Chapter 2

MULTI-CRITERIA DECISION MAKING

METHODS

2.1 BACKGROUND INFORMATION

With the continuing proliferation of decision methods and their
variants, it is important to have an understanding of their comparative value.
Each of the methods uses numeric techniques to help decision makers choose
among a discrete set of alternative decisions. This is achieved on the basis
of the impact of the alternatives on certain criteria and thereby on the overall
utility of the decision maker(s). The difficulty that always occurs when trying
to compare decision methods and choose the best one is that a paradox is
reached, i.e., What decision-making method should be used to choose the
best decision-making method? This problem is examined in Chapter 9.
Despite the criticism that multi-dimensional methods have received,
some of them are widely used. The weighted sum model (WSM) is the
earliest and probably the most widely used method. The weighted product
model (WPM) can be considered as a modification of the WSM, and has been
proposed in order to overcome some of its weaknesses. The analytic
hierarchy process (AHP), as proposed by Saaty [Saaty, 1980 and 1994], is a
later development and it has recently become increasingly popular. Professors
Belton and Gear [1983] suggested a modification to the AHP (which we will
call the revised AHP) that appears (as it is demonstrated in later chapters) to
be more consistent than the original approach. Some other widely used
methods are the ELECTRE and the TOPSIS methods.
In the section that follows these methods are presented in detail. In
Chapter 9 the methods are tested in terms of two evaluative criteria. The same
chapter also uses the test findings and examines the implication of these
findings on the effectiveness of the various decision making approaches.

2.2 DESCRIPTION OF SOME MCDM METHODS

There are three steps in utilizing any decision-making technique
involving numerical analysis of alternatives:

1) Determine the relevant criteria and alternatives.

2) Attach numerical measures to the relative importance
E. Triantaphyllou, Multi-criteria Decision Making Methods: A Comparative Study
© Springer Science+Business Media Dordrecht 2000
6 MCDM Methods: A Comparative Study, by E. Triantaphyllou

of the criteria and to the impacts of the alternatives

on these criteria.
3) Process the numerical values to determine a ranking
of each alternative.

This section is only concerned with the way the WSM, WPM, AHP,
revised AHP, ELECTRE, and TOPSIS methods process the numerical data
in step 3. The central decision problem examined in this book is described
as follows. Given is a set of m alternatives denoted as AI' A2, A3, ••• , Am and
a set of n decision criteria denoted as Cl , C2 , C3, ... , Cn. It is assumed that
the decision maker has determined (the absolute or relative) performance
value aij (for i = 1, 2, 3, ... , m andj = 1, 2, 3, ... , n) of each alternative
in terms of each criterion. That is, he/she has determined the matrix A with
the aij values, along with the criteria weights Wj (i.e., the decision matrix as
it was defined in Section 1.2). In the next chapter a number of procedures for
determining these data is discussed.
Given the aij and Wj values, then the problem examined in this chapter
is how one can rank the alternatives when all the decision criteria are
considered simultaneously. In the following sections, the criteria are assumed
to represent some kind of profit. That is, the higher the value, the better it
is. Next, a number of MCDM models for solving the above problem (i.e.,
step 3 above) is presented.

2.2.1 The WSM Method

The weighted sum model (WSM) is probably the most commonly used
approach, especially in single dimensional problems. If there are m
alternatives and n criteria then, the best alternative is the one that satisfies (in
the maximization case) the following expression [Fishburn, 1967]:
n
A~M-scor" = ~ L aijwj , for i =1, 2, 3, ... , m. (2-1)
I j = 1

where: A*WSM-score is the WSM score of the best alternative, n is the number
of decision criteria, (lij is the actual value of the i-th alternative in terms of the
j-th criterion, and Wj is the weight of importance of the j-th criterion.
The assumption that governs this model is the additive utility
assumption. That is, the total value of each alternative is equal to the sum
of the products given as (2-1). In single-dimensional cases, where all the
units are the same (e.g., dollars, feet, seconds), the WSM can be used
without difficulty. Difficulty with this method emerges when it is applied to
Chapter 2: MCDM Methods 7

multi- dimensional MCDM problems. Then, in combining different

dimensions, and consequently different units, the additive utility assumption
is violated and the result is equivalent to "adding apples and oranges".

