WORKBOOK FOR

NUR INTAN SYAFINAZ AHMAD

GHAZALI SEMIL@IS MAIL
ZOKREE ABDUL KARIM
NAZUHA MUDA @ YUSOF
WAN KAROMIAH WAN ABDULLAH

FIRST EDITION
2
 CHAPTER 4: LINEAR
PROGRAMMING

Learning Objective (LO):

After completing this chapter, students will be able to

 Formulate linear programming model
 Solve linear programming problem using graphical method
 Solve linear programming problem (maximation & minimization) using simplex
method

3
3.1 Introduction and Requirements of Linear Programming (LP)
Problem
Mathematical modeling technique in which a linear function is maximized or minimized when
subjected to various constraints. This technique has been useful for guiding quantitative
decisions in business planning, in industrial engineering, and to a lesser extent. Linear
programming (LP) is a technique that helps in resource allocation decisions

A. Requirements of LP problem

4
B. Basic Assumptions of LP

3.2 Formulating LP problems

A. LP model

 Involves developing a mathematical model to represent the managerial problem.

 The steps are:

5
 Most common LP application: product mix problems.

Use LP to decide how much of each product to make, given a series of resource restrictions

GUIDED EXAMPLE 1

As a farmer, Zoul wants to determine the number of acres of land to be planted with mangos
and guavas. He has 80 acres of land and has allocated a capital of RM10,000 to develop the
land. The expenditures needed to plant mangos and guavas per acre are RM100 and RM150
respectively. The average yields per acre are 500 kg of mangos and 300kg of guavas.

The fruits are kept in the backyard store before they are sold. The store can only hold 6000kg
of fruits at one time. The net profit of each kg of mangos and guavas are RM0.60 and RM0.90
respectively.

Formulate the above problem as a linear programming model to maximize the net profit.

6
Solution:

Step 2 & 3: Identify objective and constraints

 Objective: maximize profit
 Constraints:
 The expenditures needed to plant mangos and guavas per acre are RM100 and
RM150
 The average yields per acre are 500kg of mangos and 300kg of guavas

Step 4: Define the decision variables

 X  number of acres of land to be planted with mangos
 Y  number of acres of land to be planted with guavas
Now we can create the LP objective function in terms of X and Y .
 Maximize profit: 0.60 X  0.90Y

Step 5 & 6: Develop mathematical relationship to describe the 2 constraints.

The total expenditures

⚫ Total expenditure = (RM100/acre)(no. of acres of land to be planted with mangos) +
(RM150/acre)(no. of acres of land to be planted with guavas)
⚫ 1st constraint can be stated as:
 Expenditures needed is ≤ capital allocation to develop
 100 X  150Y  10,000 (The expenditures needed to plant mangos and guavas
per acre)

The average yields

 Total yields = (500kg/acre)(The average yields of mangos) + (300kg/acre)(The
average yields of guavas)
⚫ 2nd constraint can be stated as:
 Average yields per acre is ≤ The fruits are kept in the backyard store
 500X  300Y  6,000 (The average yields per acre of mangos and guavas)
⚫ Both constraints represent production capacity restrictions & affect the total profit

Step 7: To obtain meaningful solutions, the values of X and Y must be nonnegative

numbers.
 X  0 (number of acres of land to be planted with mangos is greater than or
equal to 0)
 Y  0 (number of acres of land to be planted with guavas is greater than or
equal to 0)

Step 8: Write the full formulation

Maximize profit Z  0.60 X  0.90Y

Subject to constraints: 100X  150Y  10,000 (expenditures needed to plant)

500X  300Y  6,000 (average yields per acre)
X,Y  0

7
 HANDS-ON EXAMPLE 1

A company makes two products P and Q using two machines A and B. Each unit of P that is
produced requires 50 minutes processing time on machine A and 30 minutes processing time
on machine B. Each unit of Q that is produced requires 24 minutes processing time on
machine A and 33 minutes processing time on machine B. Machine A is going to be available
for 40 hours and machine B is available for 35 hours. The profit per unit of P is RM25 and the
profit per unit of Q is RM30. Company policy is to determine the production quantity of each
product in such a way as to maximize the total profit given that the available resources should
not be exceeded. Formulate the problem as a linear programming model.

Solution: Ans: P  25x  30y , 50x  24y  2400 ,

30x  33y  2100 , x, y  0

HANDS-ON EXAMPLE 2

The Continuing Education Division of the UiTM Community College offers at most 30 courses
each semester. The courses offered are usually of two types: practical (such as woodworking,
wood processing and car maintenance) and humanistic (such as history, music, and fine arts).
To satisfy the demand of the community, at least 10 courses of each type must be offered
each semester. The division estimates that the profits of offering practical and humanistic
courses are approximately RM1500 and RM1000 per course, respectively. Formulate this
problem as a linear programming model.

8
Solution: Ans: Z  1500x  1000y , y  10 , y  10 , x  y  30 x, y  0

3.3 Graphical solution to LP problem(two-variable)

The graphical solution approach works only when there are 2 decisional variables, but it
provides valuable insight into how larger problems are structured. The following below are the
steps for graphical solution or method:

9
A. Feasible Solution And Corner Point Method/Iso-Profit Line Method/Iso-Cost Line
Method

1) The Feasible Solution

 Firstly, plot each of the problem’s constraint of a graph.

 Plotting involves finding points at which the line intersects the X and Y axes.
 Find the set of solution points that satisfies all the constraints simultaneously.
 The feasible region is the overlapping area of constraints that satisfies all the
restrictions on resources.

Any point in the region would be a feasible solution to the problem & any point outside the
shaded area would represent an infeasible solution

2) Corner Point Method

 Looking at the profit at every corner point of the feasible region.

 The mathematical theory behind LP is that the optimal solution must lie at one of the
corner points in the feasible region.

3) Iso-Profit (Objective) Line Method

2 Optimal
3 Solution

Feasible
Region

Iso-Profit
1 Line 4

10
4) Iso-Cost (Objective) Line Method

Iso-Cost
Line

Optimal 2
Solution
3

B. Solving Maximization and Minimization Problems

1) Solving Maximization Problem

 Maximizing an objective such as profit or revenue.

 Example of maximization problem:

11
 o A factory wishes to maximize píofits given labouí and mateíials
constíaint toincrease the production of the bowls and mugs.
 Maximization problem can be solved graphically by:
o Setting up the feasible solution region.
o Use either corner point method or iso-profit line method approach to find the values
of decision variables that yield the maximum profit.

GUIDED EXAMPLE 2

Solve the following linear programming problem by the graphical method.

Maximize profit Z  3 X  7Y
Subject to constraints: 0.10 X  0.30Y  240
X  2Y  2000
X  800
X,Y  0

Solution:

Draw the lines/equations (convert the inequalities sign into the equal sign)

Equation When Y  0 When X  0 Point

X Y  X,Y 
a) 0.10 X  0.30Y  240 2400 800 2400,0 0,800
b) X  2Y  2000 2000 1000 2000,0 0,1000
Other lines/equations: X  800, X  0,Y  0

Find the feasible region: check the constraints at origin 0,0

12
 Constraint when X  0 Result Area Colour line/arrow
&Y  0
1. 0.10 X  0.30Y  240 0  240 True left/below green
2. X  2Y  2000 0  2000 True left/below blue
3. X  800 0  800 True right yellow
4. X  0 00 True right red
5. Y  0 00 True above purple

Note: The feasible region consists of all colours (green, blue, yellow, red and purple)
constraint on the graph.

Feasible
Region

13
Method 1: Corner Point Method

Step 1: Find the coordinates of each corner point that had been labeled 1, 2, 3, 4.

Feasible
Region
4
2

*Point 1: Solve the intersection point, given that X  800, then substitute into
0.10 X  0.30Y  240
0.10(800)  0.30Y  240  Y  533.33

* Point 3: Solve the simultaneous equations:

0.10 X  0.30Y  240------ (1)
X  2Y  2000----- (2)
 X  1200,Y  400

Step 2: Evaluate the objective function at each corner point

Corner Point ( X,Y ) Z  3 X  7Y

11 (800,533.33) Z  3(800)  7(533.33)  6133.31
22 (800,0) Z  3(800)  7(0)  2400
33 (1200,400) Z  3(1200)  7(400)  6400
44 (2000,0) Z  3(2000)  7(0)  6000

Step 3: Select the corner point with the best (maximum/highest) value of the objective
function to maximize Z (profit)

Point 3
produces the highest profit of any corner point. Therefore, the value of X  1200
and Y  400 is the optimal solution. This solution yields a profit of RM6400 per production
period.

Step 4: Find the optimal solution based on the objective function to maximize Z (profit)

Optimal solution is at point 3

, where X  1200 and Y  400 with maximum profit
Z  RM 6400 . OR

14
Method 2: Iso-Profit Line Method

The optimal solution is the point lying in the feasible region that produces the highest profit.

Step 1: Gives some small number of profits. For example, Z  RM1470 is chosen (must be
in the feasible region).

