An Alternative Solution to the Analytic

Hierarchy Process
Maria Teresa Lamata*
Dep. Ciencias de la Computación e Inteligencia Artificial,
Universidad de Granada, Escuela Tecnica Superior Ingeniería Informática,
18071-Granada, Spain

Let us consider the Analytic Hierarchy Process in which the labels are structured as graded
numbers. To obtain the scoring that corresponds to the best alternative, or the ranking of the
alternatives, we need to use a total order for the graded numbers involved in the problem. In this
article, we consider a definition of such a total order, which is based upon two subjective aspects:
the degree of optimism/pessimism and the liking for risk /safety. As several operations, such as
product, quotient, and so forth, of fuzzy numbers do not preserve the triangularity, we also use
the graded numbers that are analogous to the fuzzy numbers; however, the operations with graded
numbers are carried out as a simple extension of operations with real intervals. © 2006 Wiley
Periodicals, Inc.

1. INTRODUCTION
In many situations, we use measures or quantities that are not exact but approx-
imate. In those cases, the concept of fuzzy number is more adequate than that of
real number. Since its foundation in 1965 by Zadeh,1 the Theory of Fuzzy Sets has
experienced a great growth and has obtained great importance. This fact is due,
fundamentally, to the ability of fuzzy sets to express the ambiguity and vagueness
that are inherent in the human language. Based on the usual operations with real
numbers, Zadeh’s extension principle gives rise to arithmetic with fuzzy numbers,
which is commonly used.
The Analytic Hierarchy Process ~AHP! proposed by Saaty 2 is a very popular
approach of multicriteria decision making ~MCDM! that involves qualitative data
and has been applied during the last 20 years in many situations of decision mak-
ing. The problem can be abstracted as how to derive weights, rankings, or impor-
tance for a set of activities according to their impact on the situation and the
objective of the decision to be made.
The method uses a reciprocal matrix of decisions obtained by pair compari-
sons such that the information is given in a linguistic form. The eigenvector is

*e-mail: mtl@decsai.ugr.es.

INTERNATIONAL JOURNAL OF INTELLIGENT SYSTEMS, VOL. 21, 425–441 ~2006!

© 2006 Wiley Periodicals, Inc. Published online in Wiley InterScience
~www.interscience.wiley.com!. • DOI 10.1002/int.20130
426 LAMATA
used in order to obtain the importance of the criteria/alternatives, and the additive
weightings method is used to calculate the utility for each alternative across all
criteria.
In the literature there are many authors who treat the AHP from the point of
view of fuzzy numbers, but we think that the operations are only approximate
because the composition of two fuzzy numbers is not surely a new fuzzy number.
For this, in this article, we consider the tools that are needed to solve the AHP in
which the preferences are expressed by means of linguistic labels, whose values
are stated as Zadeh’s graded numbers. These numbers are analogous to the fami-
lies of a-cuts of Zadeh’s fuzzy numbers. More specifically, each Zadeh’s graded
number is defined as a nonincreasing family of real bounded intervals, indexed by
the unit interval.3–5
Let us remark that the family of a-cuts of each Zadeh’s fuzzy number neces-
sarily satisfies the conditions imposed by the Representation Theorem of Negoita
and Ralescu,6 whereas these conditions are not necessarily satisfied by the family
of intervals that define each Zadeh’s graded number.
This difference is advantageous for the graded numbers, whose arithmetic is
a simple extension of the Interval Analysis,7,8 whereas the fuzzy numbers need
additional conditions in order to be operated via their a-cuts. Under appropriate
conditions, the properties of the fuzzy numbers can be transferred to the graded
numbers and vice versa.9,10 Thus, we can obtain analogous results when the rewards
in a decision problem are defined as Zadeh’s fuzzy numbers as well as when they
are defined as Zadeh’s graded numbers. The use of the latter is therefore because
of their simplicity. This fact was already suggested some time ago.5
Here, we continue such ideas, making now a more careful study of the rank-
ing of Zadeh’s graded numbers, because this is an essential aspect to solving the
Saaty decision problem.
To apply these methods, we need to operate and rank the graded numbers
associated with each preference. However, as we have already said for the fuzzy
numbers, the arithmetic results as a natural extension of the operations with real
numbers, whereas the situation is different for the ranking. In addition, the order
for the rewards implies an order for their corresponding alternatives and we must
finally choose “the best” alternative.
Therefore, we must use a total order for the graded numbers given as priori-
ties. The definition of such an order is the main aspect of this article. We shall
define a total order for certain graded numbers, which depends on two subjective
parameters: the “degree of optimism/pessimism” and the “liking for risk /safety.”
The former is commonly used in Decision Theory; here we use it following the
methods given by González and de Campos to partially rank fuzzy numbers.11,12
The latter parameter is introduced here as a method to finally obtain the total order
that we require.
The article is divided into six sections. First, we present the general approach
of Saaty and the principal estimators of the weights. In the third section, we briefly
review the fundamental aspects concerning Zadeh’s graded numbers, considering
only those operations that are needed to solve Saaty’s decision problem consid-
ered earlier. In the following section we consider certain orders for real intervals
International Journal of Intelligent Systems DOI 10.1002/int
 ALTERNATIVE ANALYTIC HIERARCHY PROCESS SOLUTION 427
and for graded numbers. In Section 5, we present a general problem with a graded
number, the Saaty approach being a particular case. Finally, we end with the
conclusions.

