Вы находитесь на странице: 1из 12

ISSN 1750-9653, England, UK

International Journal of Management Science


and Engineering Management, 6(1): 3-13, 2011
http://www.ijmsem.org/
A modied TOPSIS technique in presence of uncertainty
and its application to assessment of transportation
systems

Soheil Sadi-Nezhad

, Kaveh Khalili Damghani


Industrial Engineering Department, Faculty of Engineering, Sciences & Research Branch, Islamic Azad University, Tehran
1651893181, Iran
(Received 7 July 2010, Revised 19 November 2010, Accepted 12 December 2010)
Abstract. In this paper, a modied TOPSIS approach based on Preference Ratio (PR) and an ecient fuzzy distance measurement
has been proposed in an uncertain environment. As it is not proper to rank fuzzy numbers using crisp measurements, PR is supplied
to rank Generalized Fuzzy Numbers (GFNs) in a relative manner rather than absolute way. Moreover, as human logic says distances
between fuzzy numbers should not be a crisp value. So, an ecient measurement for calculating distance between fuzzy numbers
has also been utilized in the core of proposed fuzzy TOPSIS procedure. The aforementioned segments of proposed procedure made
it well-posed and ecient for modeling complicated real life problems. The performance of proposed procedure has been compare
with an existing approach in selection of dierent transportation systems modes with conicting subjective and qualitative. The
associated expert system of proposed procedure has been developed through a linkage between MS-Excel 12.0 and Visual Basic 6.0.
Keywords: fuzzy sets, transportation, multiple criteria decision making, decision support systems, TOPSIS
1 Introduction
In the early decades Fuzzy Multiple Criteria Decision
Making (FMCDM) techniques have attracted lots of re-
searches eorts due to adaptability for real conditions de-
cision making problem. Decision making in complicated
real world situations needs lots of considerations and of-
ten multi-dimensional details (Gurumurthy and Kodali,
2008 [14]). One of the well-known Multiple Attribute De-
cision Making (MADM) approaches is the Technique for
Order of Preference by Similarity to Ideal Solution (TOP-
SIS). TOPSIS was introduced by Hwang and Yoon (1981)
[15]. TOPSIS ranks alternatives according to an algorithmic
procedure. Alternatives are sorted in decreasing order of
Closeness Coecient (CC) which is calculated with respect
to distance of a given alternative from both positive and
negative ideal solution concurrently in a multi-dimensional
space.
Ambiguous data is a big challenge for Decision Makers
(DMs). This fact motivated researchers to extend MCDM
techniques, including TOPSIS, in fuzzy environment. In
general, fuzzy sets are assumed to be proper paradigm to
plot some versions of uncertainty in which qualitative and
subjective criteria are considered. An impressive variety of
fuzzy TOPSIS algorithms with dierent applications are
developed in recent years.
Some of those can be outlined as Application of TOPSIS
in evaluating initial training aircraft under a fuzzy environ-
ment by Wang and Chang (2007) [36], A note on group
decision-making based on concepts of ideal and anti-ideal
points in a fuzzy environment by Wang and Hua (2007) [38],
Compromise ratio method for fuzzy multi-attribute group
decision making by Li (2007) [23], Generalizing TOPSIS
for fuzzy multiple-criteria group decision-making by Wang
and Lee (2007) [39], Fuzzy multi-criteria evaluation of in-
dustrial robotic systems by Kahraman et al. (2007a) [19],
A two phase multi-attribute decision-making approach for
new product introduction by Kahraman et al. (2007b) [20],
Group decision-making based on concepts of ideal and anti-
ideal points in a fuzzy environment by Kuo et al. (2007)
[22], Using fuzzy number for measuring quality of service
in the hotel industry by Bentez et al. (2007) [7], Multiple-
attribute decision making methods for plant layout design
problem by Yang and Hung (2007) [43], Extensions of TOP-
SIS for large scale multi-objective non-linear programming
problems with block angular structure by Abo-Sinna et al.
(2008) [3], Extension of the TOPSIS method for decision-
making problems with fuzzy data by Jahanshahloo et al.
(2006) [18], Multiple attribute decision-making methods for
the dynamic operator allocation problem by Yang et al.
(2007) [36], A fuzzy approach for supplier evaluation and
selection in supply chain management by Chen et al. (2006)
[10], Fuzzy TOPSIS method based on alpha level sets with
an application to bridge risk assessment by Wang and Elhag
(2006) [37], An interactive algorithm for large scale multi-
ple objective programming problems with fuzzy parame-
ters through TOPSIS approach by Abo-Sinna and Abou-
El-Enien (2006) [2], The Method of Grey Related Analysis
to Multiple Attribute Decision Making Problems with In-
terval Numbers by Desheng and Olson (2006) [41], Exten-

The authors would like to acknowledge the collaborating manner of experts in Research Committee of Iranian Trac Center.

Correspondence to: E-mail address: sadinejad@hotmail.com.


International Society of Management Science
And Engineering Management
Published by World Academic Press,
World Academic Union
4 S. Sadi-Nezhad & K. Damghani: A modied TOPSIS technique in presence of uncertainty
sions of the TOPSIS for group decision-making under fuzzy
environment by Chen (2000) [9], The evaluation of airline
service quality by fuzzy MCDM by Tsaura et al. (2002) [35],
Combining Grey Relation and TOPSIS Concepts for Select-
ing an Expatriate Host Country by Chen and Tzeng (2004)
[11], Extensions of TOPSIS for multi-objective large-scale
nonlinear programming problems by Abo-Sinna and Amer
(2005) [3], and Transshipment site selection using the AHP
and TOPSIS approaches under fuzzy environment by

On ut
and Soner (2008) [30], Sadi-Nezhad and Damghani (2010)
[32] proposed a fuzzy TOPSIS method based on preference
ratio and fuzzy distance measurement in order to assess per-
formance of trac police centers, Jadidi et al. (2009) [17]
also proposed a TOPSIS extension for multi-objective sup-
plier selection problem under price breaks based on Abo-
Sinna and Amer (2005) [3] and Abo-Sinna et al. (2008) [4]
approaches. Jadidi et al. (2008) [16] proposed an optimal
grey based approach based on TOPSIS concepts for sup-
plier selection problem.
Chu and Lin (2009) [12] proposed interval arithmetic
based fuzzy TOPSIS model.

On ut et al. (2010) [28] modeled
shopping center site selection problem in Istanbul. They
proposed a combined MCDM methodology using fuzzy
AHP for assigning weights of the criteria and fuzzy TOP-
SIS to determine the most suitable alternative.

