01/02/2023

“MG5580 – Business
Modelling and Simulation
in Supply Chains”

LECTURE: Linear
Programming

Dr Behzad Hezarkhani

Agenda
❖Introductory Example
❖What is Linear Programming (LP)?
❖What Constitutes a Linear Program?
❖Standard forms of LP
❖Exercise
❖Graphical Solution to 2-Dimensional Problem
❖Possible Cases with Optimal Solutions
❖Solving Linear Programmes in General
❖Important Topics in Linear Programming
❖Working with Excel Solver

Brunel Business School 2

1
 01/02/2023

What is Linear Programming?

• Linear Programming (LP) is a tool for formulating and

solving optimization problems
• Founded by George Dantzig in 1947
• Initially developed to solve optimization problems in
military
• LP is applied everywhere today (banking, forestry,
trucking, education, etc.)
• 85% of Furtune 500 firms use LP

Brunel Business School 3

George Dantzig (1914 –2005)

Introductory Example

A company makes two products, product 1 and product 2.

The profit from selling one unit of product 1 is £5 and the
profit from selling one unit of product 2 is £10.
The company wants to optimize its profit by deciding the
number products that it must produce in the next month.

Profit = 5 𝑥1 + 10 𝑥2

Brunel Business School 4

2
 01/02/2023

Introductory Example

But there are constraints for production.

Product 1 requires 3 hours of work by one employee and product 2 requires 8 hours of
work by one employee. And there are only 240 employee hours available in the next
month.
Constraint: 3 𝑥1 + 8 𝑥2 ≤ 240

Also, the company already promised a customer 15 units of product 2. Therefore,

production number for product 2 must at least be 15 units.
Constraint: 𝑥2 ≥ 15
Obviously, production numbers must be non-negative as well.
Constraint: 𝑥1 ≥ 0
Constraint: 𝑥2 ≥ 0
Brunel Business School 5

Introductory Example

The Linear Program:

maximize 5 𝑥1 + 10 𝑥2
subject to
3 𝑥1 + 8 𝑥2 ≤ 240
𝑥2 ≥ 15
𝑥1 ≥ 0
𝑥2 ≥ 0

Brunel Business School 6

3
 01/02/2023

What constitutes a Linear Programme?

• A linear program has two parts:

• An Objective Function: a function that one tries
to optimize (that is, either to maximize or
minimize) by choosing some decision
variables,
• A set of Constraints: the conditions that need to
be met by decision variables.

Brunel Business School 7

“Linear” Programming

“Linear” in linear programming requires objective function and all constraints to

be linear:

A function 𝑓(𝑥1 , 𝑥2 , … , 𝑥𝑛 ) of 𝑥1 , 𝑥2 , … , 𝑥𝑛 is a linear function if and only if

for some set of constants 𝑐1 , 𝑐2,…, 𝑐𝑛 we have

𝑓(𝑥1 , 𝑥2 , … , 𝑥𝑛 ) = 𝑐1 𝑥1 +𝑐2 𝑥2 + ⋯ + 𝑐𝑛 𝑥𝑛

Examples:
Function Linear?
𝑥1 + 𝑥22 NO!
𝑥1 NO!
𝑥2
𝑥1 + 𝑥2 + |𝑥3 | NO!
2 Yes
−2𝑥1 + 𝑥2 − 7𝑥3
3

Brunel Business School 8

4
 01/02/2023

Standard Form of a Linear Program

Standard form of a maximization problem:

max 𝑧 = 𝑐1 𝑥1 +𝑐2 𝑥2 + ⋯ + 𝑐𝑛 𝑥𝑛
Subject To (s.t.)

𝑎1,1 𝑥1 + 𝑎1,2 𝑥2 + ⋯ + 𝑎1,𝑛 𝑥𝑛 ≤ 𝑏1

𝑎2,1 𝑥1 + 𝑎2,2 𝑥2 + ⋯ + 𝑎2,𝑛 𝑥𝑛 ≤ 𝑏2
…
𝑎𝑚,1 𝑥1 + 𝑎𝑚,2 𝑥2 + ⋯ + 𝑎𝑚,𝑛 𝑥𝑛 ≤ 𝑏𝑚
𝑥1 , 𝑥2 , … , 𝑥𝑛 ≥ 0

Brunel Business School 9

Standard Form of a Linear Program

Standard form of a minimization problem:

min 𝑧 = 𝑐1 𝑥1 +𝑐2 𝑥2 + ⋯ + 𝑐𝑛 𝑥𝑛
Subject To (s.t.)

𝑎1,1 𝑥1 + 𝑎1,2 𝑥2 + ⋯ + 𝑎1,𝑛 𝑥𝑛 ≥ 𝑏1

𝑎2,1 𝑥1 + 𝑎2,2 𝑥2 + ⋯ + 𝑎2,𝑛 𝑥𝑛 ≥ 𝑏2
…
𝑎𝑚,1 𝑥1 + 𝑎𝑚,2 𝑥2 + ⋯ + 𝑎𝑚,𝑛 𝑥𝑛 ≥ 𝑏𝑚
𝑥1 , 𝑥2 , … , 𝑥𝑛 ≥ 0

Brunel Business School 10

5
 01/02/2023

Exercise: Formulate the Program

Two Crude Petroleum runs a small refinery on the Texas coast. The refinery distils crude
petroleum from two sources, Saudi Arabia and Venezuela, into the three main products:
gasoline, jet fuel and lubricants.
The two crudes differ in chemical composition and yield different product mixes. Each
barrel of Saudi crude yields 0.3 barrel of gasoline, 0.4 barrel of jet fuel, and 0.2 barrel of
lubricants. Each barrel of Venezuelan crude yields 0.4 barrel of gasoline, 0.2 barrel of jet
fuel and 0.3 barrel of lubricants. The remaining 10% is lost to refining.
The crudes differ in cost and availability. Two Crude can purchase up to 9000 barrels per
day from Saudi Arabia at $20 per barrel. Up to 6000 barrels per day of Venezuelan
petroleum are available at the lower cost of $15 per barrel.
Two Crude’s contracts with independent distributors require it to produce 2000 barrels per
day of gasoline,1500 barrels per day of jet fuel and 500 barrels per day of lubricants. How
can these requirements be fulfilled most efficiently?

Brunel Business School 11

Solution to Two Crude Problem

min 𝑧 = 20𝑥1 +15𝑥2

s.t.

0.3𝑥1 + 0.4𝑥2 ≥ 2.0

0.4𝑥1 + 0.2𝑥2 ≥ 1.5
0.2𝑥1 + 0.3𝑥2 ≥ 0.5
𝑥1 ≤ 9
𝑥2 ≤ 6
𝑥1 , 𝑥2 ≥ 0

Brunel Business School 12

6
 01/02/2023

Graphical Solution to 2-Dimenional Programs

Suppose we only have two decision variables (𝑥1 , 𝑥2 ). Then the program can be solved
graphically by drawing the corresponding lines.

• Each linear equality in this case corresponds to a line in 2-dimensional space:

𝑐 𝑎
𝑎𝑥1 + 𝑏𝑥2 = 𝑐  𝑥2 = − 𝑥1
𝑏 𝑏

• Each line divides the spaces into two half-spaces. An inequality determines a half-
space 𝑥2

• Drawing all constraints gives the

feasible space, that is, the region 𝑎𝑥1 + 𝑏𝑥2 = 𝑐
where every point satisfies all
constraints.

• Afterwards, one must considers contours

of objective function line and optimal solutions
can be obtained by finding points lying on the 𝑥1
best objective function contour that interests
the feasible region

Brunel Business School 13

Back to Introductory Example

𝑥2

5 𝑥1 + 10 𝑥2= 400

5 𝑥1 + 10 𝑥2= 350

5 𝑥1 + 10 𝑥2= 300

5 𝑥1 + 10 𝑥2= 250

5 𝑥1 + 10 𝑥2= 200

5 𝑥1 + 10 𝑥2= 150

5 𝑥1 + 10 𝑥2= 100 𝑥2 ≥ 15

3 𝑥1 + 8 𝑥2 ≤ 240
𝑥1

Brunel Business School 14

7
 01/02/2023

Possible Cases with Optimal Solutions

No Feasible
Region
The constraints may be too limiting. If
the feasible region is empty, then the
linear program has no feasible
solution.
• In previous example, suppose we had
to produce at least 35 units of product
2.

The constraints may be too

Unbounded
permitting. In this case the linear Solution
program may be unbounded, meaning
it can be infinitely maximized or
minimized.
• In previous example, suppose the
first constraint were removed.

Brunel Business School 15

Solving Linear Programmes in General

There are many algorithms for efficiently solving linear programs:

• Simplex Method: developed by Dantzig himself, looks at corner points of
feasible regions for finding optimal solutions. This is because, optimal
solutions are always extreme (corner) points of feasible region.
• Other interior methods start from within the feasible region and find their way
to an optimal corner solution.

There are also many software tool for solving linear programs:
• Excel Solver (you need to learn this!)
• Lindo/Lingo
• IBM ILOG CPLEX
• …

Brunel Business School 16

8
 01/02/2023

Important Topics in Linear Programming

• Duality and shadow prices

• Sensitivity Analysis
• Simplex Method
• Integer Programming
• Zero-One Programming
• Mixed-Integer Programming

Brunel Business School 17

Working with Excel Solver

Brunel Business School 18

9
 01/02/2023

An Example Linear Program to Solve with Excel

Solver

max 𝑧 = 4𝑥1 +6𝑥2 +7𝑥3 +8𝑥4

s.t.

𝑥1 + 𝑥2 + 𝑥3 + 𝑥4 = 950
𝑥4 ≥ 400
2𝑥1 + 3𝑥2 + 4𝑥3 + 7𝑥4 ≤ 4600
3𝑥1 + 4𝑥2 + 5𝑥3 + 6𝑥4 ≤ 5000
𝑥1 , 𝑥2 , 𝑥3 , 𝑥4 ≥ 0

Brunel Business School 19

Thank you!

Brunel Business School 20

10