Example 2-1:
Suppose that an MCDM problem involves four criteria, which are
expressed in exactly the same unit, and three alternatives. The relative
weights of the four criteria were determined to be: wJ = 0.20, w2 = 0.15,
W 3 = 0.40, and W 4 = 0.25. Also, the performance values of the three
alternatives in terms of the four decision criteria are assumed to be as follows:

25 20 15 30
A = 10 30 20 30
30 10 30 10
Therefore, the data for this MCDM problem are summarized in the following
decision matrix:

Criteria
C1 C2 C3 C4
Alts. ( 0.20 0.15 0.40 0.25)

A1 25 20 15 30
A2 10 30 20 30
A3 30 10 30 10

When formula (2-1) is applied on the previous data the scores of the three
alternatives are:
A J• WSM-score = 25 xO.20 + 20xO.15 + 15 x 0.40 + 30xO.25 =
= 21.50.
Similarly, we get:
A 2• WSM-score = 22.00,
and A 3. WSM-score = 20.00.
Therefore, the best alternative (in the maximization case) is alternative
A2 (because it has the highest WSM score; 22.00). Moreover, the following
ranking is derived: A2 > A J > A3 (where the symbol" > " stands for "better
than") . •
8 MCDM Methods: A Comparative Study, by E. Triantaphyllou

2.2.2 The WPM Method

The weighted product model (WPM) is very similar to the WSM.
The main difference is that instead of addition in the model there is
multiplication. Each alternative is compared with the others by multiplying
a number of ratios, one for each criterion. Each ratio is raised to the power
equivalent to the relative weight of the corresponding criterion. In general,
in order to compare two alternatives AK and AL, the following product
(Bridgman [1922] and Miller and Starr [1969]) has to be calculated:
11

R(AxIAL) ;: II (aKjlaLir J, (2-2)

i = 1

where n is the number of criteria, aij is the actual value of the i-th alternative
in terms of the j-th criterion, and Wj is the weight of importance of the j-th
criterion.
If the term R(AK / AJ is greater than or equal to one, then it indicates
that alternative AK is more desirable than alternative AL (in the maximization
case). The best alternative is the one that is better than or at least equal to all
other alternatives.
The WPM is sometimes called dimensionless analysis because its
structure eliminates any units of measure. Thus, the WPM can be used in
single- and multi-dimensional MCDM. An advantage of the method is that
instead of the actual values it can use relative ones. This is true because:
11

aKjI E
i = 1
aKj
;:
I
aKj
(2-3)
11

aIJI E
i = 1
ali
A relative value d Kj is calculated using the formula:
where the aKj's are the actual values.

Example 2-2:
Consider the problem presented in the previous Example 2-1.
However, now the restriction to express all criteria in terms of the same unit
is not needed. When the WPM is applied, then the following values are
derived:
R(A/A:zJ = (25/10)°·20 x (20/30)°·15 X (15/20)°·40 X (30/30)°·25 =
= 1.007 > 1.
Similarly, we also get:
R(A/A~ = 1.067 > 1,
Chapter 2: MCDM Methods 9

and R(A/A3J = 1.059 > 1.

Therefore, the best alternative is AI' since it is superior to all the other
alternatives. Moreover, the ranking of these alternatives is as follows: Al >
A z > A3 · •

An alternative approach with the WPM method is for the decision

maker to use only products without ratios. That is, to use the following
variant of formula (2-2);
n
P(A K } = II (aKjr J, (2-4)
j = 1

In the previous expression the term P(A k ) denotes the performance value (not
a relative one) of alternative Ak when all the criteria are considered under the
WPM model. Then, when the previous data are used, exactly the same
ranking is derived. Some interesting properties of this method are discussed
in Chapter 11, Sections 11.6 and 11.7.

2.2.3 The AHP Method

The analytic hierarchy process (AHP) ([Saaty, 1980 and 1994])
decomposes a complex MCDM problem into a system of hierarchies (more
on these hierarchies can be found in [Saaty, 1980]). The final step in the AHP
deals with the structure of an mXn matrix (where m is the number of
alternatives and n is the number of criteria). The matrix is constructed by
using the relative importances of the alternatives in terms of each criterion.
The vector (ail' aiZ' ai3 , ... , ain ) for each i is the principal eigenvector of an
n X n reciprocal matrix which is determined by pairwise comparisons of the
impact of the m alternatives on the i-th criterion (more on this, and other
related techniques, is presented in Chapter 3).
The importance of the AHP, its variants, and the use of pairwise
comparisons in decision making is best illustrated in the more than 1,000
references cited in [Saaty, 1994]. A number of special issues in refereed
journals have been devoted to the AHP and the use of pairwise comparisons
in decision making. These issues are: Socio-Economic Planning Sciences
[Vol. 10, No.6, 1986]; Mathematical Modelling [Vol. 9, No. 3-5, 1987];
European Journal of Operational Research [Vol. 48, No.1, 1990]; and
Mathematical and Computer Modelling [Vol. 17, No. 4/5, 1993]. Also, four
international symposia (called ISAHP) have been dedicated on the same topic
so far and one such event is now scheduled every two years.
Some evidence is presented in [Saaty, 1980]) (see also Chapters 3 and
10 MCDM Methods: A Comparative Study, by E. Triantaphyllou