2. THE ANALYTIC HIERARCHY PROCESS:

THE MODEL OF SAATY

The method of pairwise comparisons was introduced by Fechner 13 in 1860

and worked out by Thurstone 14 in 1927. In the pairwise comparison method, stim-
uli in the field of psychophysics, or criteria and alternatives in the field of decision
making, are presented in pairs to one or more referees ~e.g., experts or decision
makers!.
Alternatives represent different choices of action available to the decision
maker, the choice of alternatives usually being assumed to be finite. It is necessary
to evaluate individual alternatives, deriving weights for the criteria, construct the
overall rating of the alternatives, and identify the best alternative.
Based in pairwise comparison, Saaty’s AHP 2,15 is a method widely used for
multicriteria decision making. It provides a means of decomposing the general
method into a hierarchy of subproblems that are easier to evaluate.
Saaty describes the seven pillars of the AHP as follows:

• ratio scales, proportionally and normalized ratio scales

• reciprocal paired comparisons
• the sensitivity of the principal eigenvector
• clustering and using pivots to extend the scale
• synthesis to create a one-dimensional ratio for representing the overall outcome
• rank preservation and reversal
• integrating group judgments

The AHP developed by Saaty is a widely used multicriteria decision method

for multiple alternative decisions. The AHP uses hierarchic or network structures
to represent a decision problem. At each level of the hierarchy, pairwise compari-
sons of decision elements ~either criteria or alternatives! are used to arrive at pri-
ority scores of the elements under consideration.
Suppose an expert elicitation of preference intensity of element A i over ele-
ment A j is a ij . For quantifying these values a ij Saaty gives the scale shown in
Table I.
Let us denote the alternatives by $ A 1 , A 2 , . . . , A n % ~n is the number of com-
pared alternatives!, their actual weights by $w1 , w2 , . . . , wn %, and the matrix of
the ratios of all weights by W ⫽ @wi /wj # . The matrix of pairwise comparisons
A ⫽ @a ij # represents the intensities of the expert’s preference between individual
pairs of alternatives ~A i vs. A j , for all i, j ⫽ 1,2, . . . , n! chosen usually from a given
scale. When there are n alternatives $ A 1 , A 2 , . . . , A n %, a decision maker compares a
pair of alternatives for all possible pairs, n~n ⫺ 1!/2, then a comparison matrix A
is obtained, where the element a ij shows the preference weight of A i obtained by
comparing with A j .
International Journal of Intelligent Systems DOI 10.1002/int
428 LAMATA
Table I. Values for Saaty’s Analytic Hierarchy Process.
Definition Scale Explanation

A i and A j are equally important.
A i is moderately more important than A j .
A i is strongly more important than A j .
A i is very strongly more important than A j .
A i is extremely more important than A j .
The scales 2, 4, 6, and 8 are also used and
represent compromises among the
tabulated scale.
1 Two elements contributed equally to the property.
3 Experience and judgment slightly favor one
element over another.
5 Experience and judgment strongly favor one
element over another.
7 An element is strongly favored and its dominance
is demonstrated in practice.
9 The evidence favoring one element over the other
is of the highest possible order of affirmation.
2, 4, Compromise is needed between two adjacent
6, 8 judgments.

冢 冣
1 a 12 J a 1j J a 1n
1/a 12 1 J a 2j J a 2n
.
A⫽
1/a 1j 1/a 2j J a ij J a in
.
1/a 1n 1/a 2n J 1/a in J 1
The elements a ij are considered to be estimates of the ratios wi /wj , where w is the
vector of actual weights of the alternative, which is what we want to find. All the
ratios are positive and satisfy the reciprocity property: a ij ⫽ 1/a ji ∀i, j ⫽ 1,2, . . . , n.
Saaty’s eigenvector solution of Aw ⫽ l max w always exists if the consistency ~or
transitivity! condition a ij * a jk ⫽ a ik ~i, j, k ⫽ 1,2, . . . , n! is satisfied.
Then, from a theoretical point of view, the AHP is based on the following
four axioms:

~1! The decision maker can provide paired comparison a ij of two alternatives i and j
corresponding to a criterion/subcriterion on a ratio scale that is reciprocal, that is,
a ij ⫽ 1/a ji ∀i, j ⫽ 1,2, . . . , n.
~2! The decision maker never judges one alternative to be infinitely better than another
corresponding to a criterion, that is, a ij ⫽ ` ∀i, j ⫽ 1,2, . . . , n.
~3! The decision problem can be formulated as a hierarchy.
~4! All criteria/subcriteria that have some impact on the given problem, and all the rele-
vant alternatives, are represented in the hierarchy in one go.