Step 2:
Objective function: 1470  3 X  7Y
Calculate the objective function by replacing X  0  Y  210  0,210  and
Y  0  X  490  490,0

Feasible
Region
4
2

Step 3:
• Construct the line on the graph with points of 0,210 and 490,0 .
• If the line does not produce the highest possible profit, try graphing 2 more lines with higher
profit than before.
• Make sure that the iso-profit line must be parallel.
• Draw a series of parallel iso-profit lines until we find the highest iso-profit line, that is, the
one with the optimal solution.
• Therefore, the highest possible iso-profit line is the one that touches the tip of the feasible
region at the corner point (1200,400) 3 and yields a profit of RM6400 .

GUIDED EXAMPLE 3

From the solution of Guided Example 1, solve the following linear programming problem by
the graphical method.

Maximize profit Z  0.60 X  0.90Y

Subject to constraints: 100X  150Y  10,000 (expenditures needed to plant)

500X  300Y  6,000 (average yields per acre)

15
 X,Y  0
Solution:

Equation When Y  0 When X  0 Point

X Y  X,Y 
a) 100 X  150Y  10000 100 66.67 100,0 0,66.67
b) 500 X  300Y  6000 12 20 12,0 0,20
Other lines/equations: X  0,Y  0

Constraint when X  0 Result Area Colour line/arrow

&Y  0
1. 100 X  150Y  10000 0  10000 True left/below blue
2. 500 X  300Y  6000 0  6000 True left/below green
3. X  0 00 True right red
4. Y  0 00 True above purple

3
1

By using the corner point method (method 1), to find the optimal solution:

Corner Point ( X,Y ) Z  0.60 X  0.90Y

11 (0,0) Z  0.60(0)  0.90(0)  0
22 (0,20) Z  0.60(0)  0.90(20)  18
33 (12,0) Z  0.60(12)  0.90(0)  7.20

2
Point produces the highest profit of any corner point.

Optimal solution is at point 2 , where X  0 and Y  20 with maximum profit

Z  RM18

16
 HANDS-ON EXAMPLE 3

Solve the following linear programming problem using the graphical method.

Maximize Z  3X  2Y
Subject to:
2X  6Y  12
5 X  4Y  40
X  2Y  12
X,Y  0

Solution: Ans: optimal soln. x  5.33, y  3.33,z  22.65

17
 HANDS-ON EXAMPLE 4

Solve the following linear programming problem graphically.

Maximize z  8x  y
Subject to:
x  y  40
2x  y  60
x,y  0

Solution: Ans: optimal soln. x  30, y  0,z  240

18
2) Solving Minimization Problem

• Minimizing an objective such as cost instead of maximizing a profit function.

• Example of minimization problem:
o A restaurant wish to develop a work schedule to meet staffing needs while minimizing
the total number of employees.
• Minimization problem can be solved graphically by:
o Setting up the feasible solution region.
o Use either corner point method or iso-cost line method approach to find the values of
decision variables that yield the minimum cost.

GUIDED EXAMPLE 4

Solve the following linear programming problem graphically.

Minimize C  12x  4y
Subject to:
x  4y  12
3x  2y  18
y 6
x, y  0

Solution:
Draw the lines/equations (convert the inequalities sign into the equal sign)

Equation When Y  0 When X  0 Point

X Y  X,Y 
a) x  4y  12 12 3 12,0 0,3
b) 3x  2y  18 6 9 6,0 0,9
Other lines/equations: Y  6, X  0,Y  0

19
Find the feasible region: check the constraints at origin 0,0

Constraint when X  0 &Y  0 Result Area Colour line/arrow

a) x  4y  12 0  12 True right/above green
b) 3x  2y  18 0  18 True right/above blue
c) y  6 06 False below yellow
d) X  0 00 True right red
e) Y  0 00 True above purple
Note: The feasible region consists of all colours (green, blue, yellow, red and purple)
constraint on the graph.

20
Feasible Region

21
Method 1: Corner Point Method

Step 1: Find the coordinates of each corner point that had been labeled 1, 2 and 3.

Feasible Region
2
3

*Point 1: Solve the intersection point, given that Y  6, then substitute into
3x  2y  18
3x  2(6)  18  x  2

* Point 2: Solve the simultaneous equations:

3x  2y  18------ (1)
x  4y  12----- (2)
 x  4.8, y  1.8

Step 2: Evaluate the objective function at each corner point

Corner Point ( X,Y ) C  12x  4y

11 (2,6) C  12(2)  4(6)  48
22 (4.8,1.8) C  12(4.8)  4(1.8)  64.8
33 (12,0) C  12(12)  4(0)  144

Step 3: Select the corner point with the best (minimum/lowest) value of the objective
function to minimize C (cost)

Point 1 produces the highest profit of any corner point. Therefore, the value of x  2
and y  6 is the optimal solution. This solution yields a cost of RM48 per production
period.

Step 4: Find the optimal solution based on the objective function to minimize C (cost)

Optimal solution is at point 1 , where x  2 and y  6 with minimum cost C  RM 48

OR

22
Method 2: Iso-Profit Line Method

The optimal solution is the point lying in the feasible region that produces the lowest cost.

Step 1: Gives some largel number of costs. For example, C  RM180 is chosen (must be
in the feasible region).

Step 2:
Objective function: 180  12x  4y
Calculate the objective function by replacing x  0  y  45  0,45 and
y  0  x  15  15,0

Feasible Region
2

Step 3:
• Construct the line on the graph with points of 0,45 and 15,0 .
• If the line does not produce the lowest possible cost, try graphing 2 more lines with lower
cost than before.
• Make sure that the iso-cost line must be parallel.
• Draw a series of parallel iso-cost lines until we find the lowest iso-cost line, that is, the one
with the optimal solution.
• Therefore, the lowest possible iso-cost line is the one that touches the bottom of the
feasible region at the corner point (2,6) 1 and yields a cost of RM 48 .

23
 HANDS-ON EXAMPLE 5

Solve the following linear programming problem graphically.

Minimize cost C  6x  4y
Subject to:
xy  4
5x  3y  15
x, y  0

Solution: Ans: optimal soln. x  1.5, y  2.5,C  19

24
 HANDS-ON EXAMPLE 6

Given the following Linear Programming model for a company:

Minimize cost C  30x  50y

Subject to
20x  10y  160
xy4
2y  12
x,y  0
Solve the LP using graphical method.

Solution: Ans: optimal soln. x  4, y  0,C  RM120

25
C. Special Cases in LP

Infeasibility:

A situation where there is no solution to the problem. In practice,

this would be a situation where conditions on costs, staffing and
resources make it impossible to produce any goods.

Unboundedness:

This is a situation where there is no limit to the optimum value

being found. In practice, this situation is likely to occur if one or
more conditions have been omitted from the problem
formulation.

Redundancy:

A redundancy constraint is one that does not affect the feasible

region. This usually occurs in large LP formulation and causes
no major difficulties.

Multiple solutions:

The optimal value of the objective function occurs at more than 1 extreme points, then the
problem has multiple optimal solution. For example: At corner point of (a,b) and (c,d ) , both
value of the objective function Z  L (equal solution).

26
CHAPTER 5: SIMPLEX
METHOD
The Simplex method is an approach to solving linear programming models by hand using
slack variables, tableaus, and pivot variables to finding the optimal solution of an optimization
problem. The simplex method systematically examines corner points, using algebraic steps,
until an optimal solution is found.
⚫ Having more than two variables- too large for the simple graphical procedure.
⚫ Yields the optimal solution to the variables & the maximum profit or minimum cost.
⚫ The approach provides valuable economic information.

A. How to set up the Initial Simplex Tableau (maximization and minimization)

Step 1: Set up the Problem in the Standard Form

Step 2: Determine Slack Variables
Step 3: Construct the Simplex Tableau
Step 4: Check Optimality
Step 5: Identify Pivot Variable
Step 6: Create the New Simplex Tableau
Step 7: Identify Optimal Values

B. Simplex Solution Procedures

The procedure to solve a linear programming model using the Simplex method the following
steps are necessary:

Step 1: Set up the Problem in the Standard Form – Standard form is the baseline format
for all linear programs before solving for the optimal solution and has three
requirements:
1) must be a maximization problem,
2) all linear constraints must be in a less-than-or-equal-to inequality,
3) all variables are non-negative.

GUIDED EXAMPLE 5

From the solution of Guided Example 1, solve the following linear programming problem by
the graphical method.

Maximize profit Z  0.60 X  0.90Y

Subject to constraints: 100 X  150Y  10,000 (expenditures needed to plant)

500X  300Y  6,000 (average yields per acre)
X,Y  0

Step 2: Determine Slack Variables – Slack variables are additional variables that are
introduced into the linear constraints of a linear program to transform them from
inequality constraints to equality constraints. This is done by adding one slack
variable for each inequality. Slack variables are needed in the constraints to transform
them into solvable equalities with one definite answer.