4 in this book) which supports the technique of pairwise comparisons for

eliciting numerical evaluations of qualitative phenomena from experts and
decision makers. However, here we are not concerned with the possible
advantages and disadvantages of the pairwise comparison and eigenvector
methods for determining the aij values. Instead, we examine the method used
in the AHP to process the aij values after they have been determined. The
entry aij' in the m x n matrix, represents the relative value of alternative Ai
n
when it is considered in terms of criterion C;. In the original AHP the sum L
aij
is equal to one. i = 1
According to the AHP the best alternative (in the maximization case)
is indicated by the following relationship (2-5):
n
A;HP-score = ~x L aijwj' for i =1, 2, 3, ,." m. (2-5)
I j = 1

The similarity between the WSM and the AHP is clear. The AHP
uses relative values instead of actual ones. Thus, it can be used in single- or
multi-dimensional decision making problems.

Example 2-3:
As before, we consider the data used in the previous two examples
(note that as in the WPM case the restriction to express all criteria in terms
of the same unit is not needed). The AHP uses a series of pairwise
comparisons (more on this can be found in Chapter 3) to determine the
relative performance of each alternative in terms of each of the decision
criteria. In other words, instead of the absolute data, the AHP would use the
following relative data:

Criteria
CI C2 C3 C4
Alts. ( 0.20 0.15 OAO 0.25)

Al 25/65 20/55 15/65 30/65

A2 10/65 30/55 20/65 30/65
A3 30/65 5/55 30/65 5/65

That is, the columns in the decision matrix have been normalized to add up
to one. When formula (2-5) is applied on the previous data, the following
scores are derived:
A 1, AHP.score = (25/65)xO.20 + (20/55)xO.15 +
+ (15/65)xOAO + (30/65)xO.25 =
Chapter 2: MCDM Methods 11

0.34.
Similarly, we get:
A 2• AHP-score = 0.35,
and A 3. AHP.score 0.31.
Therefore, the best alternative (in the maximization case) is alternative A2
(because it has the highest AHP score; 0.35). Moreover, the following
ranking is derived: A2 > Al > A3 • •

2.2.4 The Revised AHP Method

Belton and Gear [1983] proposed a revised version of the original
AHP model. They demonstrated that a ranking inconsistency can occur when
the AHP is used. A numerical example was presented that consists of three
criteria and three alternatives. In that example (which is also shown next) the
indication of the best alternative changes when an identical alternative to one
of the nonoptimal alternatives is introduced now creating four alternatives.
According to Belton and Gear the root for that inconsistency is the fact that
the relative values for each criterion sum up to one. Instead of having the
relative values of the alternatives AI> A21 A31 "'1 Am sum up to one, they
proposed to divide each relative value by the maximum value of the relative
values. In particular, they elaborated on the following example.

Example 2-4: (from [Belton and Gear, 1983], p. 228)

Suppose that the actual data of a MCDM problem with three
alternatives and three criteria are as follows:

Criteria
C1 C2 C3
Alts. (113 113 113)

1 9 8
9 1 9
1 1 1

In real life problems the decision maker may never know the previous
real data. Instead, he/she can use the method of pairwise comparisons (as
described in Chapters 3 and 4) to derive the relative data. When the AHP is
applied on the previous data, the following decision matrix with the relative
data is derived:
12 MCDM Methods: A Comparative Study, by E. Triantaphyllou

Criteria
CI C2 C3
Alts. (113 113 113)