A number of different methods have been recommended for the translation of

inconsistent judgments arranged in the pairwise comparison matrix A into a numer-
ical scale. The common feature of all the methods is that for a positive reciprocal
matrix A ⫽ @a ij # a vector w ⫽ $w1 , w2 , . . . , wn % is determined such that the matrix of
ratios @wi /wj # is a close approximation to A according to some metric.
International Journal of Intelligent Systems DOI 10.1002/int
 ALTERNATIVE ANALYTIC HIERARCHY PROCESS SOLUTION 429
The following methods can be used for finding the vector of weights w:

• In the eigenvector method 2 the vector of weights is an eigenvector corresponding to the

maximum eigenvalue l max of the matrix A. According to the Perron-Frobenius Theo-
rem, the eigenvalue l max is positive and real. Furthermore, the vector w can be chosen
with all positive coordinates. It is a normalized solution of the following equation:

Aw ⫽ l max w

• The least-squares method ~LSM !,16 which minimizes the objective function:

((
i⫽1 j⫽1
n

冉 冊 a ij ⫺
wi
wj
2

~1!

subject to
n

( wi ⫽ 1,
i⫽1
wi ⬎ 0, i ⫽ 1,2, . . . , n

But, this nonlinear optimization problem has multiple solutions. To eliminate this prob-
lem Chu et al.16 modify the objective function ~1! in the form

min ( ( ~wi ⫺ wj a ij ! 2
i j

• The geometric means method ~also known as the logarithmic least-squares method
~LLSM !!, where the approximating vector w has elements of the form

wi ⫽ 冉) 冊
i⫽1
n
a ij
1/n
, i ⫽ 1,2, . . . , n

The vector w is usually normalized so that the sum of the elements is one. It can be
proved that the vector w with elements defined by the above equations is the solution of
the problem of minimizing the sum of squares
n n

( ( ~log a ij ⫺ log wi ⫹ log wj ! 2

i⫽1 j⫽1
~2!

The associated set of normal equations can be written as

n n

( log a ij ⫽ n log vi ⫺ j⫽1

j⫽1
( log vj , i ⫽ 1,2, . . . , n ~3!

A solution to log vi is obtained by solving ~2! with an additive degree of freedom. A

particular solution is obtained by setting ( log vj ⫽ 0. An unnormalized solution to the
logarithmic least squares problem in ~1! can be explicitly written as

vi ⫽ exp 冋 1
n
n

( log a ij
j⫽1
册 n
⫽ ) a ij1/n
j⫽1
~4!

International Journal of Intelligent Systems DOI 10.1002/int

430 LAMATA
This means an estimate vi of the priority of A i is found by taking the geometric
mean of the a ij over all j ⫽ 1,2, . . . , n. These vi s can be normalized; however, for the
purpose of establishing a statistically significant rank order of alternatives, the normal-
ization is not necessary.
• The simple additive weighting method ~SAWM! is used to calculate the priorities for
each alternative across all criteria:
n

( a ij
j⫽1
wi ⫽ n n

( ( a ij
i⫽1 j⫽1

3. SOME BASIC ASPECTS OF GRADED NUMBERS

IN RELATION WITH FUZZY NUMBERS
The problem of ordering fuzzy quantities has been related by many research-
ers, with more than 35 studies in the literature. Based on the usual operations with
real numbers, Zadeh’s extension principle gives rise to arithmetic with fuzzy num-
bers, which is commonly used. Unlike, in the case of real numbers, fuzzy quanti-
ties have no natural order.
Nevertheless, the usual order of the real line R is not extended in a natural
way to the set of fuzzy numbers.17–25 These methods can be divided into two cat-
egories. On the one hand, some methods maintain the vagueness that is inherent in
the concept of a fuzzy number, thus consisting of a fuzzy order relation. On the
other hand, we have methods that give a crisp order for the fuzzy numbers.
Fuzzy decision problems require the processing of fuzzy variables and num-
bers that are representing vague or imprecise information. Many methods have
been introduced in the literature to compute information of fuzzy variables. Based
on Zadeh’s extension principle,1 several authors 26–29 suggest several algorithms to
complete the fuzzy weighted average.
The algorithms are based on the a-cut representation of fuzzy sets, nonlinear
programming implementation of the extension principle, and interval analysis. The
computation times of these algorithms are between the exponential order 2 2n for
the fuzzy weighted average ~FWA! of Dong and Wong.26 Liou and Wang 29 sug-
gested an improved fuzzy weighted average algorithm ~IFWA! to simplify the for-
mer one; the computation time is ~2 ⫹ n~n ⫹ 1!!, and logarithm time O~n{log n!
for efficient fuzzy weighted average algorithm ~EFWA! that reduces the compar-
ison and arithmetic operations of previous works.28 The computational require-
ments for the proposal of Guh et al.27 are proportional to the size n and the total
computational requirements are 2~n ⫹ 1!.
On the other hand, it is well known that the multiplication and other related
operation of two fuzzy numbers is not a new fuzzy number.30 So, if we consider
the fuzzy number related to Figure 1, for the increasing part of fuzzy number and
w 僆 @0,1#,
T1 ° T2 ⫽ @A 1 ⫹ a~M1 ⫺ A 1 !#{@A 2 ⫹ a~M2 ⫺ A 2 !#
⫽ A 1{A 2 ⫹ aA 1 ~M2 ⫺ A 2 ! ⫹ aA 2 ~M1 ⫺ A 1 ! ⫹ a 2 A 2 ~M1 ⫺ A 1 !~M2 ⫺ A 2 !
International Journal of Intelligent Systems DOI 10.1002/int
 ALTERNATIVE ANALYTIC HIERARCHY PROCESS SOLUTION 431

Figure 1. Triangular graded number.