On ut et al.
(2009) [30] developed a supplier evaluation approach based
on fuzzy analytic network process (ANP) and fuzzy TOP-
SIS methods to help a telecommunication company in the
GSM sector in Turkey. Athanasopoulos et al. (2009) [6] pre-
sented an expert system based on Max-Min set method
and fuzzy TOPSIS for coating selection of engineering part
which could handle both qualitative and quantitative vari-
ables. Malekly et al. (2010) [25] proposed a two-phased de-
cision process based on QFD and TOPSIS for selecting the
most suitable superstructure of a small to medium-span
highway bridge design. Wu et al. (2009) [40] proposed a
fuzzy MCDM procedure to evaluate banking performance.
They gathered associated criteria of banking performance
according to BSC perspectives. Then, the relative impor-
tance of the aforementioned criteria was calculated through
fuzzy AHP. Finally, three well-known MCDM approaches
including a fuzzy TOPSIS were used to evaluate three bank
performances as an illustrative example. Secme et al. (2009)
[33] proposed a fuzzy MCDM model to evaluate the perfor-
mances of banks. The proposed method used several nan-
cial and non-nancial indicators. Fuzzy AHP and TOP-
SIS integrated in the proposed model. Dursun and Karsak
(2010) [13] developed a fuzzy MCDM procedure based on
fusion of fuzzy information principles, 2-tuple linguistic
representation model, and TOPSIS to personnel selection
problem with dierent criteria such as organizing ability,
creativity, personality, and leadership which exhibit vague-
ness and imprecision. Kelemenis and Askounis (2010) [21]
also presented a fuzzy TOPSIS approach to select qualied
personnel incorporating a new concept for the ranking of
the alternatives. They applied a veto threshold. So the ul-
timate decision criterion is not the similarity to the ideal
solution but the distance of the alternatives from the veto
set by the decision makers.
As mentioned, in real cases some subjective and quali-
tative criteria may arise that can properly be represented
through fuzzy sets. Moreover, in many cases there is no
enough information about the problem this vagueness can
also be modeled through fuzzy paradigm. Beside aforemen-
tioned situations, exact data are not required or even not
easy/inexpensive to achieve in many real cases. Anyway,
fuzzy sets fuzzy sets have been proved to represent all afore-
mentioned situations properly. Most to the all aforemen-
tioned fuzzy TOPSIS procedures try to overcome the pre-
vious points through some kind of fuzzy set theory. Dif-
ferent extensions have been reported in literature of fuzzy
TOPSIS.
In fuzzy TOPSIS, the relative importance of alternatives
or rating of an alternative with respect to attributes can be
a fuzzy number. Therefore the calculations are done in fuzzy
environment and fuzzy operators are used. Main dierences
between fuzzy TOPSIS approaches can be summarized in
choosing a decision matrix normalization method, deter-
mining Fuzzy Positive Ideal solutions (FPIS) and Fuzzy
Negative Ideal solutions (FNIS), distance calculation be-
tween fuzzy numbers, and applied defuzzication method.
Most to the all of aforementioned research works presented
a procedure which uses a defuzzication method at early or
middle steps of the fuzzy TOPSIS algorithm. These meth-
ods convert fuzzy numbers to an associated crisp value
which will cause some rounding error as well as probable
and meaningful disturbance in nal ranking of alternatives.
Moreover, a most likely used technique to determine
FPIS & FNIS is to introduce (0, 0, 0, 0) as FNIS, and (1, 1,
1, 1) as FPIS for TrFNs as in Wang and Chang (2007) [36].
Determination ideal and anti-ideal point in this way may
result in following problems. These ideal and anti-ideal vec-
tors are determined without considering the real data of
the problem. Comparison of alternative with such ideal and
anti-ideal vectors may not be reasonable in real case situ-
ations. Moreover, this kind of ideal and anti-ideal determi-
nation diminishes the possibility of periodical comparison
of best and worst alternatives in order to make an overall
trend for alternative in a predened planning horizon.
Aforementioned shortages made us to modify the clas-
sic TOPSIS procedure in an uncertain environment which
has been modeled using fuzzy sets. The proposed modied
fuzzy TOPSIS procedure is assumed to have the following
advantages over existing fuzzy TOPSIS procedures in liter-
ature.
All distances between alternatives and FPIS and FNIS are
calculated using an ecient fuzzy distance measurement.
This yields more realistic properties in favor of representing
all distances between fuzzy numbers in fuzzy environment.
Although most to the all of existing fuzzy TOPSIS proce-
dures serve crisp distance between fuzzy numbers.
Comparisons of fuzzy numbers (i.e. closeness coecients)
have been accomplished using PR which has been known as
an ecient ranking method for GFNs. This makes the rank-
ing procedure more realistic in a fuzzy environment. Using
PR method for ranking, fuzzy numbers are ranked rela-
tively in an interval rather than absolutely which is usual
in existing fuzzy TOPSIS procedures.
The proposed modied fuzzy TOPSIS procedure can
properly model subjective and qualitative criteria as well
as crisp one.
The fuzzy environment is held till last step of proposed
fuzzy TOPSIS procedure. This results in reducing rounding
error and plotting uncertainty in a proper manner.
Determining the ideal and anti-ideal vectors are accom-
plished according to real data of cases.
In this paper we have presented a modied fuzzy TOP-
SIS approach based on Preference Ratio (PR) and fuzzy
distance measurement. PR was rst introduced by Modar-
res and Sadi-Nezhad (2001) [26] for fuzzy multiple crite-
ria group decision making. We used PR method for rank-
ing Fuzzy closeness coecients and ranking fuzzy numbers
IJMSEM email for contribution: submit@msem.org.uk
International Journal of Management Science and Engineering Management, 6(1): 3-13, 2011 5
to nd a proper relative order of alternatives. PR would
determine the preference of fuzzy numbers in an interval
through a relative manner rather than absolute way. It is
notable that developing methods for ranking fuzzy num-
bers has been attracted lots of research eorts. For in-
stance the eorts of Asady and Zendehnam (2007) [5] or
Abbasbandy and Asady (2006) [1] have worth to be men-
tioned. Moreover, distances between fuzzy numbers should
be a fuzzy measure. So we used an ecient fuzzy distance
measurement which was introduced by Chakraborty and
Chakraborty (2007) [8]. The calculations of distances be-
tween fuzzy numbers and ranking CCs are mainly dierent
from what are well-known in reported fuzzy TOPSIS ap-
proaches in literature.
The following sections of this paper are arranged as be-
low. Section 2 is allocated to extend some main arithmetic
operators in fuzzy environment and notations are intro-
duced. Concepts of PR are discussed in section 3. An ef-
cient fuzzy distance measure is also introduced in section
4. Due to our best knowledge this type of fuzzy distance
measure and concept of Preference Ratio are rarely used in
main core of proposed fuzzy TOPSIS algorithms. In section
5 modied fuzzy TOPSIS approach is introduced. 6th sec-
tion talks about application of recommended algorithm for
assessment of transportation systems performance in dier-
ent modes and nally in section 7, the paper will be ended
with a brief conclusion.
2 Notations and fuzzy operators
Let X be the universe of discourse, X = {x
1
, x
2
, , x
n
}.
A fuzzy set X of

A is a set of order pairs {(x
1
,

A(x
1
)
),
(x
2
,

A(x
2
)
), , (x
n
,

A(x
n
)
)} where

A
: X [0, 1] is the
membership function of

A, and

A(x
1
)
stands for the mem-
bership degree of x
i
in

A. Some denitions are presented as
follow (Zimmermann, 1991 [44]).
Denition 1. A trapezoidal fuzzy number (TrFN) can be
dened as a m = (a, b, c, d), where the membership function

m
is given by

m
(x) =
_

_
xa
ba
, a x b
1, b x c
dx
dc
, c x d
where [b, c] is called a mode interval of m, and a and d are
called lower and upper limits of m, respectively.
For sake of simplicity and without loss of generality, we
consider TrFNs throughout the paper and calculation are
extended by this assumption.
Denition 2. m = (a, b, c, d) is called a positive trapezoid
fuzzy number if a 0. a, b, c, and d are not identical.
Further detail about basic fuzzy arithmetic operators can
be found in (Zimmermann, 1991 [44]).
Lemma 1. In real cases something unusual occurs when
using arithmetic calculation for some especial negative fuzzy
numbers. For a convenient description note the following
example.
Example 1. Let

A = (0.305, 0.145, 0.894, 1.884) and

B
= {0.616, 0.491, 1.932, 3.483} as two negative TrFNs.

A()

B is calculated as below.

A()

B = (a
1
, a
2
, a
3
, a
4
)()(b
1
, b
2
, b
3
, b
4
)
=
_
a
4
b
1
,
a
3
b
2
,
a
2
b
3
,
a
1
b
4
_
, if

A < 0,

B < 0,

A = (0.305, 0.145, 0.894, 1.884),

B =(0.616, 0.491, 1.932, 3.483),

A()

B = (3.057, 1.820, 0.075, 0.088). (1)
It is clear that the result expression is not a TrFN and this
is not a coincidence at all. This result is likely when TrFNs
have the following properties:
TrFNs are negative (See denition 2.2 for positive
TrFNs).
TrFNs are asymmetric where plotting the TrFN will not
cause a symmetric trapezoid shape in 2-D surface.
There are lots of such TrFNs which hold these properties
and cause similar results.
Lemma 2. We modify such TrFNs in following manner
in order to prevent above problem in our algorithm. Let
m = (a, b, c, d) be a negative TrFN in which a is per-
manently negative and b, c and d are not equal. All
arguments are ordered ascending. A transition toward pos-
itive axis is used to prevent the aforementioned case. The
transited TrFN is named
m

= (0, |a| +b, |a| +c, |a| +d) = (0, b

, c

, d

).
Lemma 3. If two or more than two fuzzy numbers meet
the aforementioned conditions in Lemma 1, it is more rea-
sonable to transit the fuzzy numbers in a ways that their
relative situations to each other are constant after transi-
tion procedure. So we dene a unique transition value for
all of them as follow. Let