Al 1111 9/11 8/18

A2 9/11 1111 9/18
A3 1111 1111 1118

Therefore, it can be easily shown that the vector with the final AHP scores
is: (0.45, 0.47, 0.08). That is, the three alternatives are ranked as follows:
A2 > Al > A3 •
Next, we introduce a new alternative, say A 4 , which is an identical
Q!ID.'. of the existing alternative A2 (i.e., A2 == A4)' Furthermore, it is also
assumed that the relative weights of importance of the three criteria remain
the same (i.e., 113, 113, 113). When the new alternative A4 is considered,
it can be easily verified that the new decision matrix is as follows:

Criteria
CI Cz C3
Alts. ( 113 1/3 113)

Al 1/20 9/12 8/27

Az 9/20 1112 9/27
A3 1120 1112 1/27
A4 9/20 1112 9/27

Similarly as above, it can be verified that the vector with the final AHP scores
is: (0.37, 0.29, 0.06, 0.29). That is, the four alternatives are ranked as
follows: Al > A2 == A4 > A 3 • Belton and Gear claimed that this result is in
logical contradiction with the previous result (in which A2 > Aj).
When the revised AHP is applied on the last data (that is, when the
data are normalized by dividing by the largest entry in each column), the
following decision matrix is derived:

Criteria
CI C2 C3
Alts. ( 113 113 113)

Al 119 1 8/9
Az 1 119 1
A3 119 119 119
A4 1 119 1
Chapter 2: MCDM Methods 13

The vector with the final scores is: (2/3, 19/27, 1/9, 19/27). That is, the
four alternatives are ranked as follows: A2 == A4 > Al > A3 • The last
ranking is, obviously, the desired one. •

The revised AHP was sharply criticized by Saaty in [1990]. He

claimed that identical alternatives should not be considered in the decision
process. However, even earlier Triantaphyllou and Mann in [1989] had
demonstrated that similar logical contradictions are possible with the original
AHP, as well as with the revised AHP, even when non-identical alternatives
are introduced (see also Section 9.4 in Chapter 9). Some discussion on this
controversy is also presented in Chapter 11.
Two other MCDM methods are presented next. These methods are
of limited acceptance by the scientific and practitioners communities. These
are the ELECTRE and TOPSIS methods.

2.2.5 The ELECTRE Method

The ELECTRE (for Elimination and Choice Translating Reality;

English translation from the French original) method was first introduced in
[Benayoun, et al., 1966]. The basic concept of the ELECTRE method is to
deal with "outranking relations" by using pairwise comparisons among
alternatives under each one of the criteria separately. The outranking
relationship of the two alternatives Ai and Aj , denoted as A[~Aj' describes that
even when the i-th alternative does not dominate the J-th alternative
quantitatively, then the decision maker may still take the risk of regarding Ai
as almost surely better than Aj [Roy, 1973]. Alternatives are said to be
dominated, if there is another alternative which excels them in one or more
criteria and equals in the remaining criteria.
The ELECTRE method begins with pairwise comparisons of
alternatives under each criterion. Using physical or monetary values, denoted
as gJAj} and gdAJ of the alternatives Aj and Ak respectively, and by
introducing threshold levels for the difference gJAj } - gdAk }, the decision
maker may declare that he/she is indifferent between the alternatives under
consideration, that he/she has a weak or a strict preference for one of the two,
or that he/she is unable to express any of these preference relations.
Therefore, a set of binary relations of alternatives, the so-called outranking
relations, may be complete or incomplete. Next, the decision maker is
requested to assign weights or importance factors to the criteria in order to
express their relative importance.
Through the consecutive assessments of the outranking relations of the
14 MCDM Methods: A Comparative Study, by E. Triantaphyllou

alternatives, the ELECTRE method elicits the so-called concordance index,

defined as the amount of evidence to support the conclusion that alternative
Aj outranks, or dominates, alternative Ak , as well as the discordance index,
the counter-part of the concordance index.
Finally, the ELECTRE method yields a system of binary outranking
relations between the alternatives. Because this system is not necessarily
complete, the ELECTRE method is sometimes unable to identify the most
preferred alternative. It only produces a core of leading alternatives. This
method has a clearer view of alternatives by eliminating less favorable ones.
This method is especially convenient when there are decision problems that
involve a few criteria with a large number of alternatives [Lootsma, 1990].
There are many variants of the ELECTRE method. The organization
of the original version of the ELECTRE method is illustrated in the following
steps [Benayoun, et at., 1966]:

Step 1: Nonnalizing the Decision Matrix

This procedure transforms the entries of the decision matrix into
dimensionless comparable entries by using the following equation:

(2-6)

~i>~
k=l
Therefore, the normalized matrix X is defined as follows:

where m is the number of alternatives, n is the number of criteria, and xi} is

the normalized preference measure of the i-th alternative in terms of the j-th
criterion.