it is obvious that the product of two triangular fuzzy numbers is not a triangular
fuzzy number, but verifies this result for a graded number.
Zadeh’s graded numbers are defined as nonincreasing families of bounded
real intervals, indexed by the unit interval @0,1#. Its operations are defined as a
natural extension of the interval operations. So, analogous results to those, which
are commonly used for the families of a-cuts of fuzzy numbers, are easily obtained
for the graded numbers.
A detailed study of such definitions and results appears in Refs. 3 and 4. Here,
we only consider the definition of Zadeh’s graded number ~using compact inter-
vals! and the operations, which are used to solve an AHP decision problem.
A Zadeh’s graded number is defined as any mapping c : @0,1# r $@A, B# : A ⱕ
B 僆 R% that assigns to each a 僆 @0, 1# the interval @a~a!, b~a!# such that ∀a, b 僆
@0,1# , @a ⬍ b ] a~a! ⱕ a~ b! ⱕ b~ b! ⱕ b~a!# .
Hereafter, we shall only use this kind of number, which will be called graded
numbers for brevity.
Obviously, each graded number is determined by two functions a, b : @0,1# r R
that satisfy the following three conditions:

• The function a~a! is nondecreasing.

• The function b~a! is nonincreasing.
• a~1! ⱕ b~1!.

The set of graded numbers is denoted by GZ ~R!. Obviously this set extends
the real line R, because each real number P can be identified with the graded num-
ber given by the constant function 僆 R a~a! ⫽ b~a! ⫽ P, ∀a 僆 @0,1# .
In the whole set GZ ~R!, the functions a~a! and b~a! that represent each graded
number are too arbitrary. To consider graded numbers that will be really used as
linguistic terms in AHP, we shall restrict ourselves to small subsets of GZ ~R!.
For convenience, we use in this article specific kinds of graded numbers: the
“triangular graded numbers.” This kind of number is well used as membership
International Journal of Intelligent Systems DOI 10.1002/int
432 LAMATA
functions associated with linguistic labels. More precisely, we have the following
definition.

Definition 1. We give the terms:

• Triangular graded number to any graded number determined by two linear functions
a, b : @0,1# r R, whose graphs describe a triangle. More precisely, for any three real
numbers A ⱕ M ⱕ B, we have the triangular graded number @a~a!{b~a!# determined by
the following functions:

a~a! ⫽ A ⫹ a~M ⫺ A! b~a! ⫽ B ⫺ a~B ⫺ M !

Let us consider the label that corresponds to the expression “A i is moderately

more important than A j ” with associated value 3. The corresponding triangular
graded number, if the label is symmetric with amplitude 1, is the following:
a~a! ⫽ 2 ⫹ a b~a! ⫽ 4 ⫺ a

Remark. Nevertheless, it is possible to consider more general kinds of Zadeh’s

graded numbers. For example:

• The L-R-graded numbers, fixing some functions L and R as a basis for obtaining the
function a~a! and b~a!, respectively. The triangular graded numbers would result as
the particular case where a~a! and b~a! are linear, a~0! ⫽ A, a~1! ⫽ b~1! ⫽ M, and
b~0! ⫽ B. Obviously, this kind of graded number would be analogous to the known
L-R-fuzzy numbers defined by Dubois and Prade.31
• 3-graded number to Zadeh’s graded number determined by two step functions a, b:
I r R, which are constant on the intervals @0, 13_ #, ~ 13_ , 23_ #, and ~ 23_ , 1#. Thus, for any real
numbers A ⱕ A1 ⱕ A2 ⱕ B 2 ⱕ B 1 ⱕ B, we have the 3-graded number determined by the
functions

a~a! ⫽ A, b~a! ⫽ B, if 0 ⱕ a ⬍ 1
_
3

a~a! ⫽ A1, b~a! ⫽ B 1, if 1

_
3 ⱕaⱕ 2
_
3

a~a! ⫽ A2, b~a! ⫽ B 2, if 2

_
3 ⬍ a ⱕ1

Zadeh’s graded numbers offer a direct generalization of the concept of com-

pact real interval. Thus, the arithmetic with graded numbers is a direct generaliza-
tion of the arithmetic used in the Interval Analysis.

Operations. For any positive graded numbers c~a! ⫽ @~a!, b~a!# , c1 ~a! ⫽
@a 1 ~a!, b1 ~a!# , and c2 ~a! ⫽ @a 2 ~a!, b2 ~a!# and any real number P, we are inter-
ested here in the operations related to AHP problems whose values are between
1/9 and 9, and which are defined as follows:

~a! Inverse:

~c!⫺1 ~a! ⫽ $1/x : x 僆 c~a!% ⫽ @1/a 2 ~a!,1/a 1 ~a!#

International Journal of Intelligent Systems DOI 10.1002/int

 ALTERNATIVE ANALYTIC HIERARCHY PROCESS SOLUTION 433

Figure 2. Intervals associated with the partial order ⱕ.

