CE 491: Project Planning and Construction Management

Lecture on: The Linear Programming Model

Course Teacher
Professor Dr. Md. Mazharul Hoque

Dean, Faculty of Science and Engineering

Professor, Department of Civil Engineering
University of Information Technology & Sciences (UITS)

6/7/2023 1
 The Linear Programming(LP) Model

LP deals with the problem of allocating limited resources

among competing activities in an optimal manner. Linear
means that all mathematical functions in this model are
required to be linear functions.

6/7/2023 2
 The Linear Programming(LP) Model

General Form

The mathematical statement of a general form of the LP problem is the

following:
Find X1, X2, X3…………………Xn which maximizes the linear functions.

Z= C1X1 + C2X2 +……………………………….+ CnXn ... (1st)

Subject to the restrictions,

a11 x1+ a12 x2 + a13 x3+………………………………+a1n xn ≤ b1

a11 x1+ a12 x2 + a13 x3+………………………………+a2n xn ≤ b2
a11 x1+ a12 x2 + a13 x3+………………………………+a3n xn ≤ b3
… …. (2nd)
…
…
am1 x1+ am2 x2 + am3 x3+…………………………… +Cmn xn ≤ bn
6/7/2023 3
And

X1 ≥ 0 , X2 ≥ 0, …………………………………….. , Xn ≥ 0 } … (3rd )

Where aij , bi and Cj are given constants.

Special Types of linear programming problems:

- Transportation problem (for optimal shipping pattern)

- Assignment Problem ( extension of the above)
- Transshipment Problem ( extension of the above)

It is necessary to identify “Sources” and “Destinations”.

6/7/2023 4
 Definitions

The functions being maximized is called objective

functions. The restrictions are also referred to as
constraints or restraints. The variables being solved for
are called decision variables.

5
 Interpretation
Z is the chosen overall measure of effectiveness e.g. profit.
- Maximize profit
- Maximize B/C ratio
- Maximize sales
- Maximize cost
- Minimize risk

bi amount of resources is available to the n activities

aij amount of the resources “ i “ is assumed by each unit of
activity “ j ”

Left side of these inequalities is the total usage of the respective

resources. The non-negatively restrictions (xij ≥ 0 ) rule out the
possibility of negative activity levels.

Importance Properties of LP

6/7/2023 6
 Formulation of the LP Problem
Example 1: Available machine capacities are

Machine Type Available time/ Week (Hours)

Milling Machine 200

Lathe Machine 100
Grinder Machine 50

The number of the machine hours required for each unit of

the respective products is given bellow:
Machine
Product 1 Product 2 Product 3
Type
Milling 8 2 3
Lathe 4 3 -
Grinder 2 - 1
6/7/2023 7
 Formulation of the LP Problem

Sales potential for products 1 and 2 exceeds the maximum

production rate sales potential for product 3 is 20 units/ week.
The unit profit would be $20, $6 and $8 respectively on
products 1, 2, 3 respectively. The problem is to formulate a
linear programming model for determining how much of each
product the firm should produce in order to Maximize Profit.

6/7/2023 8
Formulation: LP model
Let Xi (i=1, 2, 3,) be the number of units of product i produced per
week. Since the profit has been chosen as the measure of
effectiveness,

The objective is to maximize

Z = 20 X1 + 6 X2 + 8 X3

Subject to the restrictions:

“limited resources” are available machine capacities. Sales
potential for product 3. The mathematical statements of constraints
are:
8 X1 + 2X2 + 3X3 ≤ 200
4X1 + 3X2 ≤ 100
2X1 + + X3 ≤ 50

Sales restriction
X3 ≤ 20

6/7/2023 9
Summary of LP:

Maximum Z = 20X1 + 6X2 + 8X3 ……… (1)

Subject to restriction

St
8 X1 + 2X2 + 3X3 ≤ 200
4X1 + 3X2 ≤ 100 ….. (2)
2X1 + + X3 ≤ 50
X3 ≤ 20

And
X1 ≥ 0, X2 ≥ 0, X3 ≥ 0 ………(3)

6/7/2023 10
 LP model: Solution

The “simplex method”

- Graphically
- Algebraically

The method is an algebraic procedure which progressively

approaches to the optimal solution through a well-defined
iterative process until optimally is finally reached.

6/7/2023 11
 LP model: Solution

Example
Consider the following linear programming model.
Maximize
Z = 3X1 + 5X2

Subject to the restrictions (St)

X1 ≤ 4
X2 ≤ 6
3X1 + 2X2 ≤ 18

X1 ≥ 0, X2 ≥ 0

6/7/2023 12
 LP model: Solution

How this problem be solved? For this very simple case

where there are only two variables and therefore only two
dimensions variables and therefore only two dimensions
involved, a simple graphical method can be used. This will
illustrate what simple method does.