27
Maximize profit Z  0.60 X  0.90Y  0S1  0S2

28
 Subject to : 100 X  150Y  1S1  0S2  10,000 (since 0S2  0 and 1S1  S1 )
500 X  300Y  0S1  1S2  6,000 (since 0S1  0 and 1S2  S2 )
X,Y,S1,S2  0

Step 3: Construct the Simplex Tableau - A Simplex tableau is used to perform row
operations on the linear programming model as well as to check a solution for
optimality. The tableau consists of the coefficient corresponding to the linear constraint
variables and the coefficients of the objective function.

Ci Solution mix 0.60 0.90 0 0 Quantity

(basis) X Y S1 S2
0 S1 100 150 1 0 10,000
0 S2 500 300 0 1 6,000
Zj 0 0 0 0 0

Ci  Z j 0.60 0.90 0 0

Step 4: Check Optimality

Follow these 5 steps until an optimal solution has been reached:

1. Determine which variable to enter the solution mix. Variable entering the solution has the
largest positive Ci  Z j . This is the pivot column.
2. Determine the solution mix variable to be replaced. Variable leaving the solution mix is
determined by the ratio. The following are formula for ratio :

Ratio = Quantity [Right Hand Side (RHS)] / (corresponding number in pivot column)

The row with the smallest (nonnegative) ratio will be the pivot row (will be replace in the
next tableau). The intersection of the pivot row and pivot column is referred to as the pivot
number.

R
Q
=
P C
Pivot
Column Pivot
Pivot Row Number =

Ci Solution mix 0.60 0.90 0 0 Quantity Ratio

(basis) X Y S1 S2
0 S1 100 150 1 0 10,000 66.67
0 S2 500 300 0 1 6,000 20
Zj 0 0 0 0 0

Ci  Z j 0.60 0.90 0 0

29
3. Next, setup a new basic simplex tableau. Then, in the new table calculate the new values
for the pivot row. To do this, divide every number in the pivot row by the pivot number. This
row will become the reference row to compute the new values for each remaining row.

Number in the pivot column

New Value Pivot Row 
Pivot number

4. Calculate the new values for the other row.

New row = (old row) – (corresponding number in the pivot column)(reference row)

*Reference row refer to the new value in pivot row that you have calculate in Step 3.

5. Lastly, calculate the value of Z j and Ci  Zj this new tableau. If there are any Ci  Zj
number greater than 0, return to step 1. If there are no Ci  Zj numbers that are greater
than 0, an optimal solution has been reached.

Z value is referring to value at Step 2:

Step 3: New Value for Max profit

Pivot Row (reference row)
Note: Substitute value of each solution
Step 4: New mix for each column to get the
Row value of Z for each column.

Ci Solution mix 0.60 0.90 0 0 Quantity

(basis) X Y S1 S2
0.60 S1 -150 0 1 -1/2 7,000

0.90 Y 5/3 1 0 1/300 20

Zj 0.90 0 0 3/1000 18

Ci  Z j 0.90 0 0 3/1000

Step 5: This is an optimum table since all values in Ci  Zj row contains 0 and positive value.
Therefore, the company needs to plant 20 guavas to gain a maximum profit of RM18.

30
C. Solving Maximization Problems only

Solution to the Primal LP Problem in the Final Simplex tableau

Solution to the primal problem can be obtained by looking at the variable of the final simplex
tableau in the solution mix and the corresponding values in the RHS.

Sensitivity Analysis (Shadow Price) with the Simplex Tableau

The Shadow price (solution to the dual) provides an important piece of economic information
regarding which resources can be classified as a scarce resource (unused resource = 0). It
shows the maximum amount a firm is willing to pay to make one unit of additional scarce
resource available.
If the firm is planning to increase any resources, in order to increase the production rate, the
scarce resources (the shadow price  0 or unused resources  0 , refer to the solution) will
be considered. The higher the shadow price, the more valuable the resource is.

This concept also shows how the optimal solsution and the value of its objective function
change, given changes in various inputs to the problem.

GUIDED EXAMPLE 6

Given below is a linear programming model for two different types of can food sold in a
supermarket.

Maximize Z  2X  3Y (profit in Ringgit Malaysia)

Subject to: X  3Y  180 (resource 1)

2X  3Y  360 (resource 2)
X  Y  100 (resource 3)
X,Y  0

a) Set up the initial simplex tableau for the above problem.

b) The incomplete final simplex tableau for the above problem is shown below.
i. Complete the below simplex tableau.

Ci Solution mix 2 3 0 0 0 Quantity

(basis) X Y S1 S2 S3
Y 0 1 1 0 1 40
2 2
S2 0 0 1 1 3 120
2 2
X 1 0 1 0 3 60
2 2
Zj
Ci  Z j

ii. Determine the optimal solution and the total profit.

iii. What are the shadow prices of all the resources?
iv. Is it worthwhile to purchase additional units of resource 3 at a cost of RM2.50
per unit? Why?

31
Solution:
a)
Ci Solution mix 2 3 0 0 0 Quantity
(basis) X Y S1 S2 S3
0 S1 1 3 1 0 0 180
0 S2 2 3 0 1 0 360
0 S3 1 1 0 0 1 100
Zj 0 0 0 0 0 0
Ci  Z j 2 3 0 0 0

b) i.
Ci Solution mix 2 3 0 0 0 Quantity
(basis) X Y S1 S2 S3
3 Y 0 1 1 0 1 40
2 2
0 S2 0 0 1 1 3 120
2 2
2 X 1 0 1 0 3 60
2 2
Zj 2 3 1 0 3 240
2 2
Ci  Z j 0 0 1 0 3
2 2

ii. Optimal solution: Y  40 , S2  120 , X  60 , S1  S3  0

Total profit  RM240

iii. Shadow prices: Resource 1  RM0.50

Resource 2  0
Resource 3  RM1.50

iv. It is not worthwhile to purchase since the additional cost of resource 3 is RM2.50
more than the shadow price resource 3(RM1.50).

GUIDED EXAMPLE 7

The linear programming model for the monthly production of x, y and z is given below:

Maximize profit Z  45x  60y  10z

Subject to
x  3y  2z  1800 (Item A)
4x  2y  2z  2000 (Item B)
x, y, z  0

a) Complete the optimal simplex tableau for the above problem as shown below:

32
 Ci Solution mix 45 60 10 0 0 Quantity
(basis) x y z S1 S2
y 3 2 1
0 1 5 5 10 520
1  1 3
x 1 0 5 5 10 240
Zj
Ci  Z j

b) Determine the optimal solution and the maximum profit

c) Which items are/is fully utilized?
d) If one unit of item B is added, how much will the profit increase?

Solution:
a)
Ci Solution mix 45 60 10 0 0 Quantity
(basis) x y z S1 S2
60 y 0 1
3 2 1 520
5 5 10
1 1 3
45 x 1 0 5 5 10 240
Zj 45 60 45 15
15 42,000
2
Ci  Z j 15
0 0 -45 -15 2

b) x  520 y  240 s1  0, s2  0
Maximum profit = RM42,000
c) S1 and S2 are fully utilized
d) If one unit of item B is added, profit will increase by RM7.50

HANDS-ON EXAMPLE 7

Objective function maximize the profit of Z  5x1  3x2  4x3

Subject to:
3x1  12x2  6x3  600 (machine hour )
6x1  6x2  3x3  1200 (labour hr )
6x1  9x2  9x3  900 (components)
Where x1 , x2 and x3 represent units of products A, B and C.
a) Convert these constraints and the objective function to the proper form for the use in the
initial simplex tableau.
b) Set up the initial simplex tableau for this problem.
c) Below is a simplex tableau for the above problem.

33
 Ci Solution mix 5 3 4 0 0 0 Quantity
(basis) x1 x2 x3 S1 S2 S3
0 S1 0 15/2 3/2 1 0 -1/2 150
0 S2 0 -3 -6 0 1 -1 300
5 x1 1 3/2 3/2 0 0 1/6 150
Zj
Ci  Z j

i. Complete the values for Z j and C j  Z j rows.

ii. Is this the final tableau for the problem? Explain your answer.
iii. State the optimal solution.
iv. Which resource(s) is not fully utilized? If so, how much spare capacity is left?
v. Is it worth to spend RM1.00 to buy one unit of a component? Explain.
vi. If one unit of components is added, how much will the profit increase?

Solution:
3x1  12x2  6x3  s1  600
Ans: 6x1  6x2  3x3  s2  1200 , 0 -9/2 -7/2 0 0 -5/6
6x1  9x2  9x3  s3  900
Yes, because all the entries for C j  Z j row are  0,
Pr ofit  RM750, x1  150, x2  x3  0,s1  150,s2  300,s3  0 ,
Resources that not fully utilize are machine hours of 150 hours and
labour hours of 300 hours. ( S1  150 and S2  300 ), increased by RM0.83

b) Initial simplex tableau:

Ci Solution Quantity
mix x1 x2 x3 S1 S2 S3
(basis)
S1
S2
S3
Zj
Ci  Z j

34
 Ci Solution mix 5 3 4 0 0 0 Quantity
(basis) x1 x2 x3 S1 S2 S3
0 S1 0 15/2 3/2 1 0 -1/2 150
0 S2 0 -3 -6 0 1 -1 300
5 S3 1 3/2 3/2 0 0 1/6 150
Zj
Ci  Z j

D. Interpretation of Simplex Solution

Slack Variable
Variables added to less than or equal to constraint in order to create an equality to generate
the initial solution. It represents a quantity of unused resource.