~b! Addition ~⫹!:

~c1 ⫹ c2 !~a! ⫽ $x ⫹ y : x 僆 c1 ~a!, y 僆 c2 ~a!% ⫽ @a 1 ~a! ⫹ a 2 ~a!, b1 ~a! ⫹ b2 ~a!#

~c! Product ~•!:

~c1 ° c2 !~a! ⫽ $x ° y : x 僆 c1 ~a!, y 僆 c2 ~a!% ⫽ @a 1 ~a! ° a 2 ~a!, b1 ~a! ° b2 ~a!#

~d! Quotient ~/!:

~c1 /c2 !~a! ⫽ $x/y : x 僆 c1 ~a!, y 僆 c2 ~a!% ⫽ @a 1 ~a!/b2 ~a!, b1 ~a!/a 2 ~a!#

~e! Root

~c! 1/n ~a! ⫽ $x 1/n : x 僆 c~a!% ⫽ @a 1 ~a! 1/n, a 2 ~a! 1/n #

In the same way, the problem of ranking Zadeh’s graded numbers is similar to
the problem of ranking intervals. For this reason, based upon an order between
intervals @A, B# ~with ∀A ⱕ B 僆 R! we shall consider finally, via the integration
on I, an order between graded numbers.
We must extend the usual order of the real line R. Therefore, it is fully rea-
sonable to take I1 ⫽ @A 1 , B1 # ⱕ @A 2 , B2 # ⫽ I2 when A 1 ⱕ A 2 and B1 ⱕ B2 . ~See
Figure 2.!
Nevertheless, the situation is not so obvious when one interval is included in
the other one; in this case, A 1 ⬍ A 2 ⱕ B2 ⬍ B1 . ~See Figure 3.!

Definition 2. In GZ ~R!, we define the following partial order: For any c1 , c2 僆

GZ ~R! given by c1 ~a! ⫽ @a i ~a!, bi ~a!# ~i ⫽ 1,2! we say that c1 ⱕ c2 when
∀a 僆 I a 1 ~a! ⱕ a 2 ~a! and b1 ~a! ⱕ b2 ~a!

Figure 3. Intervals associated with I2 傺 I1 and the total orders ⱕ0 ,ⱕ1 .

International Journal of Intelligent Systems DOI 10.1002/int
434 LAMATA
Considering the intervals corresponding to each level a, it is immediate to check
that this definition extends the usual order in R and that it is compatible with the
usual operations considered above.

4. RANKING OF REAL INTERVALS AND GRADED NUMBERS

4.1. Ranking of Real Intervals
In this section we begin considering several orders in the set C ⫽ $@A, B# : A ⱕ
B 僆 R% of real compact intervals. Later, we extend those orders to a subset of
GZ ~R!.
According to the subject of our study, we consider that each interval repre-
sents an imprecise determination of a linguistic term. We also consider the degree
of optimism, which is a subjective value given by the parameter l 僆 @0, 1#.
Associated with each value l and each interval I ⫽ @A, B# , we have the number
M~l! ⫽ lB ⫹ ~1 ⫺ l!A ~a precise but subjective estimation of the label!. A pessi-
mistic ~resp. optimistic! decision maker acts in accordance with the value l ⫽ 0
~resp. l ⫽ 1!. A half-way point of view will tend to compare the mean values
~A 1 ⫹ B1 !/2 and ~A 2 ⫹ B2 !/2. In this case l ⫽ _12 . First, he or she ranks the alterna-
tives looking at the value M~0! ~I ! ⫽ A ~resp. M~1! ~I ! ⫽ B!. In this way,
Ml : R 2 r R
I ⫽ @A, B# r Ml

Definition 3. For each l 僆 ~0,1!, we have the order ⱕl defined in C as follows:

@A 1 , B1 # ⱕl @A 2 , B2 # ? @A 1 , B1 # ⱕ @A 2 , B2 # or
lB1 ⫹ ~1 ⫺ l!A 1 ⬍ lB2 ⫹ ~1 ⫺ l!A 2

Moreover, for l ⫽ 0 and for l ⫽ 1, it is a total order. Indeed, for l ⫽ 0, the

case A 1 ⬍ A 2 ~resp. A 2 ⬍ A 1 ! gives rise to @A 1 , B1 # ⱕ0 @A 2 , B2 # ~resp. @A 2 , B2 # ⱕ0
@A 1 , B1 #!, whereas the case A 1 ⫽ A 2 originates the comparison through the order ⱕ
~depending only on the values B1 and B2 !. Analogously, any two intervals can be
compared for l ⫽ 1. Let us remark that ⱕ0 and ⱕ1 are lexicographic orders.

Definition 4. For each l ⫽ 0 we have the order ⱕl defined in C as follows:

@A 1 , B1 # ⱕ0 @A 2 , B2 # ? A 1 ⬍ A 2 or @A 1 ⫽ A 2 and B1 ⱕ B2 #

Definition 5. For each l ⫽ 1, we have the order ⱕl defined in C as follows:

@A 1 , B1 # ⱕ1 @A 2 , B2 # ? B1 ⬍ B2 or @B1 ⫽ B2 and A 1 ⱕ A 2 #

~See Figure 4.!

In the case of two alternatives with the same value M~l! ~Ii !, they are ranked
according to the other extreme: the interval with the greatest B ~resp. with the
greatest A! is preferred.
International Journal of Intelligent Systems DOI 10.1002/int
 ALTERNATIVE ANALYTIC HIERARCHY PROCESS SOLUTION 435

Figure 4. Representation of the different regions associated with the different orders.

