Information Technology Reference
In-Depth Information
units of measurement, the decision-making rules, and the standardisation of evaluation
results. For instance, by the type of data, decision-making methods may be classified into
three groups: 1) deterministic, 2) stochastic, and 3) fuzzy sets. There may be cases, however,
when different types of data are combined. Many authors suggest classifications which
generally differ only by the comprehensiveness and number of methods. The key difference
between classifications suggested by various authors is that some classify methods only by
the type of information about indicators, while others introduce categories of information
about alternatives (Chen & Hwang, 1991; Hwang & Yoon, 1981; Triantaphyllou, 2000).
Multiple criteria decision-making methods are most often classified into two distinct groups
with different methodology for identification of preferences and for aggregation of
information about criteria (Hwang & Yoon, 1981; Zavadskas et al., 1994). The first group
includes multiple criteria methods from the value (utility) theory based on the premise of
compensation — comparison of criteria: a possibility to fully balance the negative aspects of
one criterion with positive aspects of another . The other group includes the outranking
methods based on the concept of value without compensation and denies that criteria may
offset one another. This methods may be further classified into three subgroups: 1) selection
of the most beneficial variant using the utility function, 2) compromise models for selection
of the variant closest to ideal, 3) concordance models to determine the priority relations of
the highest compatibility (Hwang & Yoon, 1981; Zavadskas et al., 1994; Guitouni & Martel,
1998; Jeroen, 1999).
Multiple criteria analysis methods are abundant; their choice is based on the available data,
goals, desired result and participation of decision-makers in the evaluation process. We
shall proceed with a brief review of several multiple criteria methods, which are most
adequate worldwide and are best suited for environmental analysis, for evaluation of
project alternatives and technologies in the energy sector, and for integrated handling of
environmental issues.
1. Multiple criteria methods of the value (utility) theory.
This group of methods uses qualitative
input and produces quantitative output. The group has two main subgroups: analytic
hierarchy process methods and fuzzy set methods.
The
Analytic Hierarchy Process (AHP)
was
developed by the American scientist Thomas A. Saaty; lately, it is gaining popularity
worldwide and is the most frequently used method for paired comparison of indicators
(criteria, objects, features). It helps to find the weights of indicators located on the same level
of a hierarchy with respect to a higher level or weights of hierarchically unstructured
indicators. This method is based on a paired comparison matrix. Experts compare pairs of
all indicators (technologies) in question
R
R
and
ij m
,
,...,
;
here
m
is the number of
compared indicators (features).
It is a convenient method, because paired comparison of indicators is simpler than
comparison of all at once. The comparison of indicators is simple and rather reliable: it
reveals the degree to which one indicator is more important than the other. This method
enables transformation of a qualitative expert assessment of indicators into quantitative
assessment. Such comparison produces a quantum matrix
ij
. Mr
Saaty suggests evaluations using a 5-point scale (1-3-5-7-9), which is frequently used in real-
life applications (Saaty, 2000; Tam et al., 2006).
Multipurpose problems need to be separated into several components, because it helps to
simplify the problem and to structure it better. A hierarchy with different goals and/or
P
p
,
,...,
ij
Search WWH ::
Custom Search