A
1
,

A
2
,

A
3
, ,

A
n
be n TrFN
parameterized by (a
i
, b
i
, c
i
, d
i
), i = 1, 2, , n respectively.
They meet aforementioned conditions in Lemma 1. Calcu-
late | min (a
i
)|, i = 1, 2, , n and name it unique transi-
tion value (UTV). Transit all TrFNs toward positive axis
by UTV. More formally a supposed TrFN in this string
will be (a
i
+ | min (a
i
)|, b + | min (a
i
)|, c + | min (a
i
)|, d +
| min (a
i
)|), i = 1, 2, , n. It is clear that by this transi-
tion method all TrFNs will be transited in a fairly manner
and their relative positions will be constant after transition.
A direct result of this transition is that left spreads of all
transited TrFNs will be a positive or at least zero value for
those TrFNs which their left spreads are numerically equal
to UTV. Since zero value for left spreads will cause a divi-
sion by zero error during the fuzzy arithmetic, an arbitrary
very small positive value like > 0 is added to UTV. So
UTV = | min (a
i
)| +, i = 1, 2, , n is the nal UTV. The
transited TrFNs are (a
i
+ | min (a
i
)| +, b + | min (a
i
)| +
, c +| min (a
i
)| +, d +| min (a
i
)| +), i = 1, 2, , n.
3 Preference ratio
Our proposed approach is developed on the basis of Pref-
erence Ratio concept. Preference Ratio is a ranking method
introduced by Modarres and Sadi-Nezhad (2001) [26]. They
evaluate fuzzy numbers point by point and rank them at
each point in PR method. Then, the overall preference over
all points is calculated. So the preference in this way is rel-
ative rather than absolute. Suppose the objective is to rank
I fuzzy numbers. Let N
i
be the ith one dened over a real
domain S
i
R, and it is identied by a membership func-
tion (
N
i
(x)), x S
i
, with
N
i
(x) [0, 1]. Let S
i
be the
support of N
i
, one more precisely S
i
= {x,
N
i
(x) > 0},
and =

I
i=1
S
i
then is the union of the support of all
IJMSEM email for subscription: info@msem.org.uk
6 S. Sadi-Nezhad & K. Damghani: A modied TOPSIS technique in presence of uncertainty
algorithm. Let ) , , , (
~
d c b a m = be a negative TrFN in which "a" is permanently negative and "b", "c" and "d"
are not equal. All arguments are ordered ascending. A transition toward positive axis is used to prevent
the aforementioned case. The transited TrFN is named ) , , , 0 ( ) , , , 0 (
~
d c b d a c a b a m = + + + = .











Fig.1: Transition of a Negative TrFN.

Lemma 2.3. If two or more than two fuzzy numbers meet the aforementioned conditions in lemma2.1, it
is more reasonable to transit the fuzzy numbers in a ways that their relative situations to each other are
constant after transition procedure. So we define a unique transition value for all of them as follow. Let
n
A A A A
~
,...,
~
,
~
,
~
3 2 1
be n TrFN parameterized by (ai, bi, ci, di), i=1,2,,n respectively. They meet
aforementioned conditions in lemma2.1. Calculate n , 1,2, i , ) min(a
i
= and name it unique transition
value (UTV). Transit all TrFNs toward positive axis by UTV. More formally a supposed TrFN in this
string will be . ,..., 2 , 1 ), ) min( , ) min( , ) min( , ) min( ( n i a d a c a b a a
i i i i i i i i
= + + + + It is clear that by this
transition method all TrFNs will be transited in a fairly manner and their relative positions will be
constant after transition. A direct result of this transition is that left spreads of all transited TrFNs will be a
positive or at least zero value for those TrFNs which their left spreads are numerically equal to UTV.
Since zero value for left spreads will cause a division by zero error during the fuzzy arithmetic, an
arbitrary very small positive value like 0 > is added to UTV. So n , 1,2, i , ) min(a
i
= + = UTV is the
final UTV. The transited TrFNs are
. ,..., 2 , 1 ), ) min( , ) min( , ) min( , ) min( ( n i a d a c a b a a
i i i i i i i i
= + + + + + + + +
3. Preference Ratio
Our proposed approach is developed on the basis of Preference Ratio concept. Preference Ratio is a
ranking method introduced by Modarres & Sadi-Nezhad, (2001). They evaluate fuzzy numbers point by
point and rank them at each point in PR method. Then, the overall preference over all points is calculated.
So the preference in this way is relative rather than absolute. Suppose the objective is to rank I fuzzy
numbers. Let
i
N be the ith one defined over a real domain R S
i
, and it is identified by a membership
function ) ), ( (
i N
S x x
i
, with ] 1 , 0 [ ) ( x
i
N
. Let
i
S be the support of
i
N , one more precisely } 0 ) ( , { > = x x S
i
N i
,
and
i
I
i
S
1 =
= U then

is the union of the support of all fuzzy numbers. In other word, fuzzy numbers are
ranked over

. To rank fuzzy numbers, we assume their spans are not disjoint, because in that case the
ranking is clear.
A fuzzy number is evaluated by a function called Preference Function. At each point

, this function
is defined as follows:
a
b
b d a
c d
) (x
x
c
a 0
a 0
a 0
a 0
a b b c , d a
c d
) (x
x
a 0
a 0
Fig. 1 Transition of a negative TrFN
fuzzy numbers. In other word, fuzzy numbers are ranked
over To rank fuzzy numbers, we assume their spans are
not disjoint, because in that case the ranking is clear.
A fuzzy number is evaluated by a function called Pref-
erence Function. At each point, this function is dened as
follows:
G() =
_
U

(x)dx
_
U
L
(x)dx
, (2)
where (x) is the membership function of the fuzzy num-
ber, L = min {x : x } and U = max{x : x }.
This function has the same denition as 1 F() in prob-
ability theory, where F() = P[x ] is the distribution
function. At , let p() = i denote the ith fuzzy num-
ber, which is the most preferred one. Therefore p() = i,
if G
i
() = max {G

, j I}. where G
i
() is the prefer-
ence function of the jth fuzzy number. Let be the set
of point at which the ith number is ranked number one.
Then = { , p() = i}.
Denition 3. For the ith fuzzy number, R(i), the Prefer-
ence Ratio, is by denition, the percentage of that the
ith fuzzy number is the most preferred one. Then
R(i) =
|
i
|
||
, (3)
where |
i
| and || are the lengths of the real set
i
and
respectively. Modarres and Sadi-Nezhad (2001) [26] de-
veloped an algorithm for determining preference ratio in
TFN cases. An algorithm is also proposed for any contin-
uous cases by Modarres and Sadi-Nezhad (2005) [27]. Also
to make two fuzzy numbers A and B preference ratio equiv-
alent, they dene this concept as below.
Denition 4. (a) We dene that two fuzzy numbers A and
B are preference ratio equivalent if R(A) = R(B) = 0.5,
where R(A) and R(B) are preference ratio of A and B, re-
spectively. Preference ratio equivalence is shown as A
PR
B.
If
k A
PR
B,
then we say k is the Equivalence multiplier of A with
respect to B.
3.1 Ranking n fuzzy numbers
In all fuzzy TOPSIS approaches, a fuzzy dominance re-
lation is needed to be dened in order to obtain a rank
of fuzzy numbers. Modarres and Sadi-Nezhad (2001) [26]
have proposed a ranking method for fuzzy numbers that
compares them relatively, rather than preferring one num-
ber absolutely to the others. For two fuzzy number