Surplus Variable
The difference between the total value of the true (decision) variables and the
number (usually, total resource available) on the right-hand side of the equation. Thus, a
surplus variable will always have a negative value.

Basic Solution
The feasible corner-point solutions to an LP are basic feasible solutions. The solution can be
easily read from the simplex tableau constructed. A basic solution is also called as the basis.

Basic Variable
Variables in the basis are the variables that are not set to zero and have non-zero values.

35
Non-Basic Variable
Variables that are set to zero when defining the corner points (solution) are known as non-
basic variables or variables that do not appear in the basis.

Artificial Variable
Variables to be include in each equality constraint to generate the initial solution. It has no
physical meaning and will be eliminated in the process (before the final tableau)

E. Special cases in LP (simplex method)

Infeasibility (in minimization problems only):

A situation where there is no solution to the problem. This happens when LP model is
wrongly/improperly formulated.

Unbounded Solution:
Unboundedness describes linear programs that do not have finite solution and occurs in
maximization problems. A variable can be indefinitely large without violating a constraint. The
condition of unboundedness will be discovered before reaching the final tableau that is when
trying to decide which variable to remove from the solution mix. The entire ratios turn out to
be negative or undefined.

Degeneracy:
Degeneracy occurs when a problem contains a redundant constraint that is one or more of
the constraints in the formulation make another unnecessary. Degeneracy arises when there
is a tie for the smallest ratio. It can lead the situation called cycling that is the algorithm puts a
new variable in, then takes it out in the next tableau, puts it back in, and so on. One simple
way of dealing with the issue is to select either row arbitrarily. If we are unlucky and cycling
does occur, we simply go back and select the other row.

More Than One Optimal Solution:

In this case there are a number of potential solutions, all yielding the optimum value for the
objective function. Multiple solutions are spotted by looking at the final tableau. In the Cj – Zj
row, at least one non-basis variable has zero coefficients. This means that there is an
alternative solution with the same amount of profit. The column with zero coefficient will be the
new pivot column. After this, continue as before to obtain the alternative solution.

GUIDED EXAMPLE 8

The optimal simplex tableau for the linear programming problem is given in the following table:

Ci Solution mix 6 9 6 0 0 0 Quantity

(basis) X Y Z S1 S2 S3
6 X 1 0 -0.5 1 -0.5 0 40
9 Y 0 1 1 -0.5 0.5 0 60
0 S3 0 0 -3 0.5 -1.5 1 40
Zj
Ci  Z j

a) Complete the above simplex tableau.

b) State the optimal solution including the total profit.
c) Is there any alternative solution? If yes, find it.

36
Solution:

Ci Solution 6 9 6 0 0 0 Quantity
mix X Y Z S1 S2 S3
(basis)
6 X 1 0 -0.5 1 -0.5 0 40
9 Y 0 1 1 -0.5 0.5 0 60
0 S3 0 0 -3 0.5 -1.5 1 40
Zj 6 9 6 1.5 1.5 0 780
Ci  Z j 0 0 0 -1.5 -1.5 0

Optimal solution: X=40, Y=60, S3=40, Total Profit=780

Alternative solution: X=70, Z=60, S3=220, Total Profit=780

HANDS-ON EXAMPLE 8

Given the following final simplex tableau of a manufacturing firm:

Cj
Solution 150 130 250 0 0 0 0 0 Quantity
Mix X1 X2 X3 S1 S2 S3 S4 S5
0 S3 0 0 0 0.25 0.25 1 0 -0.5 56.25
150 X1 1 0 0 0 1 0 0 0 250
250 X3 0 0 1 -0.25 -0.25 0 0 0.5 143.75
0 S4 0 0 0 -0.25 -0.25 0 1 -0.5 243.75
130 X2 0 1 0 1 0 0 0 0 375
Zj 150 130 250 67.5 87.5 0 0 125 P
Cj-Zj 0 0 0 -67.5 -87.5 0 0 -125

a) What is the quantity of each product (X1, X2 and X3) that will be produced?
b) Determine the value of P.
c) Interpret the shadow price for each resource.
d) State the resource(s) that are not fully utilized and state how many units of each resource
have not been used.
e) Is there any alternative solution exists? Why?

Solution:
Ans: x1  250, x2  375, x3  143.75,P  RM122,187.50 ,
R1  RM67.50,R2  87.50,R3  R4  0,R5  125

37
 HANDS-ON EXAMPLE 8

A company produces two products ( x1 and x2 ) daily. Each product requires two resources
which are resource 1 ( S1 ) and resource 2 ( S2 ). The profits of producing each unit of x1 and
x2 are RM50 and RM80, respectively. The final simplex tableau for this linear programming
problem is shown in the following table:

Ci Solution mix Quantity

(basis) x1 x2 S1 S2
x2 0 1 0.375 -0.25 5
x1 1 0 -0.25 0.50 10
Zj
Ci  Z j

a) Complete the above simplex tableau.

b) Specify the optimal daily production levels of the two products. What is the total profit?
c) Is there any alternative solution? Explain.

38
Solution:
Ans: Ci  Zj  0,0,17.5,5, x1  10, x2  5, profit  900
No alternative solution since Ci  Zj is 0 for x1 and x2 .
Also x1 and x2 are in the solution

Ci Solution mix Quantity

(basis) x1 x2 S1 S2
x2 0 1 0.375 -0.25 5
x1 1 0 -0.25 0.50 10
Zj
Ci  Z j

F. Sensitivity Analysis – Right Hand Side Values only

39
G. Dual in LP (Interpreting the final tableau)

• In linear programming, duality implies that each linear programming problem can be
analysed in two different ways but would have equivalent solutions. Any LP problem
(either maximization and minimization) can be stated in another equivalent form based on
the same data.
• Second way of stating the same problem
• LP primal involves maximizing a profit function subject to less than or equal (≤) to
resource constraints
• LP dual involves minimizing total opportunity cost subject to greater than or equal (≥) to
product profit constraints

Solution to the Dual

As we know, the primal and dual lead to the same solution even though they are formulated
differently. In the final simplex tableau of a primal problem, the absolute values of the numbers
in the Cj – Zj row under the slack variables represent the solution to the dual problem. These
numbers are termed shadow prices. Thus the solution to the dual present the marginal profits
of each additional unit of resource that is the additional profit obtained for each additional
unit resource.

Steps to form a dual model from the primal model

Relationship between Primal and Dual

40
 GUIDED EXAMPLE 9

From the question of Guided Example 6, solve the following linear programming problem
below by write the dual for the linear programming model.

Maximize Z  2X  3Y (profit in Ringgit Malaysia)

Subject to: X  3Y  180 (resource 1)

2X  3Y  360 (resource 2)
X Y  100 (resource 3)
X,Y 0

Solution:

Minimize Cost C  180U  360V  100W

Subject to: U  2V  W  2
3U  3V  W  3
U,V,W  0

GUIDED EXAMPLE 9

From the solution of Guided Example 7,

Maximize profit Z  45x  60y  10z

Subject to
x  3y  2z  1800 (Item A)
4x  2y  2z  2000 (Item B)
x, y, z  0

Ci Solution mix 45 60 10 0 0 Quantity

(basis) x y z S1 S2
Zj 15 42,000
45 60 35 15 2
Ci  Z j 15
0 0 -35 -15 2

a) Write the dual formulation for the above LP problem.

b) State the optimal solutions for dual.

41
Solution:
a) Minimize cost C  1800U  2000V
subject to
U  4V  45
3U  2V  60
2U  2V  10
U ,V  0

b) U  15,V  7.5 and minimum cost RM42,000

HANDS-ON EXAMPLE 9

Citarasa Company produces three food paste products Rendang, Kuah Kacang and Masak
Lemak paste. Three resources required to produce the food paste product are cooking
machine time, inspection time, and paste ingredients. The manager has formulated the
problem as below.

X1  number of units of Rendang paste

X 2  number of units of Kuah Kacang paste
X 3  number of units of Masak Lemak paste

Maximise, Z  18 X1  15 X 2  10 X 3

Subject to X1  X 2  3 X 3  80 (resource cooking machine time)

2X 1  X 2  X 3  40 (resource inspection time)
2X 1  2X 2  X 3  100 (resource paste ingredients)
X 1, X 2 , X 3  0

a) State the dual to this problem.

b) The following is the final simplex tableau of the primal. Complete the tableau.

Ci Solution 18 15 10 0 0 0 Quantity
Mix X1 X2 X3 S1 S2 S3
S1 1 0 2 1 1 0
X2 2 1 1 0 1 0
S3 2 0 1 0 2 1
Zi
Cj  Z j

c) Is the solution obtained optimal? Why?

d) State the optimal solution and the total profit.
e) What is the shadow price of resource inspection time?