Step 2: Weighting the Nonnalized Decision Matrix

Next, each one of the columns of the previous X matrix is multiplied
by the associated weight of importance of the corresponding decision
criterion. These weights, denoted as (WI> W 2 , w3 , ... , wn), were determined
Chapter 2: MCDM Methods 15

by the decision maker. Therefore, the weighted matrix, denoted as Y, is:

Y = X W, or:

Yl1 Y12 Y13 YIn

Y21 Y22 Y23 Y2n

Yml Y m2 Ym3 .•• Y mn

WIX m1 W 2 X m2 W3 X m3 ... wnxmn

where:

WI 0 0 0
0 w2 0 0
n
W and LW i l.
i= I

0 0 0 ... Wn

SteR 3: Determine the Concordance and Discordance Sets

The concordance set Ckl of two alternatives Ak and AI' where m ;?;
k, I ;?; 1, is defined as the set of all the criteria for which Ak is preferred to
AI. That is, the following is true:
Ckl = {j, Ykj ;?; Yij}, for j = 1,2,3, ... , n.
The complementary subset is called the discordance set and it is described as
follows:
Dkl = {j, Ykj < Ylj}' for j = 1, 2, 3, ... , n.
16 MCDM Methods: A Comparative Study, by E. Triantaphyllou

Step 4: Construct the Concordance and Discordance Matrices

The relative value of the elements in the concordance matrix C is calculated
by means of the concordance index. The concordance index ckt is the sum
of the weights associated with the criteria contained in the concordance set.
That is, the following is true:

ekl = L Wj' for j = 1, 2, 3, ... , n.

jECk 1

The concordance index indicates the relative importance of alternative

Ak with respect to alternative At. Apparently, 0 ~ ckt ~ 1. The
concordance matrix C is defined as follows:

where the entries of matrix C are not defined when k = I.

The discordance matrix D expresses the degree that a certain
alternative Ak is worse than a competing alternative At. The elements dkt of
the discordance matrix are defined as follows:
max
JEDkl
IYkj - Ylj I
dkl = (2-7)
max
j
IYkj - Ylj I
The discordance matrix is defined as follows:

d m1 d m2 dm3 ...
as with the C matrix, the entries of matrix D are not defined when k = I.
Finally, it should be noted here that the previous two m x m matrices are
Chapter 2: MCDM Methods 17

not symmetric.

Step 5: Detennine the Concordance and Discordance Dominance

Matrices
The concordance dominance matrix is constructed by means of a
threshold value for the concordance index. For example, alternative Ak will
only have a chance to dominate alternative Al if its corresponding concordance
index Ckl exceeds at least a certain threshold value f. That is, this happens if
the following condition is true:
cld ;?! f·
The threshold value f can be determined as the average concordance
index. That is, the following relation could be true:

1 m m

" = m(m - 1)
E E
k=l 1= 1
C/d'
(2-8)
tIIt4 k., tIIt4 I.k

Based on the threshold value, the elements of the concordance dominance

matrix F are next determined as follows:
hi = 1, if CId ;?! f,
hi = 0, if cld < f·
Similarly, the discordance dominance matrix G is defined by using
a threshold value d, where d could be defined as follows:
m m
1
d =
m(m - 1) E E
k=l 1= 1
dId' (2-9)
tIIt4 k., tIIt4 I.k

and gkl = 1, if did d,

;?!

gkl = 0, if did < d·

Step 6: Determine the A22l"e2ate Dominance Matrix
The elements of the ag2l"egate dominance matrix E are next defined
as follows:
(2-10)

Step 7: Eliminate the Less Favorable Alternatives

From the aggregate dominance matrix one can derive a partial
preference ordering of the alternatives. If eld = 1, then this means that
alternative Ak is preferred to alternative Al by using both the concordance and
discordance criteria.
If any column of the aggregate dominance matrix has at least one
element equal to 1, then this column is "ELECTREally" dominated by the
corresponding row. Therefore, one can simply eliminate any column(s) which
18 MCDM Methods: A Comparative Study, by E. Triantaphyllou

have an element equal to one. Then, the best alternative is the one which
dominates all other alternatives in this manner.