Let us note that this is a lexicographic order; however, the second part is
equivalent to the following: The interval with the greatest amplitude B ⫺ A ~resp.
with the lowest amplitude B ⫺ A! is preferred.
A similar criterion must be expected for values of l near zero ~resp. one!.
Nevertheless, for other values of l, the preference about the greatest or the lowest
value of the amplitude B ⫺ A can be considered as another subjective parameter,
which enables us to rank different intervals with the same value of M~l! ~I!. A
broad interval represents a high variation in the set of possible rewards. Con-
versely, a narrow interval represents a low variation of the possible rewards, whose
values may be obtained with a higher degree of certainty.
For this reason, we think that the preference for a broad interval is associated
with a “risky” or “gambler” individual, whereas the preference for a narrow inter-
val is associated with a “safe” one. All this leads to the following definition.

Definition 6. Let us consider the parameter l 僆 @0,1#, called degree of

optimism, and the parameter k 僆 $r, s%, which means the “liking for risk” ~k ⫽ r!
or “liking for safety” ~k ⫽ s!. Associated with each value of l we define the fol-
lowing two total orders in the set C (where we consider arbitrary elements Ii ⫽
@A i , Bi #; i ⫽ 1,2 and the corresponding values M~l! ~Ii ! ⫽ lBi ⫹ ~1 ⫺ l!A i ):

_I1 ⱕlr I2 ? M~l! ~I1 ! ⬍ M~l! ~I2 ! or @Ml ~I1 ! ⫽ Ml ~I2 ! and B2 ⫺ A 2 ⱖ B1 ⫺ A 1 #

_I1 ⱕls I2 ? M~l! ~I1 ! ⬍ M~l! ~I2 ! or @Ml ~I1 ! ⫽ Ml ~I2 ! and B2 ⫺ A 2 ⱕ B1 ⫺ A 1 #

Because the safety is associated with a more extensive interval whereas the
risk is associated with a smaller interval,

• If l ⫽ 0 ~or l ⬇ 0!, then we assume that k ⫽ r.

• If l ⫽ 1 ~or l ⬇ 1!, then we assume that k ⫽ s.
• For the remaining values of l, we consider the two possible values of k.

International Journal of Intelligent Systems DOI 10.1002/int

436 LAMATA
The reflexive, antisymmetric, transitive, and convex properties ~which guar-
antee that ⱕlr and ⱕls are indeed total orders in C! are easily checked. We can
represent each @x, y# 僆 C by the point ~x, y! belonging to the half-plane y ⱖ x, thus
visualizing the orders defined above.

• For l ⫽ 0 we first rank vertical half-lines ~preferring the right lines to the left ones! and
second, inside each half-line, we rank the points ~preferring the high points to the low
ones!.
• For l ⫽ 1 we first rank horizontal half-lines ~preferring the high lines to the low ones!
and second, inside each half-line, we rank the points ~preferring the right points to the
left ones!.
• For 0 ⬍⬍ l ⬍⬍ 1, we first rank the parallel half-lines ~1 ⫺ l!x ⫹ ly ⫽ m ~according to the
values of m! and second, we rank the points inside each half-line ~which corresponds to
different intervals with the same value of Ml !. The two values of k give the two ways to
rank the points.

To apply Definition 6, in the case of graded numbers, we first summarize

each graded number c~a! ⫽ @a~a!, b~a!# into an interval, using the mean values
obtained integrating the monotonic functions a~a! and b~a!.

4.2. Ranking of Graded Numbers

Definition 7. We consider the mapping GZ ~R! r C, which maps each graded

number c~a! ⫽ @a~a!, b~a!# into the interval @A~c!, B~c!# defined as:

A~c! ⫽ 冕 0
1
a~a!d~a! B~c! ⫽ 冕
0
1
b~a!d~a!

In addition, and taking into account the last paragraph, for each l 僆 @0,1#, we
consider the value Ml ~c! :⫽ lB~c! ⫹ ~1 ⫺ l!A~c!.
First, we can partially rank the graded numbers, ranking its associated values
Ml ~c!. This would be analogous to a method given by de Campos and González 11
and González 12 to rank fuzzy numbers, which generalizes Yager’s 25 method ~cor-
responding to the case l ⫽ _12 !. Let us remark that Gonzalez uses the Stieltjes inte-
gral, with respect to any additive measure defined on @0,1#.
Second, we must rank the different graded numbers with the same Ml ~c!. If
they have associated different intervals @A, B# , then we rank these intervals using
Definition 6. This gives rise to the following partial order in GZ ~R!.

Definition 8. For each l 僆 @0,1# and each k 僆 $r, s% (assuming that k ⫽ r for
l ⬇ 0 and k ⫽ s for l ⬇ 1), we define the following partial order ⱕlk in GZ ~R!:

• If @A~c!, B~c!# ⫽ @A~f!, B~f!# , then c ⱕlk f is equivalent to @A~c!, B~c!# ⱕlk
@A~f!, B~f!#.
• If @A~c!, B~c!# ⫽ @A~f!, B~f!#, then the graded numbers c and f are not to be compared.

International Journal of Intelligent Systems DOI 10.1002/int

 ALTERNATIVE ANALYTIC HIERARCHY PROCESS SOLUTION 437
We are interested in extending this partial order to a total order. Thus, we
must rank the different graded numbers with the same Ml ~c!. If they have associ-
ated different intervals @A, B# , then we rank these intervals using Definition 2.