42
Solution:
Ans: Minimise C  80U  40V  100W , U  2V  2W  18 ,
U  V  2W  15 , 3U  V  W  10 , Ci  Z j  12,0,5,0,15,0 ,
The solution is optimal as all the values in C j  Z j row are non-positive.
X 2  40,S1  40,S3  20 , X1  X 3 S2  0 , total profit: RM600, RM15.00

Ci Solution 18 15 10 0 0 0 Quantity
Mix X1 X2 X3 S1 S2 S3
S1 1 0 2 1 1 0
X2 2 1 1 0 1 0
S3 2 0 1 0 2 1
Zi
Cj  Z j

43
 HANDS-ON EXAMPLE 10

Write the dual formulation for the below linear programming problem.

Maximize profit Z  x  6y  8z
Subject to
x  2y  1200
2y  z  1800
4x  z  3600
x, y, z  0

Solution: Ans: Minimise C  1200U  1800V  3600W , U  4W  1,

2U  2V  6 , V  W  8 ,

44
 CHAPTER 6:
TRANSPORTATION AND
ASSIGNMENT MODEL

Learning Objective (LO):

At the end of this chapter, students will be able to:

 Calculate problem involving assignment model
 Differentiate between balanced and unbalanced problem
 Differentiate between maximum and minimum problem
 Develop an initial solution for transportation model
 Solve problem involving transportation model

45
4.1 Assignment Model
 Assignment model is problem based on assigning task for object/people.
 Using this method, the task is assign in order to maximize profit or minimize the cost.
 Method that will be used to solve assignment model is Hungarian Method. This method
can ONLY be used for Minimization Problem.
 If we encounter a maximization problem, the problem should be convert into minimization
problem before we can solve it using Hungarian Method.

Approach Of The Assignment Model

 Develop a table based on the question.
 The row represents the objects or people we wish to assign
 The column represents the tasks or things we want to assign.
 The number in the table are the costs associated with each particular assignment

A. The Hungarian Method (Flood’s technique) - Minimization Problem

 This method are used to find the optimal solution without having to make a direct
comparison of every option
 The first step in this method. We will be doing matrix reduction.
Matrix reduction: by subtracting and adding appropriate number in the cost table or matrix,
we can reduce the problem to a matrix of opportunity costs
 After that we will proceed to finding an opportunity costs.
Opportunity costs: show the relative penalties associated with assigning any person to a
project as opposed to making the best or least cost assignment.

Steps in Hungarian Method:

Before using this step please check for balanced of this problem. Balanced problem means
Number Of Row = Number of Column

46
An Optimal solution can be found by assigning to:
a) Any zero which is unique to BOTH column and a row
b) Any zero which is unique to a column OR a row
c) Assignment which is not in (a) and (b)

GUIDED EXAMPLE 1

The manager of Motivation Training Agency has to assign four different facilitators to handle
four different programs. These facilitators would be paid at different rates per hour according
to their experience. The costs (RM) of assigning the facilitators are shown in the following
table. Advice the owner how to make the optimal assignment and calculate the minimum total
cost.

Facilitator Program
I II III IV
Abu 210 90 180 160
Bakar 100 70 130 200
Ishak 175 105 140 170
Musa 80 65 105 120

Solution:

Row = Column.
Therefore, it is balanced problem, we can proceed to use Hungarian Method.

Step 1:
Find the smallest number in a row

Facilitator Program Smallest

I II III IV No in a row
Abu 210 90 180 160 90
Bakar 100 70 130 200 70
Ishak 175 105 140 170 105
Musa 80 65 105 120 65

Row Substraction:

Substracting every number in a row by the smallest number such as 210  90  120...
120  65  55

Facilitator Program
I II III IV
Abu 120 0 90 70
Bakar 30 0 60 140
Ishak 70 0 35 65
Musa 15 0 40 55

47
Step 2:
Find the smallest number in a column from the row substraction

Facilitator Program
I II III IV
Abu 120 0 90 70
Bakar 30 0 60 140
Ishak 70 0 35 65
Musa 15 0 40 55
Smallest No in 15 0 35 55
Column

Column Substraction:

Substracting every number in a column by the smallest number such as 120  15  105...
55  55  0

Facilitator Program
I II III IV
Abu 105 0 55 15
Bakar 15 0 25 85
Ishak 55 0 0 10
Musa 0 0 5 0

Step 3:
Testing for optimal assignment

Facilitator Program
I II III IV
Abu 105 0 55 15
Bakar 15 0 25 85
Ishak 55 0 0 10
Musa 0 0 5 0

The minimum line that cover all zero are 3 lines.

Row = Column = 4 ≠ Minimum line (3)
Therefore, this is not optimal yet. We have to proceed to Step 4.

Step 4:
Revise the present table

Facilitator Program
I II III IV
Abu 105 0 55 15
Bakar 15 0 25 85
Ishak 55 0 0 10
Musa 0 0 5 0

Smallest Number not Covered in line = 15

48
New Revised Table :
 Smallest Number NOT covered in line – other number not covered in line
 Smallest Number NOT covered in line + intersection
 The rest just copy from the previous table

Facilitator Program
I II III IV
Abu 90 0 40 0
Bakar 75 0 10 70
Ishak 55 0 0 10
Musa 0 0 5 0

Return to Step 3:
Testing for optimal assignment

Facilitator Program
I II III IV
Abu 90 0 40 0
Bakar 75 0 10 70
Ishak 55 0 0 10
Musa 0 0 5 0

The minimum line that cover all zero are 4 lines.

Row = Column = 4 = Minimum line (4)
Therefore, this is an optimal solution.

Decision making:

Choose the program that contain cost 0 for each of the fasilitator.
Abu - Program II and Program IV
Bakar – Program II
Ishak – Program II and Program III
Musa – Program I, II and IV

Start with the fasilitator that have a less program.

Bakar – Program II
Since Bakar has been chosen for Program I delete all Program I in other fasilitator.

Abu - Program II and Program IV

Ishak – Program II and Program III
Musa – Program I, II and IV

Therefore, we will chose

Abu – Program IV
Ishak – Program III
Now, delete program III and IV from the remaining fasilitator
Musa – I

Therefore the answer is :

Abu - Program IV
Bakar – Program II
Ishak – Program III
Musa – Program I

49
Now, calculate the minimum cost for the company :
*Refer to the original table to obtain the cost for each fasilitator based on the program.
Abu - Program IV - RM 160
Bakar – Program II - RM 70
Ishak – Program III - RM 140
Musa – Program I - RM 80
Total Minimum Cost = RM 450

B. Unbalanced Assignment Problem

 Number of rows and columns are not equal
 More rows than columns  add dummy columns
 More columns than rows  add dummy rows
 Cost or time is equal to zero

GUIDED EXAMPLE 2

Ken is the owner of the Ken Electric Sdn Bhd. He receives 3 items to be repair that is blender,
microwave and toaster. Currently, he have 4 repair person that is Ahmad, Ben, Chan and
Khalid. This four repair persons have different talents and abilities. The cost of each person
repairing the item are differ based on their skill and speed. Advice Ken on how to make the
optimal assignment and calculate the minimum total cost.

Project (cost RM)

Worker 1 2 3
Ahmad 11 14 6
Ben 8 10 11
Chan 9 12 7
Khalid 10 13 8

Solution:

Row ≠ Column
Therefore add Column in order to balance the table
Cost for the dummy = 0

Worker Project
1 2 3 Dummy
Ahmad 11 14 6 0
Ben 8 10 11 0
Chan 9 12 7 0
Khalid 10 13 8 0

Row Subtraction
Worker Project
1 2 3 Dummy
Ahmad 3 4 0 0
Ben 0 0 5 0
Chan 1 2 1 0
Khalid 2 3 2 0

50
Column Subtraction :
Worker Project
1 2 3 Dummy
Ahmad 3 4 0 1
Ben 0 0 5 1
Chan 0 1 0 0
Khalid 1 2 1 0

Optimal solution:
Worker Project Cost
Ahmad 3 6
Ben 2 10
Chan 1 9
Khalid Dummy 0
Total RM25

C. Maximization assignment problem

 In order to solve maximum problem, convert the original table (maximize) to opportunity
cost (minimize) table by subtracting every element in the table from the largest number
in the table.
 Then, solve it using Hungarian Method.

GUIDED EXAMPLE 3

Four ship name Garuda, Angkasa, Tiong and Merbok have to ship goods to four different
sector that is A, B, C and D. Table below show a profit (RM ‘0000) the company will gain if he
ship the goods in time. Arrange the transportation of the goods in order for the company to
gain maximum profit and calculate the profit the company will gain.

Ship Sector
A B C D
Garuda 20 60 50 55
Angkasa 60 30 80 75
Tiong 80 100 90 80
Merbok 65 80 75 70

Solutin:

Convert the maximum problem into minimum problem.

Largest Number in the table (100) – each element in the table.

