Scribd
Upload
14 views16 pages

IMTC634 - Data Science - Chapter 9

Uploaded by

msmakkar.chief19
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
14 views16 pages

IMTC634 - Data Science - Chapter 9

Uploaded by

msmakkar.chief19
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
You are on page 1/ 16

Chapter 9: Optimization

Chapter Index
S. Reference Particulars Slide
No. No. From - To
1 Learning Objectives 3
2 Topic 1 Optimization Process 4-5
3 Topic 2 Linear Programming 6-9
4 Topic 3 Multi-criteria Optimization 10
5 Topic 4 Goal programming 11
6 Topic 5 Analytic Hierarchy Proces 12-14
s (AHP)
7 Let’s Sum Up 15
Learning Objectives

 Describe the optimization process

 € Explain linear programming

 € Describe multi-criteria optimization

 € Discuss goal programming

 € Elucidate analytic hierarchy process


1.Optimization Process

 Optimization refers to a technique used to make decisions in


organizations and analyze physical systems.
 In case of mathematics, an optimization problem is defined
as the problem of determining the best solution from the set
of all possible solutions.
 Optimization also helps in finding the highest or lowest value
of a function within some constraints. In other words,
optimization determines the best possible value for a
function from a given domain.
 The optimization process is generally used in the field of
physics and computers. In these fields, it is commonly known
as energy optimization.
1.Optimization Process

There are three steps involved in an optimization process:


 The first step is the creation of an appropriate model.
Modeling refers to the process that helps in determining and
stating objectives, variables and constraints of a problem in
mathematical terms. An objective measures the performance
of the system that we want to increase or decrease.
 The second step is to determine the category of the
developed model in the optimization process.
 The third and final step in optimization involves the selection
of software suitable for solving the optimization problem. The
optimization software is of two types, one is solver software
and other is modeling software.
2.Linear Programming

 Linear programming is also known as linear optimization. It is a

method used for achieving preeminent results which includes

maximum revenue and lowest investment in context of a

mathematical model.

 Linear programming optimizes linear objective function, depending

upon linear equality and constraints of linear inequality.

 A simple linear program involving two variables and six

inequalities can be represented using a two dimensional polytope.

 Linear programming models are used in various industries such as

manufacturing, energy, transportation, telecommunications, etc.


2.Linear Programming

Standard Form
 The standard form of a linear programming problem comprises three parts,
which are as follows:

1. Maximization of a linear function

2. Problem constraints

3. Non-negative variables
2.Linear Programming

Augmented Form
 Linear programming problems can also be represented in an

augmented form.
 In an augmented form, the problem is represented in a block

matrix form. This form has introduced non-negative slack


variables for replacing inequalities with equalities in the
constraints.
 The block matrix form for maximizing m is as follows:

where, y >= 0 and c >= 0.


2.Linear Programming

Integer Linear Programs


 A linear program in case of real variables is considered integral if
it has at least one integral optimal solution.
 Integral linear programs are highly significant in context of
combinatorial optimization since they provide an alternative
description of a problem.
 In combinatorial optimization, an optimal object is determined
from a given set of objects.
 Generally, in case of integer linear programming problem, we
used to minimize a linear cost function based on n-dimensional
vectors. These vectors are subjected to a group of linear equality,
inequality and integrity constraints on all or some of the variables.
3. Multi-criteria Optimization

 In multi-criteria optimization, the optimization problems involving


more than one objective function simultaneously are considered.
 Multi-criteria optimization is also known as the multi-objective
programming, vector optimization, multi-objective optimization,
multi-attribute optimization or Pareto optimization.
 Multi-criteria optimization finds its application in various fields
such as economics, engineering, finance, logistics, etc. In these
fields, optimal decisions require to be taken in case of two or more
objectives that are conflicting in nature.
 Multi-criteria optimization problems are not easier to solve as the
optimized objectives usually conflict with each other.
4.Goal Programming

 The occurrence of a situation when we have to deal with more than

one objective (goal), which conflict with each other but still we have to

reach a decision where we need to take all of these objectives in

account refers to a scenario of multi-criteria decision making.

 The methods of goal programming, however, don’t produce the

complete answer as whenever there are multiple goals competing with

each other, it is complex to conduct a purely objective analysis that

would produce an optimal solution.

 The primary concept behind goal programming is to specify numeric

goals for each objective and to devise a solution satisfying all the

given constraints along with minimizing the sum of deviation from the

specified goal.
5.Analytic Hierarchy Process (AHP)

 Analytic Hierarchy process (AHP) is a tool which is used to deal


with complex decision making in an effective and efficient way.
 AHP helps the decision maker to take better decisions by
setting desired goals and priorities.
 AHP, due to its specific design, checks the consistency in the
evaluations of the decision maker, avoiding any possibility of
biasing in the process of decision making.
 AHP reduces the complexity of decisions by reducing complex
decisions into sets of pairwise comparisons and then
synthesizes the result.
 AHP contains a set of evaluation constraints along with a set of
alternatives from which the decision is to be made.
5.Analytic Hierarchy Process (AHP)

Consistency in AHP Analysis


 Due to random prioritization of criteria by the decision maker
during the decision-making process, we can expect and allow
some inconsistency in the AHP analysis. However, it is a matter
of concern that how much inconsistency must be allowed in
order to maintain the quality of the analysis. To resolve this
problem, AHP is implemented with consistency ratio.
 Consistency ratio, in AHP, can be calculated by finding the ratio
of consistency index (CI) and the consistency index of a
random-like matrix (RI). We can interpret RI as the average CI
of 500 randomly submitted matrices. Mathematically,
CR=CI/RI.
5.Analytic Hierarchy Process (AHP)

Sensitivity Analysis
 The final priorities are driven by the weight given to a
particular criterion. However, it is important to know how the
results would be affected if the given criterion had different
priorities. The process of finding answers to such “what-if”
questions is called the sensitivity analysis.
 With sensitivity analysis performed, we can test the robustness
of our decision as well as identify the key factors that influence
the decision maker.
 Sensitivity analysis in an inevitable part of the whole decision-
making process and no final conclusion in the context of the
decision should be made without performing sensitivity
analysis.
Let’s Sum Up

 Optimization refers to a technique used to make decisions in organizations

and analyze physical systems.

 In case of mathematics, an optimization problem is defined as the problem

of determining the best solution from the set of all possible solutions.

 The optimization process is generally used in the field of physics and

computers. In these fields, it is commonly known as energy optimization.

 Linear programming is a method used for achieving preeminent results

which includes maximum revenue and lowest investment in context of a

mathematical model.

 Multi-criteria optimization is also known as the multi-objective

programming, vector optimization, multi-objective optimization, multi-

attribute optimization or Pareto optimization.


THANK YOU

You might also like