ADVANCED

OPERATIONS
RESEARCH
A O R
Operations Research: An overview &
Introduction to Linear Programming
By: -
Hakeem–Ur–Rehman
IQTM–PU 1
 WHAT IS OPERATIONS RESARCH?
“THE SCIENCE OF BETTER”
Operations research
 A rational, structured approach to problem solving
 Taking (often) a quantitative view of decision making
 Systems approach to problem solving
• To improve operations in manufacturing, service and
governments through the use of scientific methods
and the development of specialized techniques
Optimal decision-making in, and modeling of, deterministic and probabilistic
systems that originate from real life. These applications, which occur in
government, business, engineering, economics, and the natural and social
sciences, are largely characterized by the need to allocate limited resources.
In these situations, considerable insight can be obtained from scientific
analysis, such as that provided by Operations Research.
[By: - Hiller-Lieberman]
2
 OR: RESEARCH
 OR: a new field which started in the late 1930's and has
grown and expanded tremendously in the last 30 years

 The ACADEMIC JOURNALS: state of the art theory and

applications of OR
 Operations Research
 Management Science
 Mathematics of Operations Research
 Mathematical Programming
 European Journal of Operational Research
 Journal of the Operational Research Society
 Naval Research Logistics
 Operations Research Letters
 Computer and Operations Research
 Interfaces (applications oriented) 3
 OPERATIONS REASECH APPROACH
1. Define problem and gather data
 OR teams work in advisory capacity
 Determine appropriate objectives from management
 Concerned with the entire organization
 Data gathering

2. Formulate mathematical model

 Problem identified with DECISION VARIABLES
– How many units to buy/sell ...
– How much time to spend on a task ...
 Measure of performance is the OBJECTIVE FUNCTION
– What is the goal?;
– Usually: Max/min profit/cost/time/units
– A function of the decision variables
 Restrictions of values of decision variables set in CONSTRAINTS
– Min acceptable profit
– Max available resources
 PARAMETERS are the constants of the objective function and the constraints

4
 OPERATIONS REASECH APPROACH…
SYSTEM Vs Its MODEL

REAL  SIMPLIFICATION MODEL

SYSTEM  ABSTRACTION
 ASSUMPTIONS

5
 OPERATIONS REASECH APPROACH…
3. Develop a computer-based procedure for deriving solutions from the
model
 Mathematical representations are always approximations of the real world
 Type of model dictates the type of algorithm to use to obtain solution
 Models can be:
o DETERMINISTIC
o STOCHASTIC

4. Testing the Model

 Must ‘debug’ the model as with a computer program
 Process of testing/improving model is known as model validation

5. Preparing to Apply the Model

 Install a system for applying the model
 Usually computer-based systems that are provided with up-to-date input

6. Implementation
 OR team explains system to management
 Develop procedures required to put system into operation
 Management trains personnel
6
 MODEL & Its TYPES
 MODEL: A model is a representation of the structure of a real life
system.
– In general, models can be classified as fellows:
• Iconic models
• Analogue models
• Symbolic models

 SYMBOLIC MODELS: Symbolic (i.e., algebraic, numerical, logical)

models represent the properties of the real life system through the means
of symbols, mathematical equations, computer programs and simulation
models are also symbolic models.
– DETERMINISTIC MODELS: Deterministic models are models which
do not contain the element of probability. Deterministic models involve
optimization.
– STOCHASTIC MODELS: stochastic models are models which contain
the element of probability. Stochastic models characterize/estimate
system performance.
7
 OPERATIONS
RESEARCH: TOPICS
 Topics in Operations Research
o Deterministic models: Linear programming, integer
programming, dynamic programming, network
programming, nonlinear programming, goal programming
o Programming: planning of activities

o Stochastic models: Queuing theory, Markov decision

processes, and simulation

o Decision making models: Decision analysis, multi-

criteria and multi-objective decision making, and game
theory
8
 LINEAR PROGRAMMING:
AN INTRODUCTION
 First conceived by George B. Dantzig around 1947.

 The work of Kantorovich “Russian Mathematician” (1939)

was published in 1959.

 Dantzig’s first paper was titled “Programming in Linear

Structure”

 Koopmans Coined the term “Linear Programming” in

1948.

 Simplex Method was published in 1949 by Dantzig.

9
WHAT IS LINEAR PROGRAMMING?
A linear programming problem (LP) is an optimization problem for which:
– We attempt to maximize (or minimize) a linear function of the decision variables.
(Objective Function)
– The values of the decision variables must satisfy a set of constraints, each of
which must be a linear inequality or linear equality.
– A sign restriction on each variable. For each variable Xi the sign restriction can
either say
– Xi ≥ 0,
– Xi ≤ 0,
– Xi unrestricted.

10

ASSUMPTIONS
LINEARITY OR PROPORTIONALITY:
IN LP MODEL
– Proportionality means that the objective function and constraint coefficients are strictly
proportional to the decision variable (e.g., If the first unit of production requires ‘2’ hours of
labor so it must the 50th and 100th unit also requires ‘2’ hours of labor).
 DIVISIBILITY:
– Divisibility means that non integer (fractional) values of the decision variables are acceptable.
(e.g., LP allows a production program which uses 400 units of electronic component and
68.33 man hours of labor time to produce 40 units of black and white television sets and
25.7 units of colored television sets per week. If a fraction of a product cannot be produced,
Integer Programming is a special technique which can be used for finding non-fractional
values of resource usage and decision variables).
 CERTAINTY:
– Certainty means that the values of the parameters are known and constant
 ADDITIVITY:
– Additivity means the total effect of each decision variable (Profit, Cost, etc.) must equal the
sum of the effects contributed by each decision variable and terms of each constraint must
be additive (Total amount of resource consumed or provided) must equal the sum of the
resources used (or provided) by each decision variable.
– (e.g., if an objective is to maximize profit equal to Rs. 10 per unit of the first product made
plus Rs. 5 per unit of the second product made, and if one unit of each product is actually
produced, then the profit contributions of Rs. 10 and Rs. 5 must be added up to produce a
sum of Rs. 15).
 NON–NEGATIVITY:
– Non – negativity means that the decision variables are permitted to have only the values
11
which are greater than or equal to zero.
 DEVELOPMENT (FORMULATION) OF LP MODEL
PRODUCT (PRODUCTION) MIX PROBLEM # 1:
A firm is engaged in producing two products ‘A’ and ‘B’. Each unit of product ‘A’ requires
3Kg of raw material and 5 labor hour for processing, where as each unit of product ‘B’
requires 6Kg of raw material and 4 labor hours of the same type. Every month the firm
has the availability of 60Kg of raw material and 70 Labor hours. One unit of product ‘A’
sold earns profit Rs. 30 and one unit of product ‘B’ sold gives Rs. 40 as profit.
Formulate this problem as linear programming problem to determine as to how many
units of each of the products should be produced per month so that the firm can earn
maximum profit, assume all unit produced can be sold in the market.

CONSTRAINTS:
Decision Variables: Let X1 and X2 be the
Material Constraint: 3X1 + 6X2 ≤ 60
number of products ‘A’ and ‘B’ respectively.
Labor Constraint: 5X1 + 4X2 ≤ 70
Objective Function: Non–negativity Constraint: As X1 and X2, being the
Product – A: As, ‘X1’ are the units of product number of units produced of products ‘A’ and
‘A’, So, Product ‘A’ contributes a profit of Rs. ‘B’ cannot have negative values thus, X1≥0 and
30X1 from one unit of product. X2≥0. i.e. X1,X2 ≥ 0.
Product – B: As, X2 are the units of product ‘B’, The Complete LP problem model is:
So, Product ‘B’ contributes a profit of Rs. 40X2 Maximize: Z = 30X1 + 40X2
from one unit of product.
Subject to:
Maximize Z = 30X1 + 40X2 3X1 + 6X2 ≤ 60
5X1 + 4X2 ≤ 70
12
X1 , X2 ≥ 0
 DEVELOPMENT (FORMULATION) OF LP MODEL
PRODUCT (PRODUCTION) MIX PROBLEM # 2: Four varieties of ties produced:
1. one is an expensive, all-silk tie,
2. one is an all-polyester tie, and
3. two are blends of polyester and cotton.
The table on the following slide illustrates the cost and availability (per monthly production planning period)
of the three materials used in the production process
Monthly cost &
availability of
material

 The following table summarizes the contract demand for each of

o the four styles of ties,
o the selling price per tie, and
o the fabric requirements of each variety.

Fifth Avenue’s goal is to maximize its monthly profit. It must decide upon a policy for product mix.
LINEAR PROGRAMMING: MODEL FORMULATION
DECISION VARIABLES: Let X1 = # of all-silk ties; X2 = # polyester ties; X3 = # of
blend 1 poly-cotton ties; X4 = # of blend 2 poly-cotton ties produced per month
Calculate profit for each tie: Profit = Sales price – Cost per yard X Yards per tie
Silk ties = $6.70 – $21 x 0.125 = $4.08
Polyester = $3.55 – $6 x 0.08 = $3.07
Poly-blend 1 = $4.31 – ($6 x 0.05 + $9 x 0.05) = $3.56
Poly-blend 2 = $4.81 – ($6 x 0.03 + $9 x 0.07) = $4.00
Objective function: maximize profit = $4.08X1 + $3.07X2 + $3.56X3 + $4.00X4
Subject to:
0.125X1 ≤ 800 (Total Silk Availability Constraint)
0.08X2 + 0.05X3 + 0.03X4 ≤ 3,000 (Total Polyester Availability Constraint)
0.05X3 + 0.07X4 ≤ 1,600 (Total Cotton Availability Constraint)
X1 ≥ 6,000 (Contract Constraint – 1)
X1 ≤ 7,000 (Demand Constraint – 1)
X2 ≥ 10,000 (Contract Constraint – 2)
X2 ≤ 14,000 (Demand Constraint – 2)
X3 ≥ 13,000 (Contract Constraint – 3)
X3 ≤ 16,000 (Demand Constraint – 3)
X4 ≥ 6,000 (Contract Constraint – 4)
X4 ≤ 8,500 (Demand Constraint – 4)
X1, X2, X3, X4 ≥ 0
 LINEAR PROGRAMMING: MODEL FORMULATION
PORTFOLIO SELECTION PROBLEM:
Mr. Ali has Rs. 70, 000 to investment in several alternatives. The alternative
investments are national certificates with an 8.5% return, Defense Savings
Certificates with a 10% return, NIT with a 6.5% return, and khas deposit with a
return of 13%. Each alternative has the same time until maturity. In addition, each
investment alternative has a different perceived risk thus creating a desire to
diversify. Ali wants to know how much to invest in each alternative in order to
maximize the return.
The following guidelines have been established for diversifying the
investments and lessening the risk;
 No more than 20% of the total investment should be in khas deposit.
 The amount invested in Defense Savings Certificates should not exceed
the amount invested in the other three alternatives.
 At least 30% of the investment should be in NIT and Defense Savings
Certificates.
 The ratio of the amount invested in national certificates to the amount
invested in NIT should not exceed one to three.
Formulate the problem as a LP model.
LINEAR PROGRAMMING: MODEL FORMULATION
DECISION VARIABLES: Let X1, X2, X3 & X4 be the amount (Rs.) invested in national
certificates, Defense savings certificates, NIT, and khas deposit, respectively.

OBJECTIVE FUNCTION: Maximize (the total return from all investment):

Z = 0.085X1 + 0.100X2 + 0.065X3 + 0.130X4
CONSTRAINTS: 
X1 + X2 + X3 + X4 = 70,000 (Amount Availability Constraint)
X4 ≤ (0.20 x 70,000 = 14,000) (Investment Constraint – i)
X2 ≤ (X1 + X3 + X4)  X2 – X1 – X3 – X4 ≤ 0 (Investment Constraint – ii)
X2 + X3 ≥ (0.30 x 70,000 = 21, 000) (Investment Constraint – iii)
(X1) / (X3) ≤ 1/3  3X1 – X3 ≤ 0 (Investment Constraint – iv)
X1, X2, X3 , X4 ≥ 0 (Non–Negativity Constraint)
LP MODEL:
Maximize Z = 0.085X1 + 0.100X2 + 0.065X3 + 0.130X4 (Investment return Function)
SUBJECT TO:
X1 + X2 + X3 + X4 = 70,000 (Amount Availability Constraint)
X4 ≤ 14,000 (Investment Constraint – i)
X2 – X1 – X3 – X4 ≤ 0 (Investment Constraint – ii)
X2 + X3 ≥ 21, 000 (Investment Constraint – iii)
3X1 – X3 ≤ 0 (Investment Constraint – iv)
X1, X2, X3, X4 ≥ 0 (Non – Negativity Constraint)
 LINEAR PROGRAMMING: MODEL FORMULATION

PRODUCTION PLANNING PROBLEM:

Let us consider a company making a single product.
The estimated demand for the product for the next four months
are 1000, 800, 1200, 900 respectively.
The company has a regular time capacity of 800 per month and
an over time capacity of 200 per month.
The cost of regular time production is Rs. 20 per unit and the
cost of over time production is Rs. 25 per unit.
The company can carry inventory to the next month and the
holding cost is Rs. 3 per unit per month.
The demand has to be met every month.
Formulate linear programming problem for the above situation.
LINEAR PROGRAMMING: MODEL FORMULATION
1st Way
DECISION VARIABLES:
 Let Xj = Quantity Produced Using Regular time in month ‘j’  j = 1, 2, 3, 4
 Let Yj = Quantity Produced Using Over time in month ‘j’  j = 1, 2, 3, 4
 Let Ij = Quantity Carried at the end of month ‘j’ to the next month  j = 1, 2, 3

OBJECTIVE FUNCTION:
Minimize (the total cost):
Z = 20 (X1+X2+X3+X4) + 25(Y1+Y2+Y3+Y4) + 3 (I1+I2+I3)
CONSTRAINTS: 
Demand Constraints:
X1 + Y1 = 1000 + I1
I1 + X2 + Y2 = 800 + I2
I2 + X3 + Y3 = 1200 + I3
I3 + X4 + Y4 = 900
Capacity Constraints:
Xj ≤ 800 Where j = 1, 2, 3, 4
Yj ≤ 200 Where j = 1, 2, 3, 4
Non–Negativity Constraints:
Xj , Yj , Ij ≥ 0

12 Constraints & 11 Variables

LINEAR PROGRAMMING: MODEL FORMULATION
2nd Way
DECISION VARIABLES:
 Let Xj = Quantity Produced Using Regular time in month ‘j’  j = 1, 2, 3, 4
 Let Yj = Quantity Produced Using Over time in month ‘j’  j = 1, 2, 3, 4
OBJECTIVE FUNCTION:
Minimize (the total cost):
Z = 20 (X1+X2+X3+X4) + 25(Y1+Y2+Y3+Y4) + 3 (X1 + Y1 – 1000) + 3 (X1 + Y1 + X2
+ Y2 – 1800) + 3 (X1 + Y1 + X2 + Y2 + X3 + Y3 – 3000) + 3 (X1 + Y1 + X2 + Y2 +
X3 + Y3 + X4 + Y4 – 3900)

CONSTRAINTS: 
Demand Constraints:
X1 + Y1 ≥ 1000
X1 + Y1 – 1000 + X2 + Y2 ≥ 800
X1 + Y1 + X2 + Y2 – 1000 – 800 + X3 + Y3 ≥ 1200
X1 + Y1 + X2 + Y2 + X3 + Y3 – 1000 – 800 – 1200 + X4 + Y4 ≥ 900
Capacity Constraints:
Xj ≤ 800 Where j = 1, 2, 3, 4
Yj ≤ 200 Where j = 1, 2, 3, 4
Non–Negativity Constraints:
Xj , Yj , Ij ≥ 0
12 Constraints & 8 Variables
LINEAR PROGRAMMING: MODEL FORMULATION
3 rd Way
DECISION VARIABLES:
 Let Xijk = Quantity Produced in month ‘i’ (i = 1, 2, 3, 4) to meet the demand of month ‘j’ (j =
1, 2, 3, 4) using production type ‘k’ (k = 1, 2)
OBJECTIVE FUNCTION: Minimize (the total cost):
Z = 20 (X111+X121+X131+X141 + X221+X231+X241+X331+X341+X441)+ 25(X112+X122+X132+X142+
X222+X232+X242+X332+X342+X442) + 3 (X121+X122+X231+X232+X341+X342)+ 6 (X131+X132+X241+X242) +
9 (X141+X142)
CONSTRAINTS: 
Demand Constraints:
X111 + X112 = 1000
X121 + X122 + X221 + X222 = 800
X131 + X132 + X231 + X232 + X331 + X332 = 1200
X141 + X142 + X241 + X242 + X341 + X342 + X441 + X442 = 900
Capacity Constraints:
X111 + X121 + X131 + X141 ≤ 800
X221 + X231 + X241 ≤ 800
X331 + X341 ≤ 800
X441 ≤ 800
X112 + X122 + X132 + X142 ≤ 200
X222 + X232 + X242 ≤ 200
X332 + X342 ≤ 200
X442 ≤ 200
Non–Negativity Constraints:
Xjjk ≥ 0
LINEAR PROGRAMMING: MODEL FORMULATION
The Medequip Company produces precision medical diagnostic
equipment at two factories. Three medical centers have placed orders
for this month’s production output. The below table shows what the
cost would be for shipping each unit from each factory to each of these
customers. Also shown are the number of units that will be produced
at each factory and the number of units ordered by each customer.

A decision now needs to be made about the shipping plan for

how many units to ship from each factory to each customer.
Formulate a linear programming model for this problem.
LINEAR PROGRAMMING: MODEL FORMULATION
 LP SOLUTION: GRAPHICAL METHOD
A Carpenter wants to maximize his profit; the LP model shows
his objective and constraints:
Maximize: 5 X1 + 3 X2
(Where X1 = Number of Chairs, X2 = Number of Tables)
Subject to:
2 X1 + X2 ≤ 40
X1 + 2 X2 ≤ 50
X1, X2 ≥ 0
Solve the following LP problem graphically:
Minimize: 6 X1 + 14 X2
Subject to:
5X1 + 4X2 ≥ 60
3X1 + 7X2 ≤ 84
X1 + 2X2 ≥ 18
X1, X2 ≥ 0
CLASSIFICATION OF LP SOLUTION
CLASSIFICATION OF LP SOLUTION
 FEASIBLE SOLUTION REGION / FEASIBLE REGION: A
point is feasible solution if it satisfies all constraints. The
feasible solution region is the set of all feasible solution
points.

 BOUNDED FEASIBLE REGION: A feasible region is that

can be enclosed in a circle (mean’s closed area) is to be
bounded feasible region.

 UNBOUNDED FEASIBLE REGION: An unbounded

feasible region that can not be enclosed in a circle (mean’s
at least open (area) from one side).

 OPTIMAL SOLUTION: A feasible solution point where the

objective function meet the desired objective / goal (i.e.:
maximization or minimization) to be optimal solution.
 CLASSIFICATION OF LP SOLUTION
• MULTIPLE / ALTERNATE OPTIMUM SOLUTION:
– More than One Optimal Solution:
• Two or more optimal solutions may exist, and
• This actually allows management great flexibility in deciding
which combination to select.

Maximize: Z = 4X1 + 4X2

Subject to:
X1 + 2X2 ≤ 10
6X1 + 6X2 ≤ 36
X1 ≤ 4
X1, X2 ≥0

As, at point ‘A’ (2,4) the value of the objective

function is: Z = 4(2) + 4(4) = 24, and at
point ‘B’ (4,2) the value of the objective
function is: Z = 4(2) + 4(4) = 24.
 CLASSIFICATION OF LP SOLUTION
• UNBOUNDED SOLUTION: In some LP models, the
values of the variables may be increased indefinitely without
violating any of the constraints in this result the solution space
becomes unbounded so As a result the value of the objective
function may increase (in Maximization Case) or decrease (in
Minimization Case) indefinitely. But it is wrong to conclude that
just because the solution space is unbounded then solution also
unbounded. The solution space may be unbounded but the
solution may be finite.

Maximize: Z = 2X1 + X2
Subject to: X1 – X2 ≤ 10
2X1 ≤ 40
X1, X2 ≥ 0
 CLASSIFICATION OF LP SOLUTION
• UNBOUNDED SOLUTION: (EXAMPLE # 2)

Minimize: 6X1 + 4X2

Subject to:
3X1 + 3X2 ≥ 40
3X1 + X2 ≥ 40
2X1 + 5X2 ≥ 44
X1, X2 ≥ 0

Optimal value to be 88, at X1 = 12, and X2 = 4.

 CLASSIFICATION OF LP SOLUTION
• DEGENERACY: In LP problems, intersection of two
constraints will define a corner point of the feasible region.
But if more than two constraints pass through any one of
the corner points of the feasible region, excess constraints
will not serve any purpose, and therefore they act as
redundant constraints, so redundant constraints create the
Degeneracy.
Maximize: Z = 4X1 + 6X2
Subject to:
6X1 + 4X2 ≤ 24
X2 ≤ 3
5X1 + 10X2 ≤ 40
X1, X2 ≥ 0

Optimal value to be 26, at X1 = 2, and X2 = 3.

 CLASSIFICATION OF LP SOLUTION
• NON–EXISTING / INFEASIBLE SOLUTION: A
linear programming problem is infeasible if there is no
solution exists that satisfies all of the constraints.

Maximize: Z = 3X1 + 2X2

Subject to:
2X1 + X2 ≤ 2
3X1 + 4X2 ≥ 12
X1, X2 ≥ 0
QUESTIONS

31