Definition 9. For each l 僆 @0,1# and each k 僆 $r, s% (assuming that ak exists,
k ⫽ r for l ⬇ 0 and k ⫽ s for l ⬇ 1!, we define the following total order ⱕlk :

• If @A~c!, B~c!# ⫽ @A~f!, B~f!#, then we use these intervals to rank the graded numbers
c and f, according to the total order ⱕlk given in Definition 2.
• If @A~c!, B~c!# ⫽ @A~f!, B~f!#, then we use the intervals c~ak ! and f~ak ! to rank c
and f (using also the total order ⱕlk given in Definition 2).

5. CASE STUDY

5.1. Energy Allocation

We propose this example for simplicity. In this example we are concerned

with finding allocation weights for several large users of energy according to their
overall contribution to different objectives of society. The representation is shown
in Figure 5.
There are three large users of energy in the United States: C1 ⫽ household
users, C2 ⫽ transportation, and C3 ⫽ power generating plants. These comprise the
third or lower level of the hierarchy.
The objectives against which these energy users will be evaluated are: contri-
bution to economic growth ~EG!, contribution to environmental quality ~EQ!, and
contribution to national security ~NS!, which comprise the second level of the
hierarchy. We construct the figure according Saaty’s solution, which, in our case,
corresponds to the second component of the TFN. ~See Figure 5.!

Figure 5. Hierarchical representation with values for the energy allocation problem.
International Journal of Intelligent Systems DOI 10.1002/int
438 LAMATA
Table II. Pairwise comparison matrix for the three criteria.
EG EQ NS LLMS

EG @1, 1# @4.5, 5.5# @2.5, 3.5# @0.528, 0.783#

EQ @11/60, 9/40# @1, 1# @5/12, 3/4# @0.100, 1.161#
NS @7/24, 5/12# @1.5, 2.5# @1, 1# @0.179, 0.296#

5.2. Social and Political Advantage

We construct the pairwise comparison matrix of these three objectives
~Table II! according to their impact on the overall objective of social and political
advantage.
Now the decision maker, after a thorough study, has also made the following
assessment of the relative importance of each user from the standpoint of the econ-
omy, the environment, and national security ~the second hierarchy level!.
The matrix that collects these judgments is given in Table III.
C1 ⫽ @C1A , C1B # ⫽ ~ @0.506, 0.800# ⫹ @0.465, 0.726# ⫹ @0.415, 0.740# !

* ~ @0.528, 0.783# ! ⫽ @0.388, 0.962#

C2 ⫽ @C2A , C2B # ⫽ ~ @0.180, 0.305# ⫹ @0.269, 0.438# ⫹ @0.228, 0.329# !

* ~ @0.100, 0.161# ! ⫽ @0.163, 0.407#

C3 ⫽ @C3A , C3B # ⫽ ~ @0.100, 0.167# ⫹ @0.063, 0.089# ⫹ @0.101, 0.121# !

* ~ @0.179, 0.296# ! ⫽ @0.077, 0.181#

~See Table IV.!

Table III. Pairwise comparison matrix for the three alternatives related to the criteria.
C1 C2 C3 ~LLMS!

EG
C1 @1, 1# @2.5, 3.5# @4.5, 5.5# @0.506, 0.800#
C2 @7/24, 5/12# @1, 1# @1.5, 2.5# @0.180, 0.305#
C3 @11/60, 9/40# @5/12, 3/4# @1, 1# @0.100, 0.167#
EQ
C1 @1, 1# @1.5, 2.5# @6.5, 7.5# @0.465, 0.726#
C2 @5/12, 3/4# @1, 1# @4.5, 5.5# @0.269, 0.438#
C3 @15/112, 13/84# @11/60, 9/40# @1, 1# @0.063, 0.089#
NS
C1 @1, 1# @1.5, 2.5# @2.5, 3.5# @0.415, 0.740#
C2 @5/12, 3/4# @1, 1# @1.5, 2.5# @0.228, 0.329#
C3 @7/24, 5/12# @5/12, 3/4# @1, 1# @0.101, 0.121#

International Journal of Intelligent Systems DOI 10.1002/int

 ALTERNATIVE ANALYTIC HIERARCHY PROCESS SOLUTION 439
Table IV. Final weights related to the example.
LLMS TFN

A @0.388, 0.962#
B @0.163, 0.407#
C @0.077, 0.181#

We obtain three graded numbers, one for each alternative; the upper value is
in accordance with the characteristic of optimism l ⫽ 1, whereas the lower values
represent the pessimism l ⫽ 0.
We see that the alternative C1 dominates the other two alternatives and C2
dominates C3 , and then it is possible to see, in this case, that the preference order
is the following: C1 Ɑ C2 Ɑ C3 . ~See Figure 6.!

6. CONCLUSIONS

We present a new approach for the analytic hierarchy process in the general
case in which the linguistic labels are know as graded numbers. We have demon-
strated that the model is more general than the model proposed by Saaty.
Taking into account the definition of division given in Section 3, it is natural
that the final graded numbers have deviation toward the right; they are not
symmetric.
In the particular case in which we work only with the central value of the
solution, the result is the same as the corresponding in the original work of Saaty,
but this central value is not for the value l ⫽ 0.5, because, as we have said, the
final results are not symmetric.
We have the possibility to introduce the OWA operator 32 ; in this case the
three final values are taking account, introducing for this the degree of optimism.
We exhibit the advantages for the graded numbers, whose arithmetic is a sim-
ple extension of the Interval Analysis, whereas the fuzzy numbers need additional
conditions in order to be operated via their a-cut.