N
1
and

N
2
they developed an algorithm to nd scalar K such as

N
1
PR
K

N
2
.
Modarres and Sadi-Nezhad (2005) [27]. An algorithm for
ranking n TFNs was proposed by Sadi-Nezhad and Ghaleh-
Assadi (2008) [31]. They also have developed an algorithm
for Max operator of two TFNs by preference ratio. They
considered 3 initial estimation methods for Max operator
and ve cases of two TFN positions in relation to each
other. They have proven that the nal results are depen-
dent to initial estimation method.
4 Fuzzy distance measure
Distance measures have become important due to the
signicant applications in diverse elds like remote sensing,
data mining, pattern recognition, multivariate data anal-
ysis, MADM methods and etc. Most of proposed meth-
ods measure the distances of fuzzy numbers as a crisp
value, although human intuitions oer a fuzzy measure
is more acceptable and reasonable for measuring the dis-
tances of fuzzy numbers. Answering to this logical ques-
tion that if the numbers themselves are not known ex-
actly, how can the distance between them be an exact
value? made researchers eager to develop several fuzzy dis-
tance measures. Tran and Duckstein (2002) [34] developed
a method for comparison of fuzzy numbers using a fuzzy
distance measure. Voxman introduced a fuzzy distance for
generalized fuzzy numbers (GFN) using -cut. The Voxman
new distances measurement method has been improved by
Chakraborty and Chakraborty (2007) [8]. They have intro-
duced a novel fuzzy distance measure and have proven the
quality of their method is over the Voxman method through
ambiguity and fuzziness which are two main attributes of
fuzzy numbers.
We have used theirs as the main core of distance cal-
culation step in our modied TOPSIS method as a val-
idated method. In the next paragraph this method has
been presented briey to make a comfort sense. Details can
be founded in Chakraborty and Chakraborty (2007) [8] or
Sadi-Nezhad and Damghani (2010) [32].
4.1 Dening fuzzy distance for GFNs
Let us consider two GFNs as

A
1
= (
1
,
2
;
1
,
1
) and

A
2
= (
3
,
4
;
2
,
2
). Therefore the -cut of

A
1
and

A
2
rep-
resents two intervals, respectively [

A
1
]

= [A
L
1
(), A
R
1
()]
and [

A
2
]

= [A
L
2
(), A
R
2
()], for all [0, 1]. Since it
IJMSEM email for contribution: submit@msem.org.uk
International Journal of Management Science and Engineering Management, 6(1): 3-13, 2011 7
may be possible to obtain the distance between two inter-
val numbers by means of their dierence, they employ the
interval-distance operation for interval [A
L
1
(), A
R
1
()] and
[A
L
2
(), A
R
2
()] to formulate the fuzzy distance between

A
1
and

A
2
. So the distance between [

A
1
]

and [

A
2
]

for every
[0, 1] can be one of the following:
d(

A
1
,

A
2
)

=
_
_
_
[

A
1
]

[

A
2
]

, if
A
L
1
(1)+A
R
1
(1)
2

A
L
2
(1)+A
R
2
(1)
2
[

A
2
]

[

A
1
]

, if
A
L
1
(1)+A
R
1
(1)
2
<
A
L
2
(1)+A
R
2
(1)
2
(4)
In order to consider both notations they use a 0-1 variable
lambda. So:
[d
L

, d
R

] = ([

A
1
]

[

A
2
]

) + (1 )([

A
2
]

[

A
1
]

), (5)
=
_
_
_
1, if
A
L
1
(1)+A
R
1
(1)
2

A
L
2
(1)+A
R
2
(1)
2
0, if
A
L
1
(1)+A
R
1
(1)
2
<
A
L
2
(1)+A
R
2
(1)
2
(6)
where
d
L

=[A
L
1
() A
L
2
() + A
R
1
() A
R
2
()]
+ [A
L
2
() A
R
1
()], (7)
d
R

=[A
L
1
() A
L
2
() + A
R
1
() A
R
2
()]
+ [A
R
2
() A
L
1
()]. (8)
Therefore the fuzzy distance between

A
1
and

A
2
is dened
by:

d(

A
1
,

A
2
) = (d
L
=1
, d
R
=1
; , ), (9)
where
= d
L
=1
max
_
_
1
0
d
L

d
_
and =
_
1
0
d
R

d d
R
=1
.
(10)
Left and right spread of the fuzzy distance between

A
1
and

A
1
are represented by , , respectively. Using these
initial materials, we develop a modied TOPSIS method
based on preference ratio and fuzzy distance measurement
in next section.
Lemma 4. As mentioned before fuzzy numbers are as-
sumed to be TrFNs throughout the paper. Let us extend
the proposed fuzzy distance measurement for TrFNs. Gen-
erally, two main categories may be occurred for distance
calculation. In rst case the compared TrFNs have some
overlap while in second case TrFNs are assumed to be ex-
actly distinct. These categories are represented in 11 dif-
ferent cases. These could be found for two arbitrary TrFNs
like m
1
= (a
1
, b
1
, c
1
, d
1
) and m
2
= (a
2
, b
2
, c
2
, d
2
) in Fig. 2.
(a) For an arbitrary -cut level like
0
in the rst category
(i.e. graphs VIII and IX in Fig. 2) or in the second category
(i.e. graph X in Fig. 2) we have:

d(

A
1
,

A
2
) = (d
L
=1
, d
R
=1
; , )
= (b
2
c
1
, c
2
b
1
; b
2
c
1
max{0.5[(a
2
d
1
)
+ (b
2
c
1
)], 0}, 0.5[(d
2
a
1
) + (b
1
c
2
)]).
(11)
(b) And through the same procedure for an arbitrary -cut
level like
0
in the rst category (i.e. graphs IV and V in
Fig. 2) or in the second category (i.e. graph XI in Fig. 2)
we have:

d(

A
1
,

A
2
) = (d
L
=1
, d
R
=1
; , )
= (b
1
c
2
, c
1
b
2
; b
1
c
2
max{0.5[(a
1
d
2
)
+ (b
1
c
2
)], 0}, 0.5[(d
1
a
2
) + (b
2
c
1
)])
(12)
The proofs of (11) and (12) can be found in Sadi-Nezhad
and Khalili Damghani (2010) [32]. The procedure is the
same for all other cases and categories shown in Fig. 2.
5 Proposed TOPSIS algorithm
In this section we revisit a fuzzy TOPSIS approach de-
veloped by Sadi-Nezhad and Khalili Damghani (2010) [32].
Then, the modied fuzzy TOPSIS is applied in selection
of dierent transportation systems modes with conicting
subjective and qualitative. They developed a modied fuzzy
TOPSIS using PR and fuzzy distance measurement. So, a
brief introduction is supplied here. Details can be found
in Sadi-Nezhad and Khalili Damghani (2010) [32]. First of
all let us plan the MADM problem more formally. Suppose
there are m alternative and n attribute. We ask k dier-
ent experts about scoring each alternative with respect to
attributes using a mixed qualitative-quantitative measure-
ment scale. Let x
k
i j
be the score which kth expert have allo-
cated to ith alternative with respect to jth attribute. The
weights of attributes have been asked or available in form
of fuzzy numbers. By these denitions the problem can be
represented more formally as follow.

D
k
=
_

_
x
k
11
x
k
12
x
k
1j
x
k
21
x
k
22
x
k
2j
.
.
.
.
.
.
.
.
.
.
.
.
x
k
i1
x
k
i2
x
k
i j
_

W
k
= [ w
k
1
, w
k
2
, , w
k
j
]. (13)

D
k
is the fuzzy decision matrix of kth expert. It has i row
and j column which are related to alternatives and at-
tributes, respectively. The fuzzy weighs of attributes for kth
expert is shown in form of

W
k
matrix which is extracted
from questionnaires.

D =
_

_

11

12

1j

21

22

2j
.
.
.
.
.
.
.
.
.
.
.
.

i1

i2

i j
_

_
(14)
where
i j
=

1
i j
+
1
i j
++
k
i j
k
.

W = [ w
1
, w
2
, , w
j
], (15)
where w
j
=
w
1
j
+ w
1
j
++ w
k
j
k
.
Without lose of generality, all data are assumed to be
TrFNs. By these assumptions the steps of proposed fuzzy
algorithm are as below.
Step 1. In order to achieve a more smooth decision matrix,
we apply a columnar normalization at rst step. In fact this
step guarantees all data of a certain column in decision ma-
trix are laid down in a reasonable interval. We use linear
scales transform normalization function in this study which
preserves the property that the ranges of normalized TrFN
to be included in [0, 1] interval. We name the normalized
decision matrix

N where:

N =
_

_
r
11
r
12
r
1j
r
21
r
22
r
2j
.
.
.
.
.
.
.
.
.
.
.
.
r
i1
r
i2
r
i j
_

_
(16)
where
IJMSEM email for subscription: info@msem.org.uk
8 S. Sadi-Nezhad & K. Damghani: A modied TOPSIS technique in presence of uncertainty
)]) ( ) [( 5 . 0 }, 0 )], ( ) [( 5 . 0 max{ ; , ( ) , ; , ( )
~
,
~
(
~
1 2 2 1 2 1 2 1 2 1 2 1 2 1 1 1 2 1
c b a d c b d a c b b c c b d d A A d
R L
+ + = =
= =




The proofs of (11) and (12) can be found in Sadi-Nezhad and Khalili Damghani, (2010). The procedure is
the same for all other cases and categories shown in figure2.


