6/7/2023 13
 10

X1 = 4
7
c X2 = 6
d
X2 6

b 3X1 + 2X2 = 18
3

a
0
0 1 2 3 4 5 6 7 8 9 10
X1

Figure: Lp Model Graphical Solution

Shaded areas shows permissible values of (X1, X2)
14
 Important
a) Shaded area is the Feasible solution
b) The Basic Feasible solutions correspond the “extreme”
points of the set of feasible solutions= (o,a,b,c,d)
c) An optional solution is a Feasible solution which
maximizes or minimizes objective function.
d) One of the Extreme Points will be the optimal Solution
(C) = (X1 = 2 , X2 = 6)
i.e. Z = 36 (X1 = 2 , X2 = 6)

15
 Graphical Procedure
1st step: Plot graphically the values of (X1, X2 ) which are
permitted by the restrictions

2nd Step: Pick out the points in this region which maximizes the
value of
This is done by trial and error procedure.

Try Z=0, by a line ( Z = 3X1 + 5X2 )

Draw lines, m= - 3/5

Z = 0 (X1=0, X2=0); Z= 30 (X1=0, X2=6)

*Z=36 (X1 = 2, X2 = 6); Z = 27 (X1 =4, X2 = 3)

Z=12 (X1=4, X2=0)

Optimal Solution Z = 36 (X1=2, X2 =6)

6/7/2023 16
 Simplex Method (Algebraic Procedure)

It was pointed out that the simplex method is an algebraic

procedure, whereas the proceeding discussion has discussed
only geometrical concepts involving extreme points. How
does one translate this conceptually geometrical procedure
into a usable algebraic procedure?

6/7/2023 17
 Simplex Method (Algebraic Procedure)

The crux of the problem is to identify extreme points (basic

feasible solutions) algebraically.

The first is to convert the model into an equivalent model

containing no in equalities using “slack variable”

Consider the inequality X1 ≤ 4 . How this can be replaced by an

equivalent equation?

Let, X3 = 4 – X1, so that X3 is just the slack between the two

sides of the inequality, X3 is called a slacked variable.
Therefore,
X1 + X3 = 4

6/7/2023 18
 Simplex Method (Algebraic Procedure)

The original constraint X1 ≤ 4 holds whenever. Hence will be replaced

by
X1 + X3 = 4
X1 + X3 = 4 and X3 ≥ 0

Introducing slack variables in the restrictions, the original LP model can

now be replaced by equivalent model as

Max Z = 3X1 + 5X2

Or Z - 3X1 - 5X2 =0

S.t.
X1 + X3 =4
X2 + X4 =6
3X1 + 2X2 + X5 = 18

And Xj ≥ 0 for j = 1, 2, 3 ………… 5

6/7/2023 19
 LP (Algebraic Solution)

The following is the augmented form of the “Simplex Tableau” as it

would appear for the initial iteration.
Coefficients of
Eqn. Basic RHS
operations
No variables Z X1 X2 X3 X4 X5 of Eqn

I Z 1 -3 -5 0 0 0 0
Initial
II X3 0 1 0 1 0 0 4
iteration
III X4 0 0 1 0 1 0 6
IV X5 0 3 2 0 0 1 18

Identity Matrix
This is called the simplex tableau:
Solution: Z = 0, X3 = 4, X4 = 6, X5 = 18, X1 = 0, X2 =0
Now we have to proceed further for optimal equation by choosing
“entering” and “leaving” basic variables.
6/7/2023 20
 Algebraic Solution

Select the smallest coefficient of row I, i.e. (-5), so X2

would be new entering variable. Excluding row I, go down
through X2 column for (bi/ai) ratios i.e. 6/1, 18/2, 4/0.
Select the lowest ratio; here it is 6/1; so X4 must be
leaving variable.
Then eliminate X2 from other equations except III. This
can be done by subtracting (-5), 0, & 2 times row II from
these rows respectively.

6/7/2023 21
 Algebraic Solution
Eqn. Basic Coefficients of RHS of
operations
No variables Z X1 X2 X3 X4 X5 Eqn
I-(-5)III I1 Z 1 -3 0 0 5 0 30
II-0(III) II2 X3 0 1 0 1 0 0 4
III3 X2 0 0 1 0 1 0 6
IV-2(III) IV4 X5 0 3 0 0 -2 1 6

2nd Iteration: Solution, Z=30, X3=4, X2= 6, X5=6, X1=0, X4 =0

Assignment:

• Find entering basic variable and leaving basic variable

• Find optimal Solution

6/7/2023 22
 Algebraic Solution (continue)

Again consider row I1 (objective function). Since one of the coefficient

i.e. (-3) is negative, the solution cannot be optimal. So another iteration is
required.

X1 must be entering basic variable X5 must be leaving basic variable.

Eliminate X1 form the equations except IV1

Get the optimal Solution

Z=36 X1=
X2=
X3 = ??????
X4 = ??????
X5 = ??????
X

6/7/2023 23