Figure 6. The three intervals associated with the three alternatives.

International Journal of Intelligent Systems DOI 10.1002/int
440 LAMATA
Acknowledgments
We are thankful to José Antonio Herencia for helpful comments and the anonymous refer-
ees for their valuable comments that improved the quality of this paper. This work was partially
supported by the DGICYT under project PB98-1305 and TIC2002-0341.

References
1. Zadeh LA. The concept of a linguistic variable and its application to approximate reason-
ing. Inform Sci 1975;8:199–249.
2. Saaty TL. The analytic hierarchy process. New York: McGraw-Hill; 1980.
3. Herencia JA. Graded sets and points: A stratified approach to fuzzy sets and points. Fuzzy
Set Syst 1996;77:191–202.
4. Herencia JA. Graded numbers and graded convergence of fuzzy numbers. Fuzzy Set Syst
1997;88:183–194.
5. Herencia JA, Lamata MT. A total order for the graded numbers used in decision problems.
Int J Uncertainty Fuzziness Knowl Based Syst 1999;7:267–276.
6. Negoita CV, Ralescu D. Applications of fuzzy sets to systems analysis. Basel: Birkhäuser
Verlag; 1975.
7. Moore RE. Interval analysis. Englewood Cliffs, NJ: Prentice-Hall; 1966.
8. Moore RE. Methods and applications of interval analysis. Philadelphia: SIAM; 1979.
9. Bouchon-Meunier B, Kosheleva O, Kreinovich V, Nguyen HT. Fuzzy numbers are the
only fuzzy sets that keep invertible operations invertible. Fuzzy Set Syst 1997;91:155–163.
10. Ralescu D. A survey of the representation of fuzzy concepts and its applications. In: Gupta
MM, Ragade RK, Yager RR, editors. Advances in fuzzy sets theory and applications.
Amsterdam: North-Holland; 1979. pp 77–91.
11. de Campos LM, González A. A subjective approach for ranking fuzzy numbers. Fuzzy Set
Syst 1989;29:145–153.
12. González A. A study of the ranking function approach through mean values. Fuzzy Set
Syst 1990;35:29– 41.
13. Fechner GT. Elements of psychophysics, vol 1. New York: Holt, Rinehart & Winston;
1965.
14. Thurstone LL. A law of comparative judgments. Psych Rev 1927;34:273–286.
15. Saaty TL. A scaling method for priorities in hierarchical structure. J Math Psych
1977;15:234–281.
16. Chu A, Kabala R, Springarn K. A comparison of two methods for determining the weights
of belonging to fuzzy sets. J Opt Theor Appl 1979;27:531–541.
17. Adamo JM. Fuzzy decision trees. Fuzzy Set Syst 1980;4:207–219.
18. Baas SM, Kwakernaak H. Rating and ranking of multiple-aspect alternatives using fuzzy
sets. Automatica 1977;13:47–58.
19. Baldwin JF, Guild NCF. Comparison of fuzzy sets on the same decision space. Fuzzy Set
Syst 1979;2:213–233.
20. Bortolan G, Degani R. A review of some methods for ranking fuzzy subsets. Fuzzy Set
Syst 1985;15:1–19.
21. Delgado M. Verdegay JL, Vila MA. A procedure for ranking fuzzy numbers using fuzzy
relations. Fuzzy Set Syst 1988;26:49– 62.
22. Matarazzo B, Munda G. New approaches for the comparison of L-R fuzzy numbers: A
theoretical and operational analysis. Fuzzy Set Syst 2001;118:407– 418.
23. Wang X, Kerre EE, Ruan D. Consistency of judgement matrix and fuzzy weights in fuzzy
analytic hierarchy process. Int J Uncertainty Fuzziness Knowl Based Syst 1995;3:35– 46.
24. Wang X, Kerre EE. Reasonable properties for the ordering of fuzzy quantities. Fuzzy Set
Syst 2001;118:375–385.
25. Yager RR. A procedure for ordering fuzzy subsets of the unit interval. Inform Sci
1981;24:143–161.

International Journal of Intelligent Systems DOI 10.1002/int

 ALTERNATIVE ANALYTIC HIERARCHY PROCESS SOLUTION 441
26. Dong WM, Wong FS. Fuzzy weighted averages and implementation of the extension prin-
ciple. Fuzzy Set Syst 1987;21:183–199.
27. Guh YY, Hon CC, Wang KM, Lee ES. Fuzzy weighted average: A max–min paired elim-
ination method. Comput Math Appl 1996;32:115–123.
28. Lee DH, Park D. An efficient algorithm for fuzzy weighted average. Fuzzy Set Syst
1997;87:39– 45.
29. Liou TS, Wang MJ. Fuzzy weighted average: An improved algorithm. Fuzzy Set Syst
1992;49:307–315.
30. van Laarhoven PJM, Pedrycz W. A fuzzy extension of Saaty’s priority theory. Fuzzy Set
Syst 1983;11:229–241.
31. Dubois D, Prade H. Fuzzy sets and systems: Theory and applications. New York: Aca-
demic Press; 1980.
32. Yager RR. Ranking fuzzy numbers using a-weighted valuations. Int J Uncertainty Fuzzi-
ness Knowl Based Syst 2000;8:573–591.

International Journal of Intelligent Systems DOI 10.1002/int