Ship Sector
A B C D
Garuda 80 40 50 45
Angkasa 40 70 80 25
Tiong 20 0 10 20
Merbok 35 20 25 30

51
Solve it using Hungarian Method

Row Substraction:
Ship Sector
A B C D
Garuda 40 0 10 5
Angkasa 15 45 55 0
Tiong 20 0 10 20
Merbok 15 0 5 10

Column Substraction:
Ship Sector
A B C D
Garuda 25 0 5 5
Angkasa 0 45 50 0
Tiong 5 0 5 20
Merbok 0 0 0 10

Test Optimality
Ship Sector
A B C D
Garuda 25 0 5 5
Angkasa 0 45 50 0
Tiong 5 0 5 20
Merbok 0 0 0 10

Optimal Solution
Ship Sector Answer Profit
Sector (RM ‘0000)
Garuda B, D D 55
Angkasa A, D A 60
Tiong B B 100
Merbok A, B, C C 75
Maximum Profit 290

HANDS-ON EXAMPLE 1

Tech Enterprise has four technicians to be assigned to repair four machines for their clients.
The four technician are paid hourly. Below are the wages of the four technician, the wages
are calculated based on their experience, skill and speed.

Technician Machine
A B C D
Ganesh 24 16 28 22
Chandran 20 18 20 16
Maxden 28 16 14 22
Danial 12 20 16 18

Find the optimal assignment that will minimize the total cost for the company to pay their
technician. What is the total hourly profit?

52
Solution: Ans: Ganesh - C, Chandran - B, Maxden - A, Danial - D,
Total Min Cost = RM 92)

53
 HANDS-ON EXAMPLE 2

Frenz Marketing Research has four project leaders available for assignment to three clients.
Find the assignment of project leaders to clients that will minimize the total time to complete
all projects. The estimated project completion times in days are as follows:

Project leader Client 1 Client 2 Client 3

Azmi 10 15 9
Basir 9 18 5
Chen 9 14 3
Danny 8 16 6

Solution: Ans: Azmi – Clien 2, Basir-Dummy,Chen – Client

3,Danny- Client 1, Total Min Cost = 26

54
 HANDS-ON EXAMPLE 3

The Malaysian Royal Navy wishes to assign four warships to patrol four sectors of the
Malaysian Costal Areas on the lookout for illegal fishing activities and pirates. The relative
efficiencies of these four warships set by the commander are shown in the following table.
Find the optimal assignment to maximize efficiencies and calculate the total efficiencies.

Ship Sector
A B C D
1 90 65 95 40
2 70 60 80 75
3 85 40 80 60
4 55 80 65 55

Solution: Ans: Ship 1- Sector C, Ship 2 – Sector D, Ship 3 – Sector A,

Ship 4 – Sector B, Total max Cost =335

55
4.2 Transportation Model
 Transportation model are used to find the best way to distribution goods from several
points of supply (sources) to a number of points of demand (destinations).
 A product is transported from a number of sources to a number of destinations at the
minimum possible cost (to minimize the cost).
 Each source is able to supply a fixed number of units of the product and each destination
has a fixed demand for the product
 All constraints are equalities in a balanced transportation model where supply equals to
demand.
 Constraints contain inequalities in unbalanced models where supply does not equal to
demand
 Solution steps:
 Define problem
 Set up transportation tableau
 Develop initial solution using least cost method
 Find optimal solution using MODI method (modified distribution)
 There are 2 step involve in transportation model. The first one is developing Initial
solution and the second one is finding an optimal solution.
 Before we start with the calculation, please make sure to always check whether the
problem is balanced or unbalanced problem.
 Balanced problem meaning that number of supply = number of demand

Developing an Initial Solution

 There are 3 method to find the initial solution
a) Method I: Least Cost Method-LCM
b) Method II: Northwest Corner Rule
c) Method III: Vogel’s Approximation Method (VAM) (For Minimization Problem)
 You can choose any suitable method for developing initial solution.

A. Method I: Least Cost Method-LCM (For Minimization Problem)

1. Check if the problem is balance (Total Supply SS =Total Demand DD)

If the problem is unbalance, then either one of the following cases will happen:
a. Total SS > Total DD, add 1 dummy destination(location)
b. Total SS < Total DD, add 1 dummy source
2. Set up the transportation table
3. Find the 1st solution using LCM. This method starts by assigning goods to
the cell with the least cost (excluding zero cost assigned to each
dummy)

56
 GUIDED EXAMPLE 4

From/To A B C D Supply
X 28 18 32 36 300

Y 22 16 14 32 420

Z 32 24 20 36 600

Demand 280 200 360 480 1320

Solution:

1. Choose the lowest cost available. That is 14. 14 is in row Y and column C. Now choose
that lowest value between number of supply row Y = 420 and number of demand in column
C = 360.
Therefore we will be choose 360. This means that the maximum value you can filled in is
360. Now since we already fill up the box of the lowest cost (14) with value of 360, all of
the column C have been filled. Means you cannot filled in other number. So write X to
indicidate that this column no longer available.

From/To A B C D Supply
X 28 18 32 36 300
X
Y 22 16 14 32 420
360
Z 32 24 20 36 600
X
Demand 280 200 360 480 1320

2. Choose the lowest cost from available box. The lowest cost now is 16. 16 located at row
Y with supply 420 and column B with demand 200. Now, choose the lowest number
between 420 and 200. 200 will be chosen. This means that the maximum value you can
filled in is 200. Before you fill the box with cost 16 with 200, take a look at the row. The
total supply for row Y is 420 but we already fill in the number of supply for the column C.
Therefore the calculation will be 420 – 360 = 60. Now, all row Y have been fulfilled.

From/To A B C D Supply
X 28 18 32 36 300
140 X
Y 22 16 14 32 420
X 60 360 X
Z 32 24 20 36 600
X X
Demand 280 200 360 480 1320

57
3. Next choose the lowest cost that left. We will choose 18. 18 is located in column B with
demand 200 and row X with supply 300. Now, choose the lowest number between 200
and 300. 200 will be chosen. This means that the maximum value you can filled in is 200.
Before you fill the box with cost 18 with 200, take a look at the column B. The total supply
for column B is 200 but we already fill in the number of supply for the row Y with 60.
Therefore the calculation will be 200 – 60 = 140. Now, all column B have been fulfilled.

From/To A B C D Supply
X 28 18 32 36 300
140 X
Y 22 16 14 32 420
X 60 360 X
Z 32 24 20 36 600
X X
Demand 280 200 360 480 1320

4. Next choose the lowest cost that left. We will choose 28. 28 is located in column A with
demand 280 and row X with supply 300. Now, choose the lowest number between 280
and 300. 280 will be chosen. This means that the maximum value you can filled in is 280.
Before you fill the box with cost 28 with 280, take a look at the row X. The total demand
for column is 280 but we already fill in the number demand for the column B with 140.
Therefore the calculation will be 300 – 140 = 160. We can choose the answer. Now, all
row X have been fulfilled.

From/To A B C D Supply
X 28 18 32 36 300
160 140 X X
Y 22 16 14 32 420
X 60 360 X
Z 32 24 20 36 600
X X
Demand 280 200 360 480 1320

5. Next choose the lowest cost that left. We will choose 32. 32 is located in column A with
demand 280 and row Z with supply 600. Now, choose the lowest number between 280
and 600. 280 will be chosen. This means that the maximum value you can filled in is 280.
Before you fill the box with cost 32 with 280, take a look at the row X. The total demand
for column A is 280 but we already fill in the number of demand for the column A with 280.
Therefore the calculation will be 280 – 160 = 120.

From/To A B C D Supply
X 28 18 32 36 300
160 140 X X
Y 22 16 14 32 420
X 60 360 X
Z 32 24 20 36 600
120 X X
Demand 280 200 360 480 1320

58
6. Now the only box that have not been filled is cost 36. Fill in the box with the remaining
supply.
600 – 120 = 480

From/To A B C D Supply
X 28 18 32 36 300
160 140 X X
Y 22 16 14 32 420
X 60 360 X
Z 32 24 20 36 600
120 X X 480
Demand 280 200 360 480 1320

Total Transportation = (160 x RM 28) + (140 x RM 18) + (60 x RM 16) + (360 x RM 14) +
(120 x RM 32) + (480 X RM 36)
= RM34,120.

B. Method II: Northwest Corner Rule

i. The starting point is the northwest corner of the table with the maximum quantity
allocated to the cell.
ii. If the supply at each row is allocated fully, allocation is then moved to the next row. This
begins with the first row.
iii. Similarly, if the demand of each column is allocated fully, the demand is then moved to
the next column. Again, beginning with the first column
iv. Finally all the supplies and demands are ensured to be satisfied.

Solution:

From/To A B C D Supply
X 28 18 32 36 300
1
280 20
2
x x
Y 22 16 14 32 420
x 180
3 4
240 x
Z 32 24 20 36 600
x x 5
120 6
480
Demand 280 200 360 480 1320

Total Transportation = (280 x RM 28) + (20 x RM 18) + (180 x RM 16) + (240 x RM 14) +
(120 x RM 20) + (480 x RM 36)
= RM34,120.