Fig. 2: Categories for fuzzy distance calculation.
5. Proposed TOPSIS Algorithm
In this section we revisit a fuzzy TOPSIS approach developed by Sadi-Nezhad and Khalili Damghani,
(2010). Then, the modified fuzzy TOPSIS is applied in selection of different transportation systems
modes with conflicting subjective and qualitative. They developed a modified fuzzy TOPSIS using PR
and fuzzy distance measurement. So, a brief introduction is supplied here. Details can be found in Sadi-
Nezhad and Khalili Damghani, (2010). First of all let us plan the MADM problem more formally.
Suppose there are m alternative and n attribute. We ask k different experts about scoring each alternative
with respect to attributes using a mixed qualitative-quantitative measurement scale. Let
k
ij
x
~
be the score
which kth expert have allocated to ith alternative with respect to jth attribute. The weights of attributes
have been asked or available in form of fuzzy numbers. By these definitions the problem can be
represented more formally as follow.
[ ]
k
j
k k k
k
ij
k
i
k
i
k
j
k k
k
j
k k
k
w w w W
x x x
x x x
x x x
D
~
...
~ ~
~
,
~
...
~ ~
~
...
~ ~
~
...
~ ~
~
2 1
2 1
2 22 21
1 12 11
=

=
M M M M

k
D
~
is the fuzzy decision matrix of kth expert. It has i row and j column which are related to alternatives
and attributes, respectively. The fuzzy weighs of attributes for kth expert is shown in form of
k
W
~
matrix
which is extracted from questionnaires.
) 13 (
) (x
x
) (x
x
Subscribed as 1
Subscribed as 2
) (x
x

) (x
x
) (x
x
) (x
x
) (x
x
) (x
x
) (x
x
) (x
x
) (x
x



) 12 (
Fig. 2 Categories for fuzzy distance calculation
r
i j
=
_

_
_
a
i j
d
+
j
,
b
i j
d
+
j
,
c
i j
d
+
j
,
d
i j
d
+
j
,
_
,
if j is a benet attribute
_
a

i j
d
i j
,
a

i j
c
i j
,
a

i j
b
i j
,
a

i j
ai j
,
_
,
if j is a cost attribute and d
+
j
is not zero
_
1
a
i j
d
+
j
, 1
b
i j
d
+
j
, 1
c
i j
d
+
j
, 1
d
i j
d
+
j
,
_
,
if j is a cost attribute and d
+
j
is not zero
(17)
and
d
+
j
= max(d
i j
), a

i j
= min(d
i j
), i = 1, 2, , m. (18)
Step 2. According to dierent importance of the attributes,
construct the weighted normalized decision matrix as:

V =
_

_
v
11
v
12
v
1j
v
21
v
22
v
2j
.
.
.
.
.
.
.
.
.
.
.
.
v
i1
v
i2
v
i j
_

_
v
i j
= r
i j
() w
i j
= (a
i j
, b
i j
, c
i j
, d
i j
), (19)
where, v
i j
are a normalized TrFN and their arguments vary-
ing in closed interval [0, 1].
Step 3. Dene a fuzzy positive-ideal solution and a fuzzy
negative-ideal solution as follow. Name them S
+
and S

,
respectively.
S
+
= ( v
+
i1
, v
+
i2
, , v
+
i j
), (20)
where v
+
i j
= max v
i j
= (max a
i j
, max b
i j
, max c
i j
, max d
i j
),
i = 1, 2, , m, if j is a benet attribute, v
+
i j
= min v
i j
=
(min a
i j
, min b
i j
, min c
i j
, min d
i j
), i = 1, 2, , m, if j is a
cost attribute.
S

= ( v

i1
, v

i2
, , v

i j
), (21)
where v

i j
= min v
i j
= (min a
i j
, min b
i j
, min c
i j
, max d
i j
),
i = 1, 2, , m, if j is a benet attribute, v

i j
= min v
i j
=
(max a
i j
, max b
i j
, min c
i j
, max d
i j
), i = 1, 2, , m, if j is a
cost attribute.
Step 4. Calculate fuzzy distance of each alternative from S
+
and S

. Call these distances positive fuzzy distance (PFD)


and negative fuzzy distance (NFD), respectively.

PFD
i
=

d(A
i
, S
+
), i = 1, 2, , m, (22)

NFD
i
=

d(A
i
, S

), i = 1, 2, , m. (23)
These distance are calculate using fuzzy distance measure
(i.e. subsection 4.1).
Step 5. Apply Lemma 3 on PFDs and NFDs.
Step 6. Dene a fuzzy closeness coecient (FCC) as follows.

FCC
i
=

d(A
i
, S

d(A
i
, S

) +

d(A
i
, S
+
)
=

NFD
i

NFD
i
+

PFD
i
, i = 1, 2, , m. (24)
Obviously whatever the

FCC
i
, i = 1, 2, , m value is
close to unit, the utility of the associated alternative is
higher for decision makers group. But it is clear that

FCC
i
s,
i = 1, 2, , m are TrFNs and generally a TrFN is not big-
ger than the other in an absolute manner. TrFNs are com-
pared relatively in a proposed interval using PR method.
Step 7. Apply step 1 and step 2 of algorithm III from
Modarres and Sadi-Nezhad (2005) [27] for achieving CC
i
,
i = 1, 2, , m crisp value which are associated to each

