 0 -1/6 1 -1/2
600 X 300 1 0 0 -1/2 0 1/2
Zj 180,000 600 2,500/6 500 100/3 0 50
Cj - Zj 0 - 0 - 0 -50
1,300/6 100/3

Row 2 of table III:

Multiply the entries in the replacing row by the additive inverse of a

number in the optimum column (-1/4), and add the old entries of Row 2
in table II.

100 (-1/4) + 125 = 100

0 (-1/4) + 0 = 0
5/6 (-1/4) + 7/8 = 2/3
1 (-1/4) + ¼ = 0
2/3 (-1/4) + 0 = -1/6
0 (-1/4) + 1 = 1
-1/2 (-1/4) + (-1/8) = -1/2

Row 3 of table III:

Multiply the entries in the replacing row by the additive inverse of a

number in
the optimum column (-3/4), and add the old entries of Row 3 in table II.

100 (-3/4) + 375 = 300

0 (-3/4) + 1 = 1
5/6 (-3/4) + 5/8 = 0
1 (-3/4) + 3/4 = 0
2/3 (-3/4) + 0 = -1/2
0 (-3/4) + 0 = 0
-1/2 (-3/4) + (1/8) = ½

Decision:
Maximum Profit: PhP180,500
X = 300 pcs. of Cabinet
Y = 0 no. of Chair
Z = 100 pcs. of Tables

The Adlex Furniture Shop should produce 300 pieces of cabinets and
100 pieces of tables in order to realize a maximum profit of PhP180,000.
 Module 1 – Optimization of Discrete Model

2. Maximize Z: 10𝑥 + 12𝑦

𝑥≤2
Subject to: { 𝑦≥6
𝑥 + 𝑦 ≤ 10
𝑥≥0
𝑦≥0

Multiply the constraint with greater than or equal to (≥) symbol by (-1) to
reverse into less than or equal to (≤) symbol, then add a slack variable.
−(𝑦 ≥ 6) −𝑦 ≤ −6
Maximize Z: 10𝑥 + 12𝑦 + 0𝑆1 + 0𝑆2 + 0𝑆3
𝑥 + 𝑆1 = 2
Subject to: { −𝑦 + 𝑆2 = 6
𝑥 + 𝑦 + 𝑆3 = 10
𝑥≥0
𝑦≥0
𝑆1 ≥ 0
𝑆2 ≥ 0
𝑆3 ≥ 0

Table I. Initial Feasible Solution

Cj 10 12 0 0 0
SVar SVal X1 X2 S1 S2 S3
0 S1 2 1 0 1 0 0

0 S2 -6 0 -1 0 1 0

0 S2 10 1 1 0 0 1

Zj 0 0 0 0 0 0

Cj - Zj 10 12 0 0 0

Divide all entries in Row 3 by 1, a non-zero and non-negative entry

in the optimum column to get the replacing row.

(10 1 1 0 0 1) ÷ 1 (10 1 1 0 0 1) replacing row

Table II. Second Feasible Solution

Cj 10 12 0 0 0
SVar SVal X1 X2 S1 S2 S3
0 S1 2 1 0 1 0 0

0 S2 4 1 0 0 0 1

12 X2 10 1 1 0 0 1

Zj 120 12 12 0 0 12

Cj - Zj -2 0 0 0 -12
 Module 1 – Optimization of Discrete Model

Row 2 of Table II:

10 (1) + (- 6) = 4
1 (1) + 0 = 1
1 (1) + (-1) = 0
0 (1) + 0 = 0
0 (1) + 0 = 0
1 (1) + 1 = 1

Decision :
Maximum Z: 120
x = 0 ; y = 10

3. Minimize Z: 3𝑥1 + 5𝑥2 + 5𝑥3

𝑥1 + 𝑥2 + 𝑥3 ≤ 60
𝑥1 + 𝑥3 ≤ 40
3𝑥1 + 2𝑥2 ≥ 30
Subject to:
𝑥1 ≥ 0
𝑥2 ≥ 0
{ 𝑥3 ≥ 0

New program with Slack and Artificial Variables:

Minimize Z: 3𝑥1 + 5𝑥2 + 5𝑥3 + 0𝑆1 + 0𝑆2 + 0𝑆3 + 10𝐴1
𝑥1 + 𝑥2 + 𝑥3 + 𝑆1 = 60
𝑥1 + 𝑥3 + 𝑆2 = 40
3𝑥1 + 2𝑥2 − 𝑆3 + 𝐴1 = 30
𝑥1 ≥ 0
Subject to:
𝑥2 ≥ 0
𝑥3 ≥ 0
𝑆1 , 𝑆2 , 𝑆3 ≥ 0
{ 𝐴1 ≥ 0

Table I. Initial Feasible Solution

Cj 3 5 5 0 0 0 10
SVar SVal X1 X2 X3 S1 S2 S3 A1
0 S1 60 1 1 1 1 0 0 0
0 S2 40 1 0 1 0 1 0 0
10 A1 30 3 2 0 0 0 -1 1
Zj 300 30 20 0 0 0 -10 10
Cj - Zj -27 -15 5 0 0 10 0

Row 3 a replacing row:

(30 3 2 0 0 0 -1 1) ÷ 3 (10 1 2/3 0 0 0 -1/3 1/3)
 Module 1 – Optimization of Discrete Model

Table II. Second Feasible Solution

Cj 3 5 5 0 0 0 10
SVar SVal X1 X2 X3 S1 S2 S3 A1
0 S1 50 0 1/3 1 1 0 1/3 -1/3
0 S2 30 0 -2/3 1 0 1 1/3 -1/3
3 X1 10 1 2/3 0 0 0 -1/3 1/3
Zj 30 3 2 0 0 0 -1 1
Cj - Zj 0 3 5 0 0 1 9

Row 1 of table II: Row 1 of table II:

10 (-1) + 60 = 50 10 (-1) + 40 = 30
1 (-1) + 1 = 0 1 (-1) + 1 = 0
2/3(-1) + 1 = 1/3 2/3(-1) + 0 = -2/3
0 (-1) + 1 = 1 0 (-1) + 1 = 1
0 (-1) + 1 = 1 0 (-1) + 0 = 0
0 (-1) + 0 = 0 0 (-1) + 1 = 1
-1/3 (-1) + 0 = 1/3 -1/3 (-1) + 0 = 1/3
1/3 (-1) + 0 = -1/3 1/3 (-1) + 0 = -1/3

Decision:
X1 = 10, and
Minimum Z = 30

Exercises:

In each of the following, write the new program with slack(Sn) and/or artificial
(An) variables , then solve by simplex method.

1. Maximize: 100𝑥1 + 200𝑥2 New Program with Sn Variables:

Subject to: 3𝑥1 + 𝑥2 ≤ 6 Maximize:
2𝑥1 + 3𝑥2 ≥ 12 Subject to:
𝑥1 ≥ 0
𝑥2 ≥ 0

2. Minimize: 2𝑥1 + 10𝑥2 New Program with Sn and An Variables:

Subject to: 𝑥1 + 𝑥2 ≥ 3 Minimize:
𝑥1 + 𝑥2 ≤ 6 Subject to:
𝑥1 ≥ 0
𝑥2 ≥ 0

Seatwork:

Solve the following problem by simplex method.

Alex company’s profit is P = 300X + 200Y + 100Z where X, Y, and Z are three
types of products produced. The profit function is subject to the following
constraints: 𝑋 + 𝑌 ≤ 200, 𝑍 ≤ 50, 𝑋 + 𝑍 ≥ 150, 𝑋 ≥ 0, 𝑌 ≥ 0, 𝑍 ≥ 0. Find
the values of X, Y, and Z to realize the maximum profit.
 Module 1 – Optimization of Discrete Model

Problems Set:
I.
In each of the following, write the new program with slack(Sn) and/or artificial
(An) variables , then solve by simplex method.

1. Maximize: 50𝑥 + 40𝑦 New Program with Slack Variables:

Subject to: 2𝑥 + 𝑦 = 12 Maximize:
2𝑥 + 3𝑦 ≤ 24 Subject to:
2𝑥 + 9𝑦 ≥
𝑥≥0
𝑦≥0

2. Maximize: 8𝑥 + 7𝑦 New Program with Slack Variables:

Subject to: 𝑥 + 𝑦 ≤ 10 Maximize:
𝑥+𝑦 ≥2 Subject to:
𝑥≤5
𝑦≤6
𝑥≥0
𝑦≥0

3. Minimize: 10𝑥1 + 2𝑥2 New Program with (Sn) and An Variables:

Subject to: 𝑥1 + 𝑥2 ≥ 10 Minimize:
3𝑥1 + 𝑥2 ≤ 12 Subject to:
𝑥1 ≥ 0
𝑥2 ≥ 0
II.
Solve the following problems using simplex method.

1. A Marine Biologist team which is composed of a senior marine

biologist, an assistant marine biologist and a secretary who receive
PhP800, PhP600 and PhP350 per day respectively must finish a
particular aqua marine project within a period of 100 days. However, it
is stipulated that the senior biologist and the assistant biologist render
at least 12 days and 25 days respectively, and that the secretary be on
the no more than 70 days. Set up a linear programming model that will
optimize the cost of the project.

2. Mr. Bundy owns a 10-hectare farm, which he plants peanuts and corn.
The peanut needs 4 units of insecticides and 3 units of fertilizers per
hectare, while the corn needs 3 units of insecticides and 6 units of
fertilizers. There are available 36 units of insecticides and 48 units of
fertilizers. The peanuts average contribution to profit is PhP20,800 per
hectare while the corn is
PhP15,000 per hectare. His objective is to optimize his profit. How
many hectares should he plant with peanuts and with corn.

3. Three foods are to be purchased in amounts X1, X2, and X3. Their
calories and vitamin contents, minimum requirements and prices are
shown below:
X1 X2 X3 Minimum
Requirements
Calorie 4 2 5 26
Vitamin 1 4 4 13
Price per unit PhP18 PhP90 PhP130

How much of X1, X2, and X3 must be purchased to minimize the cost?
Module 1 – Optimization of Discrete Model
Module 1 – Optimization of Discrete Model