Chapter 1

INTRODUCTION TO MULTI-CRITERIA
DECISION MAKING

1.1 MULTI-CRITERIA DECISION MAKING:

A GENERAL OVERVIEW

The analysis of the way people make decisions (prescriptive theories)

or the way people ought to make decisions (normative theories) is perhaps as
old as the recorded history of mankind. Of course, not all these analyses
were characterized by the rigorous scientific approaches we see in the
literature today. Therefore, it is not surprising that the literature in decision
making is humongous and continuously increasing. At the same time,
however, the development of the perfect decision making method for rational
real life decision making still remains an elusive goal. This contradiction
between the extensiveness of the study on this subject and the elusiveness of
the final goal of the real life applicability of the findings, constitutes in a way
the ultimate decision making paradox.
Multi-criteria decision making (MCDM) is one of the most well
known branches of decision making. According to many authors (see, for
instance, [Zimmermann, 1996]) MCDM is divided into multi-objective
decision making (MODM) and multi-attribute decision making (MADM).
However, very often the terms MADM and MCDM are used to mean the
same class of models (Le., MCDM).
MODM studies decision problems in which the decision space is
continuous. A typical example is mathematical programming problems with
multiple objective functions. The first reference to this problem, also known
as the "vector-maximum" problem, is attributed to Kuhn and Tucker [1951].
On the other hand, MCDM/MADM concentrates on problems with discrete
decision spaces. In these problems the set of decision alternatives has been
predetermined.
Although MCDM methods may be widely diverse, many of them have
certain aspects in common [Chen and Hwang, 1991]. These are the notions
of alternatives and attributes (also often called goals or decision criteria) as
described next.

Alternatives:
Usually alternatives represent the different choices of action available to the
decision maker. In this book the set of alternatives is assumed to be finite,

E. Triantaphyllou, Multi-criteria Decision Making Methods: A Comparative Study

Springer Science+Business Media Dordrecht 2000
2 MCDM Methods: A Comparative Study, by E. Triantaphyllou

ranging from several to hundreds. They are supposed to be screened,

prioritized, and eventually ranked.

Multiple Attributes:
Each MCDM problem is associated with multiple attributes. Attributes are
also referred to as "goals" or .,decision criteria". Attributes represent the
different dimensions from which the alternatives can be viewed.

In cases in which the number of criteria is large (e. g., more than a
dozen), criteria may be arranged in a hierarchical manner. That is, some
criteria may be major ones. Each major criterion may be associated with
several sub-criteria. Similarly, each sub-criterion may be associated with
several sub-sub-criteria and so on. Although some MCDM methods may
explicitly consider a hierarchical structure in the criteria of a decision
problem, most of them assume a single level of criteria (e.g., no hierarchies).

Conflict among Criteria:

Since different criteria represent different dimensions of the alternatives, they
may conflict with each other. For instance, cost may conflict with profit etc.
In this book no such conflict is assumed unless it is explicitly stated
otherwise.

Incommensurable Units:
Different criteria may be associated with different units of measure. For
instance, in the case of buying a used car, the criteria "cost" and "mileage"
may be measure~ in terms of dollars and thousands of miles, respectively.
It is this nature of having to consider different units which makes MCDM
problems intrinsically hard to solve.

Decision Weights:
Most of the MCDM methods require that the criteria be assigned weights of
importance. Usually, these weights are normalized to add up to one. How
these weights can be determined is described in Chapters 4 and 5.

Decision Matrix:
An MCDM problem can be easily expressed in a matrix format. A decision
matrix A is an (m X n) matrix in which element au indicates the performance
of alternative Ai when it is evaluated in terms of decision criterion C; (for i =
1,2,3, ... , m, andj = 1,2,3, ... , n). It is also assumed that the decision
maker has determined the weights of relative performance of the decision
criteria (denoted as wj ' for j = 1, 2, 3, ... , n). This information is best
summarized in Definition 1-1 (according to [Zimmermann, 1996]) and Figure
Chapter 1: Introduction to MCDM 3

1-1. Please note that in this definition the term "goals" is used instead of the
usual term "criteria".

Definition 1-1:
Let A = {Ai' for i = 1, 2, 3, ... , n} be a (finite) set of decision alternatives
and G = {gj' for j = 1, 2, 3, ... , m} a (finite) set of goals according to
which the desirability of an action is judged. Determine the optimal
alternative A with the highest degree of desirability with respect to all
relevant goals gj'

Criteria
CI C2 C3 Cn
Alts. ( WI w2 W3 wn )

Al all a l2 aJ3 a ln
A2 a21 a22 a23 a2n

Am amI am2 amJ amn

Figure 1-1: A Typical Decision Matrix.

1.2 CLASSIFICATION OF MCDM METHODS

As it was stated in the preface, there are many MCDM methods
available in the literature. Each method has its own characteristics. There
are many ways one can classify MCDM methods. One way is to classify
them according to the type of data they use. That is, we may have
deterministic, stochastic, or fuzzy MCDM methods (for an overview of
fuzzy MCDM methods see [Chen and Hwang, 1991]). However, there may
be situations which involve combinations of all the above (such as stochastic
and fuzzy) data types.
Another way of classifying MCDM methods is according to the
number of decision makers involved in the decision process. Hence, we have
single decision maker MCDM methods and group decision makers MCDM
(for more information on the later class, the reader may want to check the
journal of Group Decision Making). In this book we concentrate our attention
to single decision maker deterministic MCDM methods.
In [Chen and Hwang, 1991] deterministic -- single decision maker --
MCDM methods were also classified according to the type of information and
4 MCDM Methods: A Comparative Study, by E. Triantaphyllou

the salient features of the information. A taxonomy of a number of MCDM

methods is given in Figure 1-2 (as adopted from [Chen and Hwang, 1991]).
For a brief description of the methods mentioned in Figure 1-2, the interested
reader may want to consult with [Hwang, 1987] or [Chen and Hwang, 1991].
The weighted sum model (WSM), the analytic hierarchy process (AHP), the
revised AHP, the weighted product model (WPM), the ELECTRE, and the
TOPSIS method are described in later sections and are the ones which are the
most widely used. Finally, it should be stated here that there are many other
alternative ways for classifying MCDM methods [Chen and Hwang, 1991].
However, the previous ones are the most common schemes.

Type of Information Salient Feature Major Classes of Methods

From the Decision of Information
Maker

Dominance
No Information r---------t~ Maximin
7'-------.J Maximax

Multi-Attribute Conjunctive Method

Decision (Satistying Method)
Making
Disjunctive Method

Elimination by Aspect
Information Lexicographic Semi order
on the l...-_---'r---+i Lexicographic Method
Attributes

Weighted Sum Model

Weighted Product Model
\---+1 Analytic Hierarchy Process
l...-_---' ELECTRE
TOPSIS

Figure ]-2: A Taxonomy of MCDM Methods (according to Chen and Hwang [1991]).