59
C. Method III: Vogel’s Approximation Method (VAM) (For Minimization Problem)

i. Compute opportunity cost (difference between two lowest cost cell of the row/column)
for each row/column.
ii. Select the highest opportunity cost and assign maximum units to the lowest cost cell of
the row/column. Eliminate the row/column and recalculate the opportunity cost.
iii. Repeat step 1 and 2 until the table is fully filled.

Solution:

From/To A B C D Supply
X 28 18 32 36 300 1
10 4 8
X 200
1
x 100
5 36 536
Y 22 16 14 32 420 2 8 310
3
280 x x 6
140 32 32
Z 32 24 20 36 600 4 212 4
x x 360 2
2404 4
36 x
Demand 280 200 360 480 1320
66 6 xx 2xx x x 66xx x 44 4 4 4
** 1200, number 1 represent the first quantity been filled.

Total Transportation = (200 x RM 18) + (100 x RM 36) + (280 x RM 22) + (140 x RM 32) +
(360 x RM 20) + (240 x RM 36)
= RM 33 680.

Comparison between the three methods.

Method Transportation Cost
LCM RM 34120
Northwest Corner Rule RM 34120
VAM RM33680 (lowest)

* VAM is the best method to choose.

D. Finding an Optimal Solution: MODI Method

Testing for optimal solution

1. Make sure: No. of occupied cells = no. of rows + no. of column – 1 (to ensure no
degeneracy problem)
Compute the cost value for each row and column, for those squares that are currently used
or occupied by setting
Ri  K j  Cij
where Ri = value assigned to row I, Kj = value assigned to column j, Cij = cost in square ij
2. After all the equation has been written, set R1 = 0.
3. Solve the system of equation for all R and K values.
4. Next, calculate the improvement index using formula for unoccupied/unused cells only:
Improvement Index, Iij  Cij - Ri - K j
Write the improvement index at the bottom left corner of each unused cell.
5. A solution is optimum when all Iij  0 . If at least one Iij  0 , get a better solution.

60
Finding a better solution
1. Find the entering cell - the cell with the largest negative Improvement Index
2. Find the leaving cell.
a. Make a closed path for the entering cell. (this path must be as simple as possible)
• Start and end at the entering cell.
• Use vertical & horizontal line.
• Lines can change only at occupied cells – not all occupied cells have to be in the
closed path.
b. Place (+) & (-) sign alternately at each corner of the path. Start with the (+) sign at the
entering cell.
c. Choose the lowest number of units from cells containing (-) sign, say Q. This is the
leaving cell. (Do the same for maximization problem)
3. Set up a new table.
• Add Q to any cell with (+) sign and subtract Q to any cell with (-) sign.
4. Compute new improvement indices for this new solution to test either the solution is
optimal or not. Repeat all the process until the optimal solution is found.

E. The Unbalanced Transportation Problem

 When demand exceeds supply, a dummy row is added to the tableau

From/To A B C Supply
1 6 8 10
150
2 7 11 11
175
3 4 5 12
275
Dummy 0 0 0
50
Demand 200 100 350 650

 When supply exceeds demand, a dummy column is added to the tableau

From/To A B C Dummy Supply

1 6 8 10 0
150
2 7 11 11 0
175
3 4 5 12 0
375
Demand 200 100 300 100 700

All steps applied in order to obtain optimal solution are same as previous

61
F. Alternative Optimal Solution

Multiple solutions are possible when one or more improvement indices in the optimal solution
stages equals zero. This means that it is possible to assign alternative transportation routes
with the same total transportation cost.

Steps to find the alternative solution are:

i. Choose the cell with Iij  0 as the entering cell.
ii. Use the same procedure as in step 2 and 3 (in finding a better solution)
iii. Write down the new solution

Notes: If there is 1 cell with Iij  0 , the problem has only one alternative solution. If there are
two cells with Iij  0 , the problem has two alternative solutions.

GUIDED EXAMPLE 5

Find an alternative optimal solution

From/To KA=6 KB=7 KC=10 Supply

A B C
R1=0 1 6 8 10
25 +1 125 150
R2=1 2 7 11 11
0 +3 175 175
R3=-2 3 4 5 12
175 100 +4 275
Demand 200 100 300 600

G. Degeneracy

 In a transportation tableau with m rows and n columns, there must be m + n - 1 cell with
allocations; if not, it is degenerate.
 The tableau in the figure does not meet the condition since 3 + 3 -1 = 5 cells and there are
only 4 cells with allocations.

From/To A B C Supply
1 6 8 10
100 50 150
2 7 11 11
250 250
3 4 5 12
200 200
Demand 200 100 300 600

 In a degenerate tableau, all the MODI equations cannot be developed

 To rectify a degenerate tableau, an empty cell must artificially be treated as an
occupied cell and we place a zero (representing a fake shipment).
 Place zero in one of the lowest cost unused cell and then treat the square as if it
were occupied cell. If still unable to solve for R and K, change to the next unused
least cost cell.

62
 From/To A B C Supply
1 6 8 10
100 50 150
2 7 11 11
250 250
3 4 5 12
200 0 200
Demand 200 100 300 600

GUIDED EXAMPLE 6

Zamcomp Sdn. Bhd. produces one type of computer table at three factories located at different
towns. Factory I, II, and III can manufacture 300, 200, and 100 units respectively. These
computer tables are distributed to three major outlets, A, B and C. The demands from the
outlets are 200, 150 and 250 units respectively. The cost of producing one unit of computer
table varies due to different production technologies.

The production and distribution cost per unit (RM) from factories to outlets are as follows.

OUTLETS
FACTORY A B C
I 37 34 36
II 36 32 35
II 42 35 41

a) Set up a transportation tableau and determine the initial solution using LCM.
b) Is the initial solution from (a) optimal? Why?
c) Calculate the total production and distribution cost for the initial solution.
d) Solve the optimal solution, list the solution and calculate the minimum total production
and transportation cost.
e) Explain why there is an alternative optimal solution, find it and fill in the table below.

Solution:

a) Initial Using LCM

 Check for balance/unbalanced problem. Balance problem Demand = Supply

From/To A B C Supply
I 37 34 36
100 X 200 300
II 36 32 35 200
X 150 50
III 42 35 41 100
100 X X
Demand 200 150 250 600

Total Transportation = (100 x RM 37) + (200 x RM 36) + (150 x RM 32) + (50 x RM 35)
(100 x RM 42)
= RM21,650

63
* check number of quantity that have been filled to make sure we did not encounter
degeneracy problem.
Row + Column – 1 = Number of box filled
3 + 3 -1 = 5 √ (not degeneracy problem)

b) Check for Optimality using MODI

From/To A B C Supply
= = =
I 37 34 36
= 100 1 200 300
II 36 32 35 200
��� = −�� 0 150 50
III 42 35 41 100
= 100 -3 0
Demand 200 150 250 600

Ri  K j  Cij (for current box filled)

1 + 1 = 37 Solve for all R and K value and fill in the above tableu

�1 + �3 = 36 ��� = �� , ��� = ����, ��� = ����, ��� = −��, ��� = ��, ��� = ����

2 + 2 = 32

2 + 3 = 35

3 + 1 = 42

Then, solve index improvement for each of unused cell. Iij  Cij  Ri  K j

��� = 34 − 0 − 33 = ��

���� = 36 − (−1) − 37 = ��

����� = 35 − 5 − 33 = −��

����� = 41 − 5 − 36 = ��

Since, one of the value of index improvement is not larger and equal to zero. Therefore it is
not optimal.

c) Total Transportation = (100 x RM 37) + (200 X RM 36) + (150 X RM 32) + (50 X RM 35)
(100 X RM 42)
= RM21,650

64
d) Finding a better solution.

From/To A B C Supply
= = =
I (+) 37 34 (-) 36
= 100 1 200 300
II 36 32 35 200
��� = −�� 0 150 (-) 50 (+)
III (-) 42 35 41 100
= 100 -3 (+) 0
Demand 200 150 250 600

 Choose the most negative value for index improvement to be entering cell. Cell with -3 as
improvement index are chose.
 Make a close path
 Place (+) & (-) sign alternately at each corner of the path. Start with the (+) sign at the
entering cell.
 Choose the lowest number of units from cells containing (-) sign, say Q. This is the leaving
cell. (Quantity = 100)
 Set up a new table.
• Add Q to any cell with (+) sign and subtract Q to any cell with (-) sign.

From/To A B C Supply
I 37 34 36
200 100 300
II 36 32 35 200
50 150
III 42 35 41 100
100
Demand 200 150 250 600

Now check for optimality using MODI

From/To A B C Supply
= = =
I 37 34 36
= 200 1 100 300
II 36 32 35 200
��� = −�� 0 50 150
III 42 35 41 100
= 5 100 5
Demand 200 150 250 600

Ri  Kj  Cij (for current box filled)

1 + 1 = 37 Solve for all R and K value and fill in the above tableu

�1 + �3 = 36 ��� = �� , ��� = ����, ��� = ����, ��� = −��, ��� = ��, ��� = ����

2 + 2 = 32

2 + 3 = 35

3 + 2 = 33

65
Then, solve index improvement for each of unused cell. Iij  Cij Ri K j

��� = 34 − 0 − 33 = ��

���� = 36 − (−1) − 37 = ��

����� = 42 − 0 − 37 = ��

����� = 41 − 0 − 36 = ��

Since, index improvement is larger and equal to zero. Therefore it is optimal.