FCC
i
, i = 1, 2, , m.
Step 8. Order the CC
i
, i = 1, 2, , m in non-increasing
mode. Choose the alternative with biggest CC
i
index.
IJMSEM email for contribution: submit@msem.org.uk
International Journal of Management Science and Engineering Management, 6(1): 3-13, 2011 9
Table 1 Performance indicators & indices for dierent transportation modes
Mode No. Mode Name
General Indicators
Detailed
Indices
Service Quality Outcome Cost eciency
Mode 1 Walking
Sidewalk/path supply
Pedestrian LOS
Crosswalk conditions
Pedestrian mode
split Avg. annual
walk distance
Pedestrian crash rates
Cost per sidewalk-km
Cost per walk-km
Cost per capita
Mode 2 Cycling
Bike path and
lane supply Cycling
LOS Path conditions
Bicycle mode split
Avg. annual cycle
distance Cyclist
crash rates
Cost per path-km
Cost per cycle-km
Cost per capita
Mode 3 Automobile
Roadway supply
Roadway pavement
condition Roadway
LOS Parking
availability
Avg. auto trip
travel time Vehicle
energy consumption
and pollution
emissions Motor
vehicle crash rates
Cost per lane-km
Cost per vehicle-km
User cost per
capita External
cost per capita
Mode 4 Public transit
Transit supply
Transit LOS
Transit stop and
station quality Fare
aord ability
Transit mode split
Per capita transit
travel Avg. transit
trip travel time
Transit crash and
assault rates
User cost per
pass.-km User
cost per capita
Subsidy per capita
Mode 5 Taxi
Taxi supply
Average response
time
Taxi use Taxi
crash and assault
rates
Cost per taxi-trip
External costs
Mode 6 Multi-modal
Transport system
integration Accessibility
from homes to
common destinations
User survey results
Total transportation
costs Total average
commute time Total
crash casualty rates
Total cost passenger-km
Total cost per
capita External cost
per capita
Mode 7 Aviation
Airport supply Air
travel service
frequency Air
travel reliability
Air travel use
Air travel
crash rates
Cost per trip
External costs
Airport subsidies
Mode 8 Rail
Rail line supply
Rail service speed
and reliability
Rail mode split
Rail trac volumes
Rail crash rates
Cost per rail-km
Cost per tonne-km
External costs
Mode 9 Marine
Marine service supply
Marine service speed
and reliability
Marine mode split
Marine trac
volumes Marine
accident rates
Cost per tonne-km
Subsidies External
costs
6 Assessment of transportation systems
performance
Let us represent this multi attribute group decision mak-
ing problem. Assessment and evaluation of Transportation
Systems Performance is essential. Eciency assessment has
dierent usages like recognizing amount of performance de-
crease or increase, a proper guideline for new investments
and development plans, selecting best improvement pro-
cedure for having a totally ecient Transportation Sys-
tems and so on. Tab. 1 represents dierent transporta-
tion modes (Litman, 2008 [24]). Each transportation modes
is assessed through dierent indices. Ranking these trans-
portation modes subject to multi attributes will be con-
ducted to a MADM problem. We proposed each trans-
portation modes as an alternative. Alternatives are scored
subject to each attributed in an ambiguous environment.
Without lose of generality, we consider TrFNs to model
ambiguity in data.
Three general types of performance indicators can gen-
erally be categorized in a transport system as follow.
Service quality: These reect the quality of service expe-
rienced by users.
Outcomes: These reect outcomes or outputs.
Cost eciency: These reect the ratio of inputs (costs)
to outputs (desired benets).
IJMSEM email for subscription: info@msem.org.uk
10 S. Sadi-Nezhad & K. Damghani: A modied TOPSIS technique in presence of uncertainty
Fig. 3 Interface of developed application software for PR methodology
Table 2 Linguistic variable for importance of each attribute
Linguistic variable TrFN
Extreme low (EL) (0, 0, 0.1, 0.2)
Very low (VL) (0.1, 0.2, 0.3, 0.4)
Low (L) (0.2, 0.3, 0.4, 0.5)
Medium low (ML) (0.3, 0.4, 0.5, 0.6)
Medium (M) (0.4, 0.5, 0.6, 0.7)
Medium high (MH) (0.5, 0.6, 0.7, 0.8)
High (H) (0.6, 0.7, 0.8, 0.9)
Very high (VH) (0.7, 0.8, 0.9, 1.0)
Extreme high (EH) (0.8, 0.9, 1.0, 1.0)
Table 3 Linguistic variable for rating of each alternative with
respect to each attribute
Linguistic variable TrFN
Extreme low (EL) (0, 0, 1, 2)
Very low (VL) (1, 2, 3, 4)
Low (L) (2, 3, 4, 5)
Medium low (ML) (3, 4, 5, 6)
Medium (M) (4, 5, 6,7)
Medium high (MH) (5, 6, 7, 8)
High (H) (6, 7, 8, 9)
Very high (VH) (7, 8, 9, 0)
Extreme high (EH) (8, 9, 10, 10)
Each type is important. Service quality reects users
perspectives. Outcomes reect planning objectives. Cost ef-
ciency reects economic performance. As mentioned be-
fore, Tab. 1 illustrates examples of these indicators for var-
ious transport modes. Note that all attributes are equally
scaled to comparing the eciency of Transportation Sys-
tems in relatively homogeneous conditions. Moreover, there
is no doubt that other factors are participant. It is notable
that addition of other attributes has no disturbance for the
logic of proposed procedure. It is obvious that in our case a
transport mode, is considered as an ecient/ideal one when
with a high service quality (attribute 1) and a high outcome
(attribute 2), has a low cost (attribute 3). Attribute 1 and
attribute 2 have a rough and ambiguous nature that can
appropriately be covered by a fuzzy numbers. Attribute 3
is basically crisp values but are treated as TrFNs due to
dierent reasons. So, a fuzzy environment which has been
equipped by a group of expert ideas can properly overcome
this ambiguous, complicated and vague situation. It is clear
that attribute 1 and attribute 2 are qualitative while at-
tribute 3 is reasonable to be proposed quantitative. This
attributes are described properly using linguistic variable
and eciently can be transformed to associated fuzzy num-
bers. More formally our case is a fuzzy MADM problem
with 9 alternatives as states and 3 attribute. We asked 30
transportation experts for rating quantitative attribute by
TrFNs. This rating was done by linguistic variable mea-
sure and a transformation phase of linguistic variable to
associate TrFNs for qualitative attributes. Each attribute
IJMSEM email for contribution: submit@msem.org.uk
International Journal of Management Science and Engineering Management, 6(1): 3-13, 2011 11
Table 4 Final aggregation of experts rating
Attribute 1 Attribute 2 Attribute 3
Mode 1 (5.25, 6, 7, 7.5) (5.75, 6.75, 7.75, 8.5) (3.5, 4.25, 5.25, 6.25)
Mode 2 (2.25, 3, 4, 5) (2.5, 3.5, 4.5, 5.5) (3.25, 4.25, 5.25, 6.25)
Mode 3 (4, 4.75, 5.75, 6.5) (6.25, 7.25, 8.25, 8.75) (4, 5, 6, 7)
Mode 4 (2.5, 3.25, 4.25, 5.25) (2.75, 3.75, 4.75, 5.75) (4.75, 5.75, 6.75, 7.75)
Mode 5 (4, 5, 6, 7) (2.5, 3.25, 4.25, 5.25) (7, 8, 9, 9.5)
Mode 6 (3.75, 4.75, 5.75, 6.75) (2.5, 3.5, 4.5, 5.5) (3.75, 4.75, 5.75, 6.75)
Mode 7 (3, 3.5, 4.5, 5.5) (1.5, 2, 3, 4) (4.5, 5.25, 6.25, 7)
Mode 8 (6.5, 7.5, 8.5, 9) (5.25, 6.25, 7.25, 8) (4.75, 5.75, 6.75, 7.5)
Mode 9 (7.25, 8.25, 9.25, 9.75) (1.75, 2.5, 3.5, 4.5) (2.5, 3.5, 4.5, 5.5)
Table 5 Final importance of attributes
Attribute 1 Attribute 2 Attribute 3
(0.425, 0.5, 0.6, 0.675) (0.25, 0.325, 0.425, 0.525) (0.4, 0.5, 0.6, 0.7)
Table 6 Normalized decision matrix
Attribute 1 Attribute 2 Attribute 3
Mode 1 (0.724, 0.727, 0.757, 0.769) (0.92, 0.931, 0.939, 0.971) (0.5,0.531, 0.583, 0.658)
Mode 2 (0.31, 0.364, 0.432, 0.513) (0.4, 0.483, 0.545, 0.629) (0.464,0.531, 0.583, 0.658)
Mode 3 (0.552, 0.576, 0.622, 0.667) (1, 1, 1, 1) (0.571, 0.625,0.667, 0.737)
Mode 4 (0.345, 0.394, 0.459, 0.538) (0.44, 0.517, 0.576,0.657) (0.679, 0.719, 0.75, 0.816)
Mode 5 (0.552, 0.606, 0.649, 0.718) (0.4, 0.448, 0.515, 0.6) (1, 1,1, 1)
Mode 6 (0.517, 0.576, 0.622, 0.692) (0.4, 0.483, 0.545,0.629) (0.536, 0.594, 0.639, 0.711)
Mode 7 (0.414, 0.424, 0.486, 0.564) (0.24, 0.276, 0.364,0.457) (0.643, 0.656, 0.694, 0.737)
Mode 8 (0.897, 0.909, 0.919, 0.923) (0.84, 0.862, 0.879,0.914) (0.679, 0.719, 0.75, 0.789)
Mode 9 (1, 1, 1, 1) (0.28, 0.345, 0.424, 0.514) (0.357, 0.438, 0.5,0.579)
Table 7 Weighted normalized decision matrix
Attribute 1 Attribute 2 Attribute 3
Mode 1 (0.308, 0.364, 0.454, 0.519) (0.23, 0.303, 0.399, 0.51) (0.2,0.266, 0.35, 0.461)
Mode 2 (0.132, 0.182, 0.259, 0.346) (0.1, 0.157, 0.232, 0.33) (0.186,0.266, 0.35, 0.461)
Mode 3 (0.234, 0.288, 0.373, 0.45) (0.25, 0.325, 0.425,0.525) (0.229, 0.313, 0.4, 0.516)
Mode 4 (0.147, 0.197, 0.276, 0.363) (0.11, 0.168, 0.245,0.345) (0.271, 0.359, 0.45, 0.571)
Mode 5 (0.234, 0.303, 0.389, 0.485) (0.1, 0.146, 0.219, 0.315) (0.4,0.5, 0.6, 0.7)
Mode 6 (0.22, 0.288, 0.373, 0.467) (0.1, 0.157, 0.232, 0.33) (0.214,0.297, 0.383, 0.497)
Mode 7 (0.176, 0.212, 0.292, 0.381) (0.06, 0.09, 0.155, 0.24) (0.257,0.328, 0.417, 0.516)
Mode 8 (0.381, 0.455, 0.551, 0.623) (0.21, 0.28, 0.373, 0.48) (0.271,0.359, 0.45, 0.553)
Mode 9 (0.425, 0.5, 0.6, 0.675) (0.07, 0.112, 0.18, 0.27) (0.143,0.219, 0.3, 0.405)
Table 8 Fuzzy positive ideal solution & fuzzy negative ideal solution
Attribute 1 Attribute 2 Attribute 3
FPIS (0.425, 0.5, 0.6, 0.675) (0.25, 0.325, 0.425, 0.525) (0.143, 0.218, 0.3, 0.405)
FNIS (0.132, 0.181, 0.259, 0.346) (0.06, 0.087, 0.154, 0.24) (0.4, 0.5, 0.6, 0.7)
has dierent importance weight which also has been asked
through a TrFN. We used the following tables for repre-
senting the importance of each attribute and rating of each
alternative with respect to each attribute. We ask 30 ex-
perts about rating 9 alternatives with respect to the 3 at-
tributes using linguistic variable mentioned in Tab. 3. The
aggregation of experts ratings according to pre-step 1 of
algorithm for k = 30 (i.e. Eq. (14) and Eq. (15)) is shown
in Tab. 4. The same process was implemented for impor-