2.2.6 The TOPSIS Method

TOPSIS (for the Technique for Order Preference by Similarity to

Ideal Solution) was developed by Yo on and Hwang [1980] as an alternative
to the ELECTRE method and can be considered as one of its most widely
accepted variants. The basic concept of this method is that the selected
alternative should have the shortest distance from the ideal solution and the
farthest distance from the negative-ideal solution in some geometrical sense.
The TOPSIS method assumes that each criterion has a tendency of
monotonically increasing or decreasing utility. Therefore, it is easy to define
the ideal and negative-ideal solutions. The Euclidean distance approach was
proposed to evaluate the relative closeness of the alternatives to the ideal
solution. Thus, the preference order of the alternatives can be derived by a
series of comparisons of these relative distances.
The TOPSIS method evaluates the following decision matrix which
refers to m alternatives which are evaluated in terms of n criteria:

where xij denotes the performance measure of the i-th alternative in terms
of the J-th criterion. Next, the steps of the TOPSIS method are presented.

Step 1: Construct the Nonnalized Decision Matrix

The TOPSIS method first converts the various criteria dimensions into
non-dimensional criteria as was the case with the ELECTRE method (i.e.,
relation (2-6». An element rij of the normalized decision matrix R is thus
calculated as follows:
Chapter 2: MCDM ¥ethods 19

(2-11)

~i;x~
&=1

As a remark, it should be stated here that in the ELECTRE and

TOPSIS methods the Euclidean distances defined in expressions (2-6) and (2-
11), respectively, represent some plausible assumptions. Other alternative
distance measures could be used as well, in which case it is possible for one
to get different answers for the same problem.

Step 2: Construct the Weighted Normalized Decision Matrix

A set of weights W = (wI> w2 , W 3 , ... , wn ), (where: Ew; = 1) defined
by the decision maker is next used with the decision matrix to generate the
weighted normalized matrix V as follows:

Step 3: Determine the Ideal and the Negative-Ideal Solutions

The ideal, denoted as A*, and the negative-ideal, denoted as A-,
alternatives (solutions) are defined as follows:
A" = { (max vii I jEJ), (min vii I j E JI), i = 1,2,3, ... ,m}

= { VI*, V 2*, ... , vn* }. (2-12)

A- = { (min vii I jEJ), (max vii I j E JI), i = 1,2,3, ... ,m}

= { VI_, v2_, ... , vn_ }. (2-13)

where: J = {j = 1, 2, 3, ... , n and j is associated with benefit criteria},
J I = {j = 1, 2, 3, ... , nand j is associated with cost/loss criteria}.

The previous two alternatives are fictitious. However, it is reasonable

to assume here that for the benefit criteria, the decision maker wants to have
a maximum value among the alternatives. For the cost criteria, the decision
20 MCDM Methods: A Comparative Study, by E. Triantaphyllou

maker wants to have a minimum value among the alternatives. From the
previous definitions it follows that alternative A * indicates the most preferable
alternative or the ideal solution. Similarly, alternative A- indicates the least
preferable alternative or the negative-ideal solution.

Step 4: Calculate the Separation Measure

The n-dimensional Euclidean distance method is next applied to
measure the separation distances of each alternative from the ideal solution
and negative-ideal solution. Thus, for the distances from the ideal solution
we have:

n
L (vij - Vj J2, for i = 1,2,3, ... , m, (2-14)
j = 1

where Si* is the distance (in the Euclidean sense) of each alternative from the
ideal solution.
Similarly, for the distances from the negative-ideal solution we have:

n
S.1 - L (vij - Vj _)2, for 1, 2, 3, ... , m, (2-15)
j = 1

where Si_ is the distance (in the Euclidean sense) of each alternative from the
negative-ideal solution.

Step 5: Calculate the Relative Closeness to the Ideal Solution

The relative closeness of an alternative Ai with respect to the ideal
solution A* is defined as follows:

c. =
S.'
1*
Si. + 1-

and i = 1, 2, 3, ... , m. (2-16)

Apparently, Ci* = 1, if Ai = A*, and Ci_ = 0, if Ai = A-.
Chapter 2: MCDM Methods 21

Step 6: Rank the Preference Order

The best (optimal) alternative can now be decided according to the
preference rank order of C,"*. Therefore, the best alternative is the one that
has the shortest distance to the ideal solution. The previous definition can
also be used to demonstrate that any alternative which has the shortest
distance from the ideal solution is also guaranteed to have the longest distance
from the negative-ideal solution.