Answer:
Transport from I to A = 200
Transport from I to C = 100
Transport from II to B = 50
Transport from II to C = 150
Transport from III to B = 100

Total Transportation = (200 x RM 37) + (100 X RM 36) + (50 X RM 32) + (150 X RM 35)
(100 X RM 35)
= RM21,350

e) Yes, it has alternative optimal solution because one of the improvement index is equal to
zero.

The method to find alternative optimal solution is the same as finding an optimal solution.
The different is we choose index improvement = 0 as our entering cell.

From/To A B C Supply
= = =
I (-) 37 34 (+) 36
= 200 1 100 300
II 36 32 (-) 35 200
��� = −�� 0 (+) 50 150
III 42 35 41 100
= 5 100 5
Demand 200 150 250 600

 Make a close path

 Place (+) & (-) sign alternately at each corner of the path. Start with the (+) sign at the
entering cell.
 Choose the lowest number of units from cells containing (-) sign, say Q. This is the leaving
cell. (Quantity = 150)
 Set up a new table.
• Add Q to any cell with (+) sign and subtract Q to any cell with (-) sign.

66
 From/To A B C Supply
I 37 34 36
50 250 300
II 36 32 35 200
150 50
III 42 35 41 100
100
Demand 200 150 250 600

Answer:
Transport from I to A = 50
Transport from I to C = 250
Transport from II to A = 150
Transport from II to B = 50
Transport from III to B = 100

Total Transportation = (50 x RM 37) + (250 x RM 36) + (150 x RM 36) + (50 x RM 32)
(100 x RM 35)
= RM21,350

* the total transportation is the same as (d) but the transportation item from factory to
outlet is different.

Maximization Transportation Problem

 The optimal solution to a maximization problem has been found when all improvement
indices are negative or zero.
 If any index is +ve, the cell with the largest +ve improvement index is selected to be filled
using a MODI method.
 This new solution evaluated and the process continues until there are no +ve improvement
indices.

GUIDED EXAMPLE 7

Refer to Guided-Example 6:
Calculate the total profit if the computer tables are sold at RM320 each of the outlets.

Solution:

Initial Using LCM

 Check for balance/unbalanced problem. Balance problem Demand = Supply

From/To A B C Supply
I 283 286 284
100 X 200 300
II 284 288 285 200
X 150 50
III 278 285 279 100
100 X X
Demand 200 150 250 600

67
Total Transportation = (100 x RM 283) + (200 x RM 284) + (150 x RM 288) + (50 x RM 285)
(100 x RM 278)
= RM170,350

* check number of quantity that have been filled to make sure we did not encounter
degeneracy problem.
Row + Column – 1 = Number of box filled
3 + 3 – 1 = 5 √ (not degeneracy problem)

Check for Optimality using MODI

From/To A B C Supply
= = =
I 283 286 284
= 100 -1 200 300
II 284 288 285 200
= 0 150 50
III 278 285 279 100
��� = −�� 100 3 0
Demand 200 150 250 600

Ri  Kj  Cij (for current box filled)

1 + 1 = 283 Solve for all R and K value and fill in the above tableu

�1 + �3 = 284 ��� = �� , ��� = ������, ��� = ������, ��� = ��, ��� = −��, ��� = ������

2 + 2 = 288

2 + 3 = 285

3 + 1 = 278

Then, solve index improvement for each of unused cell. Iij  Cij  Ri  K j

��� = 286 − 0 − 287 = −��

���� = 284 − 1 − 283 = ��

����� = 285 − (−5) − 287 = ��

����� = 279 − (−5) − 284 = ��

Since, one of the value of index improvement is larger and equal to zero. Therefore it is not
optimal.

68
Finding a better solution.

From/To A B C Supply
I (+) 283 286 (-) 284
100 -1 200 300
II 284 (-) 288 285 200
0 150 50 (+)
III (-) 278 (+) 285 279 100
100 3 0
Demand 200 150 250 600

 Choose the most positive value for index improvement to be entering cell. Cell with 3 as
improvement index are chose.
 Make a close path
 Place (+) & (-) sign alternately at each corner of the path. Start with the (+) sign at the
entering cell.
 Choose the lowest number of units from cells containing (-) sign, say Q. This is the leaving
cell. (Quantity = 100)
 Set up a new table.
 Add Q to any cell with (+) sign and subtract Q to any cell with (-) sign.

From/To A B C Supply
I 283 286 284
200 100 300
II 284 288 285 200
50 150
III 278 285 279 100
100
Demand 200 150 250 600

Now check for optimality using MODI

From/To A B C Supply
= = =
I 283 286 284
= 200 -1 100 300
II 284 288 285 200
= 0 50 150
III 278 285 279 100
��� = −�� -3 100 -3
Demand 200 150 250 600

Ri  Kj  Cij (for current box filled)

1 + 1 = 283 Solve for all R and K value and fill in the above tableu

�1 + �3 = 284 ��� = �� , ��� = ������, ��� = ������, ��� = ��, ��� = −��, ��� = ������

2 + 2 = 288

2 + 3 = 285

3 + 2 = 285

69
Then, solve index improvement for each of unused cell. Iij  Cij Ri K j

��� = 286 − 0 − 287 = −��

���� = 284 − 1 − 283 = ��

����� = 278 − (−2) − 283 = −��

����� = 279 − (−2) − 284 = −��

Since, index improvement is not larger and equal to zero. Therefore it is optimal.

Answer:
Transport from I to A = 200
Transport from I to C = 100
Transport from II to B = 50
Transport from II to C = 150
Transport from III to B = 100

Total Transportation = (200 x RM 283) + (100 X RM 284) + (50 x RM 288) + (150 x RM 285)
(100 x RM 285)
= RM170,650

HANDS-ON EXAMPLE 4

A plywood producer manufactures plywood at three factories to meet the demand of three construction
sites. The factories are located at different locations. The production capacity at each factory
and the demand from each construction site are shown in the following tables:

Factory Capacity (units) Site Demand (units)

X 300 P 200
Y 200 Q 150
Z 100 R 230

The transportation cost (RM) per unit from each factory to each site is shown in the following
table:
Factory Site
P Q R
X 7 4 6
Y 6 2 5
Z 12 5 11

a) Find an initial solution for this transportation problem.

b) Determine an optimal solution and the associated total transportation cost.
c) Which factory will have the surplus? State the surplus.
d) Is there any alternative optimal solution to this problem? If yes, find the alternative optimal
solution.

70
Solution: Ans: Total Cost = RM 3430, Total Cost=RM 3190,
Factory Z will have the surplus of 20 units
From/To P Q R Dummy Supply
X 7 4 6 0 300
70 230
Y 6 2 5 0 200
130 70
Z 12 5 11 0 100
80
Demand 200 150 230 20 600

71
72
HANDS-ON EXAMPLE 5

Sasuki Manufacturing Company has orders for three kinds of product manufactured in their
factory located in Pasir Gudang.

Product Orders (Unit)

A 1300
B 800
C 1200

Three machines are available for the manufacturing operations. All three machines can
produce all the products at the same production rate. However, due to varying defect
percentages of each machine, the unit costs of products vary depending on the machine used.
Machine capacities for the next week, and the unit costs, are as follows:

Capacity (Units) Product

Machine
Machine A B C
1 1300 1 7 6 6
2 1800 2 12 7 2
3 900 3 8 5 9

a) Find the minimum-cost production schedule for the products and machines by using the
Stepping-Stone Method.
b) Is the solution unique? Explain.

Solution:
Ans: Total Min. Cost = RM15,500, the solution is unique

73
74
 HANDS-ON EXAMPLE 6

Ceria Company manufactures tires in two factories XY and PQ and ships them to four
distribution centres in Johor Bahru, Penang, Petaling Jaya and Kuala Terengganu. The
following table shows the cost (RM) per unit, capacity and demand.

To Distribution Centre
Johor Bahru Penang Petaling Jaya Kuala Terengganu Capacity
From
XY 7 11 5 12 300
PQ 10 15 8 16 200
Demand 150 120 50 180

a) Find the optimal solution to the problem.

b) Calculate the total transportation cost.

Solution:
Ans:
To Distribution Centre
Johor Bahru Penang Petaling Jaya Kuala Terengganu Capacit
From (K1=7) (K2=11) (K3=5) (K4=12)
XY 7 11 (120) 5 12 300
(R1=0) [0] (0)** (180)
PQ 10 15 8 16 200
(R2=3) (150) [1] (50) [1]
Demand 150 120 50 180 500
Total transportation=RM5380

75
76
77