tance weight of attributes according to Tab. 2. The nal
importance weights of attributes are also presented in Tab.
5. Note that all table contents are TrFNs.
Now we can start our algorithm. The input matrixes of
our algorithm are Tabs. 4 and 5 as fuzzy decision matrix
and fuzzy weights of attributes. By applying step 1 on Tab.
3 and considering that attribute 1 and attribute 2 are in-
herently benet oriented and attribute 3 is mainly behaved
as cost oriented one we will have Tab. 6.
Direct result of algorithms step 2 using Tabs. 5 and 6 is
summarized in Tab. 7. Results of step 3 are shown in Tab.
8. Calculating fuzzy distance between each alternative and
FPIS and FNIS (PFD & NFD) is shown in Tab. 9.
Absolute value of UTVs for fuzzy positive distances
and fuzzy negative distances are o.838 and 0.554 respec-
IJMSEM email for subscription: info@msem.org.uk
12 S. Sadi-Nezhad & K. Damghani: A modied TOPSIS technique in presence of uncertainty
Table 9 FPD & FND for each alternative
Fuzzy Positive Distance Fuzzy Negative Distance
Mode 1 (0.559, 0.063, 0.49, 0.98) Mode 1 (0.109, 0.402, 0.916, 1.337)
Mode 2 (0.221, 0.299, 0.818, 1.286) Mode 2 (0.415, 0.075, 0.554, 0.999)
Mode 3 (0.477, 0.04, 0.593, 1.088) Mode 3 (0.217, 0.299, 0.814, 1.255)
Mode 4 (0.167, 0.364, 0.891, 1.372) Mode 4 (0.501, 0.001, 0.49, 0.945)
Mode 5 (0.13, 0.417, 0.958, 1.423) Mode 5 (0.552, 0.065, 0.437, 0.908)
Mode 6 (0.313, 0.217, 0.745, 1.235) Mode 6 (0.364, 0.147, 0.636, 1.091)
Mode 7 (0.094, 0.407, 0.921, 1.337) Mode 7 (0.466, 0.029, 0.447, 0.872)
Mode 8 (0.562, 0.04, 0.522, 1.019) Mode 8 (0.148, 0.371, 0.894, 1.34)
Mode 9 (0.532, 0.037, 0.494, 0.967) Mode 9 (0.096, 0.398, 0.89, 1.31)
Table 10 Transmitted FPD & FND for each alternative
Transmitted Fuzzy Positive
Distance by UTV+0.0001
Transmitted Fuzzy Negative
Distance by UTV+0.0001
Mode 1 (0.003, 0.499, 1.052, 1.542) Mode 1 (0.443, 0.954, 1.468, 1.889)
Mode 2 (0.341, 0.861, 1.38, 1.848) Mode 2 (0.137, 0.626, 1.106, 1.55)
Mode 3 (0.085, 0.602, 1.155, 1.65) Mode 3 (0.334, 0.851, 1.366, 1.806)
Mode 4 (0.395, 0.926, 1.453, 1.934) Mode 4 (0.051, 0.553, 1.041, 1.497)
Mode 5 (0.432, 0.979, 1.52, 1.985) Mode 5 (0, 0.486, 0.988, 1.459)
Mode 6 (0.249, 0.779, 1.307, 1.797) Mode 6 (0.188, 0.699, 1.188, 1.643)
Mode 7 (0.468, 0.969, 1.483, 1.899) Mode 7 (0.086, 0.523, 0.999, 1.424)
Mode 8 (0, 0.522, 1.084, 1.581) Mode 8 (0.404, 0.922, 1.446, 1.892)
Mode 9 (0.03, 0.525, 1.056, 1.529) Mode 9 (0.455, 0.95, 1.442, 1.862)
tively. Application of Transition phase by these UTVs and
= 0.0001, will cause Tab. 10.
Fuzzy Closeness Coecients are calculated as Tab. 11.
Ranking alternatives by Preference Ratio concept in a
relative manner will cause Tab. 12.
Table 11 Fuzzy Closeness Coecients
Mode CC
i
Mode 1 (0.129, 0.379, 1.01, 4.242)
Mode 2 (0.04, 0.252, 0.743, 3.241)
Mode 3 (0.097, 0.337, 0.94, 4.304)
Mode 4 (0.015, 0.222, 0.704, 3.357)
Mode 5 (0, 0.194, 0.675, 3.377)
Mode 6 (0.055, 0.28, 0.804, 3.761)
Mode 7 (0.026, 0.211, 0.67, 2.57)
Mode 8 (0.116, 0.365, 1.001, 4.68)
Mode 9 (0.134, 0.38, 0.977, 3.839)
Table 12 Final ranking of alternatives by preference ratio
Mode 1/K-value K-value Rank
Mode 1 0.914 1.094 3
Mode 2 0.688 1.453 8
Mode 3 0.917 1.091 2
Mode 4 0.705 1.418 6
Mode 5 0.703 1.422 7
Mode 6 0.796 1.256 5
Mode 7 0.549 1.821 9
Mode 8 0.997 1.003 1
Mode 9 0.835 1.198 4
It is notable that the proposed approach has been imple-
mented using MS-Excel 12.0 and VB 6.0 which were applied
to code Fuzzy TOPSIS algorithm and PR methodology, re-
spectively. Fig. 3 represents the interface of developed ap-
plication in VB 6.0 for PR.
7 Conclusion
We proposed a modied TOPSIS method in fuzzy envi-
ronment. In our proposed algorithm several decision makers
had direct eect on fuzzy decision matrix calculation for a
typical MADM problem. Distances were calculated through
an ecient fuzzy distance measure which causes more re-
alistic assumptions. Moreover, Preference Ratio was eec-
tively used for ranking fuzzy closeness coecients of each
alternative at nal steps. The ideal and anti-ideal vectors
were calculated considering real data in favor of supplying a
proper infrastructure for multi-period comparison of alter-
native. We modied preference ratio for manipulating neg-
ative fuzzy numbers by a transition procedure. The overall
properties of proposed procedure made it robust and inter-
esting assessment tools for real life problems. The proposed
algorithm was successfully applied for a real case study of
transportation system modes of Iran. The transportation
system modes problem is a proper testimonial of compli-
cated and real life MADM problem in uncertain situation.
The proposed modied fuzzy TOPSIS algorithm can easily
be re-implemented in other applications.
References
[1] Abbasbandy, S. and Asady, B. (2006). Ranking of fuzzy num-
bers by sign distance. Information Sciences, 176:24052416.
[2] Abo-Sinna, M. and Abou-El-Enien, T. (2006). An interac-
tive algorithm for large scale multiple objective programming
problems with fuzzy parameters through TOPSIS approach.
Applied Mathematics and Computation, 177(2):515527.
[3] Abo-Sinna, M. and Amer, A. (2005). Extensions of TOPSIS
for multi-objective large-scale nonlinear programming prob-
lems. Applied Mathematics and Computation, 162:243256.
[4] Abo-Sinna, M., Amer, A., and Ibrahim, A. (2008). Exten-
sions of TOPSIS for large scale multi-objective non-linear pro-
gramming problems with block angular structure. Applied
Mathematical Modelling, 32(3):292302.
IJMSEM email for contribution: submit@msem.org.uk
International Journal of Management Science and Engineering Management, 6(1): 3-13, 2011 13
[5] Asady, B. and Zendehnam, A. (2007). Ranking fuzzy num-
bers by distance minimization. Applied Mathematical Mod-
elling, 31:25892598.
[6] Athanasopoulos, G., Riba, C., and Athanasopoulou, C.
(2009). A decision support system for coating selection based
on fuzzy logic and multi-criteria decision making. Expert Sys-
tems with Applications, 36:1084810853.
[7] Bentez, J., Martn, J., and Roman, C. (2007). Using fuzzy
number for measuring quality of service in the hotel industry.
Tourism Management, 28:544555.
[8] Chakraborty, C. and Chakraborty, D. (2007). A theoretical
development on a fuzzy distance measure for fuzzy numbers.
Mathematical and Computer Modelling, 43:254261.
[9] Chen, C. (2000). Extensions of the TOPSIS for group deci-
sion making under fuzzy environment. Fuzzy Sets and Sys-
tems, 114:19.
[10] Chen, C., Lin, C., and Huang, S. (2006). A fuzzy approach
for supplier evaluation and selection in supply chain man-
agement. International Journal of Production Economics,
102:289301.
[11] Chen, M. and Tzeng, G. (2004). Combining Grey Relation
and TOPSIS Concepts for Selecting an Expatriate Host Coun-
try. Mathematical and Computer Modeling, 40:14731490.
[12] Chu, T. and Lin, Y. (2009). An interval arithmetic based
fuzzy TOPSIS model. Expert Systems with Applications,
36:1087010876.
[13] Dursun, M. and Karsak, E. (2010). A fuzzy MCDM ap-
proach for personnel selection. Expert Systems with Applica-
tions, 37:43244330.
[14] Gurumurthy, A. and Kodali, R. (2008). A multi-criteria
decision-making model for the justication of lean manufac-
turing systems. International Journal of Management Science
and Engineering Management, 3(2):100118.
[15] Hwang, C. and Yoon, K. (1981). Multiple Attribute Deci-
sion Making. Springer-Verlag, Berlin.
[16] Jadidi, O., Hong, T., and et al. (2008). An optimal grey
based approach based on TOPSIS concepts for supplier selec-
tion problem. International Journal of Management Science
and Engineering Management, 4(2):104117.
[17] Jadidi, O., Hong, T., and Firouzi, F. (2009). TOPSIS ex-
tension for multi-objective supplier selection problem under
price breaks. International Journal of Management Science
and Engineering Management, 4(3):217229.
[18] Jahanshahloo, G., Lot, F., and Izadikhah, M. (2006). Ex-
tension of the TOPSIS method for decision-making problems
with fuzzy data. Applied Mathematics and Computation,
181:15441551.
[19] Kahraman, C., B uy ukozkan, G., and Ates, N. (2007a). A
two phase multi-attribute decision-making approach for new
product introduction. Information Sciences, 177:15671582.
[20] Kahraman, C., C evik, S., and et al. (2007b). Fuzzy multi-
criteria evaluation of industrial robotic systems. Computers
& Industrial Engineering, 52:414433.
[21] Kelemenis, A. and Askounis, D. (2010). A new TOPSIS-
based multi-criteria approach to personnel selection. Expert
Systems with Applications, 37:49995008.
[22] Kuo, M., Tzeng, G., and Huang, W. (2007). Group decision-
making based on concepts of ideal and anti-ideal points in a
fuzzy environment. Mathematical and Computer Modeling,
45:324339.
[23] Li, D. (2007). Compromise ratio method for fuzzy multi-
attribute group decision making. Applied Soft Computing,
7:807817.
[24] Litman, T. (2008). A Good Example of Bad Transporta-
tion Performance Evaluation: A Critique of the Fraser In-
stitute Report, Transportation Performance of the Canadian
Provinces; Victoria Transport Policy Institute; Presented at
the Institute of the Transportation Engineers. Vancouver Is-
land Section, Victoria, British Columbia.
[25] Malekly, H., Mousavi, S., and Hashemi, H. (2010). A fuzzy
integrated methodology for evaluating conceptual bridge de-
sign. Expert Systems with Applications, 37:49104920.
[26] Modarres, M. and Sadi-Nezhad, S. (2001). Ranking fuzzy
numbers by preference ratio. Fuzzy Set and Systems, 118:429
436.
[27] Modarres, M. and Sadi-Nezhad, S. (2005). Fuzzy simple ad-
ditive weighting method by preference ratio. Intelligent Au-
tomation and Soft Computing, 11(4):235244.
[28]

On ut, S., Efendigil, T., and Kara, S. (2010). A combined
fuzzy MCDM approach for selecting shopping center site: An
example from Istanbul, Turkey. Expert Systems with Appli-
cations, 37:19731980.
[29]

On ut, S. and Kara, S. (2008). Transshipment site selection
using the AHP and TOPSIS approaches under fuzzy environ-
ment. Waste Management, 28(9):15521559.
[30]

On ut, S., Kara, S., and Isik, E. (2009). Long term supplier
selection using a combined fuzzy MCDM approach: A case
study for a telecommunication company. Expert Systems with
Applications, 36:38873895.
[31] Sadi-Nezhad, S. and Assadi, R. (2008). Preference ratio-
based maximum operator approximation and its application
in fuzzy ow shop scheduling. Applied Soft Computing, 8:759
766.
[32] Sadi-Nezhad, S. and Damghani, K. (2010). Application of a
fuzzy TOPSIS method base on modied preference ratio and
fuzzy distance measurement in assessment of trac police cen-
ters performance. Applied Soft Computing, 10:10281039.
[33] Secme, N., Bayrakdaroglu, A., and Kahraman, C. (2009).
Fuzzy performance evaluation in Turkish Banking Sector us-
ing Analytic Hierarchy Process and TOPSIS. Expert Systems
with Applications, 36:1169911709.
[34] Tran, L. and Duckstein, L. (2002). Comparison of fuzzy
numbers using a fuzzy distance measure. Fuzzy Sets and Sys-
tems, 130:331341.
[35] Tsaura, S., Chang, T., and Yen, C. (2002). The evaluation
of airline service quality by fuzzy MCDM. Tourism Manage-
ment, 23:107115.
[36] Wang, T. and Chang, T. (2007). Application of TOPSIS in
evaluating initial training aircraft under a fuzzy environment.
Expert Systems with Applications, 33:870880.
[37] Wang, Y. and Elhag, T. (2006). Fuzzy TOPSIS method
based on alpha level sets with an application to bridge risk
assessment. Expert Systems with Applications, 31:309319.
[38] Wang, Y. and Hua, Z. (2007). A note on group decision-
making based on concepts of ideal and anti-ideal points in a
fuzzy environment. Mathematical and Computer Modeling,
46:12561264.
[39] Wang, Y. and Lee, H. (2007). Generalizing TOPSIS for
fuzzy multiple-criteria group decision-making. Computers and
Mathematics with Applications, 53:17621772.
[40] Wu, H., Tzeng, G., and Chen, Y. (2009). A fuzzy MCDM
approach for evaluating banking performance based on bal-
anced scorecard. Expert Systems with Applications, 36:10135
10147.
[41] Wu, J. and Olson, D. (2006). The Method of Grey Re-
lated Analysis to Multiple Attribute Decision Making Prob-
lems with Interval Numbers. