Scribd
Upload
59 views

Study Unit 3 - Linear Programming - LP

An oil company produces two grades of gasoline by mixing oils. An LP is formulated to determine the minimum cost of production given constraints on the minimum production amounts for each grade and constraints limiting the percentage of each ingredient used in the mixtures. The decision variables represent the barrels of each oil used to produce each gasoline grade. The objective is to minimize the total cost based on barrel amounts and per barrel costs of each oil.

Uploaded by

Marthinus Marais
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
59 views

Study Unit 3 - Linear Programming - LP

An oil company produces two grades of gasoline by mixing oils. An LP is formulated to determine the minimum cost of production given constraints on the minimum production amounts for each grade and constraints limiting the percentage of each ingredient used in the mixtures. The decision variables represent the barrels of each oil used to produce each gasoline grade. The objective is to minimize the total cost based on barrel amounts and per barrel costs of each oil.

Uploaded by

Marthinus Marais
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 22

INGB 472: DECISION-SUPPORT SYSTEMS

Study Unit 3: Linear Programming (LP)


MULTI VARIATE LP’S
Solving a two-dimensional LP is easy in the sense that the feasible region can be
visualised on a x-y plane.
When solving for a three-dimensional LP the region becomes more complex,
since a z-plane is added.
What about solving for an n-dimensional LP where n≥ 4?
MULTI VARIATE LP’S
Before solving higher order LP’s it is important to know how to formulation of the
objective and constraint functions should/may look.

Let’s consider a generalised form of an LP:


LINEAR PROGRAMMING
Irrespective of the number of variables, the following must hold:

Objective function (either minimise or maximise)


Subject to constraints (the rules, the physics).

Therefore, the concept stays unchanged.


LINEAR PROGRAMMING
Let’s consider the following generalisation:

There exists an n-dimensional cost vector 𝒄𝒄 = 𝑐𝑐1 , 𝑐𝑐2 , … , 𝑐𝑐𝑛𝑛 ⊤ exerting costs (or
weights) on the n-dimensional 𝐱𝐱 = 𝑥𝑥1 , 𝑥𝑥2 , … , 𝑥𝑥𝑛𝑛 ⊤ variable vector.
The 𝑚𝑚 × 𝑛𝑛 matrix 𝑨𝑨 contains constraint coefficients, with the m-dimensional vector
𝒃𝒃 the constraint requirements.
LINEAR PROGRAMMING (MINIMIZATION)
minimise 𝑐𝑐1 𝑥𝑥1 + 𝑐𝑐2 𝑥𝑥2 + ⋯ + 𝑐𝑐𝑛𝑛 𝑥𝑥𝑛𝑛
subject to 𝑎𝑎11 𝑥𝑥1 + 𝑎𝑎12 𝑥𝑥2 + ⋯ + 𝑎𝑎1𝑛𝑛 𝑥𝑥𝑛𝑛 ≥ 𝑏𝑏1
𝑎𝑎21 𝑥𝑥1 + 𝑎𝑎22 𝑥𝑥2 + ⋯ + 𝑎𝑎2𝑛𝑛 𝑥𝑥𝑛𝑛 ≥ 𝑏𝑏2
⋮
𝑎𝑎𝑚𝑚1 𝑥𝑥1 + 𝑎𝑎𝑚𝑚𝑚 𝑥𝑥2 + ⋯ + 𝑎𝑎𝑚𝑚𝑚𝑚 𝑥𝑥𝑛𝑛 ≥ 𝑏𝑏𝑚𝑚
𝑥𝑥1 , 𝑥𝑥2 , … , 𝑥𝑥𝑛𝑛 ≥ 0.
LINEAR PROGRAMMING
In matrix form:
minimise 𝑐𝑐 𝑇𝑇 𝑥𝑥
subject to 𝐴𝐴𝐴𝐴 ≥ 𝑏𝑏
𝑥𝑥 ≥ 0
LINEAR PROGRAMMING
Note that this is a ‘neat’ generalisation and that you have been following this all
along:
LINEAR PROGRAMMING (MINIMIZATION)
minimise 𝑐𝑐1 𝑥𝑥1 + 𝑐𝑐2 𝑥𝑥2 + ⋯ + 𝑐𝑐𝑛𝑛 𝑥𝑥𝑛𝑛
subject to 𝑎𝑎11 𝑥𝑥1 + 𝑎𝑎12 𝑥𝑥2 + ⋯ + 𝑎𝑎1𝑛𝑛 𝑥𝑥𝑛𝑛 ≥ 𝑏𝑏1
𝑎𝑎21 𝑥𝑥1 + 𝑎𝑎22 𝑥𝑥2 + ⋯ + 𝑎𝑎2𝑛𝑛 𝑥𝑥𝑛𝑛 ≥ 𝑏𝑏2
⋮
𝑎𝑎𝑚𝑚1 𝑥𝑥1 + 𝑎𝑎𝑚𝑚𝑚 𝑥𝑥2 + ⋯ + 𝑎𝑎𝑚𝑚𝑚𝑚 𝑥𝑥𝑛𝑛 ≥ 𝑏𝑏𝑚𝑚
𝑥𝑥1 , 𝑥𝑥2 , … , 𝑥𝑥𝑛𝑛 ≥ 0.
LINEAR PROGRAMMING
Note that all the inequalities in the previous example were “≥”. You have to
ensure this hold in order to write out the general matrix format as provided.
Therefore if one of the inequalities was “≤”; the constraint should firstly have been
multiplied by “-1” throughout.
LINEAR PROGRAMMING
For maximisation the following must hold:
maximise 𝑐𝑐 𝑇𝑇 𝑥𝑥
subject to 𝐴𝐴𝐴𝐴 ≤ 𝑏𝑏
𝑥𝑥 ≥ 0
EXAMPLES
Now that the rules are provided, formulate the following linear problems:
EXAMPLE
Company A places advertisements. A budget of 8000 is available and the
maximum audience number must be reached.
The following information is available:
EXAMPLE

Given that at least 5 radio spots should be placed and radio advertisements may
not exceed 1800 in cost.
What is the maximum number of audience that could be reached?
EXAMPLE FORMULATION
𝑋𝑋1 = number of TV.
𝑋𝑋2 = number of newspaper.
𝑋𝑋3 = number of radio I.
𝑋𝑋4 = number of radio II.
EXAMPLE FORMULATION
max(5000𝑋𝑋1 + 8500𝑋𝑋2 + 2400𝑋𝑋3 + 2800𝑋𝑋4 )
subjected to
𝑋𝑋1 ≤ 12
𝑋𝑋2 ≤ 5
𝑋𝑋3 ≤ 25
𝑋𝑋4 ≤ 20
800𝑋𝑋1 + 925𝑋𝑋2 + 290𝑋𝑋3 + 380𝑋𝑋4 ≤ 8000
𝑋𝑋3 + 𝑋𝑋4 ≥ 5
290𝑋𝑋3 + 380𝑋𝑋4 ≤ 1800
𝑋𝑋1 , 𝑋𝑋2 , 𝑋𝑋3 , 𝑋𝑋4 ≥ 0
OIL EXAMPLE
An oil company produces 2 grades of gasoline; R (regular) and E (economy) by
mixing oils I and II.
Oil Ingredient Ingredient Cost/barrel
A% B%
I 35 55 30.00
II 60 25 34.80

A minimum of 25000 barrels of R and 32000 of E.


At least 45% of each barrel of R must contain ing A. Max 50% of E barrels from ing
B.
Set up an LP to determine the minimum cost.
OIL EXAMPLE
𝑋𝑋1 = barrels of I to produce R.
𝑋𝑋2 = barrels of I to produce E.
𝑋𝑋3 = barrels of II to produce R.
𝑋𝑋4 = barrels of II to produce E.
Determine constraints:
R must contain at least 45% of A
∴ 0.35𝑋𝑋1 + 0.6𝑋𝑋3 ≥ 0.45(𝑋𝑋1 + 𝑋𝑋3 )
∴ −0.1𝑋𝑋1 + 0.15𝑋𝑋3 ≥ 0
OIL EXAMPLE
E may contain a max of 50% from B
∴ 0.55𝑋𝑋2 + 0.25𝑋𝑋4 ≤ 0.5(𝑋𝑋2 + 𝑋𝑋4 )
∴ −0.05𝑋𝑋2 + 0.25𝑋𝑋4 ≥ 0
OIL EXAMPLE
min 30𝑋𝑋1 + 30𝑋𝑋2 + 34.80𝑋𝑋3 + 34.80𝑋𝑋4
subjected to:
𝑋𝑋1 + 𝑋𝑋3 ≥ 25000
𝑋𝑋2 + 𝑋𝑋4 ≥ 32000
−0.1𝑋𝑋1 + 0.15𝑋𝑋3 ≥ 0
−0.05𝑋𝑋2 + 0.25𝑋𝑋4 ≥ 0
𝑋𝑋1 , 𝑋𝑋2 , 𝑋𝑋3 , 𝑋𝑋4 ≥ 0
Homework
A company explores the possibility to invest in turbines for energy recovery from
excess steam. No more than two turbines may be invested in and R100 million is
available. Which turbine selections will maximise the net present value (NPV) of
the funds available, given the following information?
HOMEWORK
Turbine Cost (R Income year Income year Income year
million) I (R million) II (R million) III (R million)
I 20 5 10 11
II 30 7 16 17
III 50 10 23 26
IV 60 15 28 35
V 80 19 35 40

You might also like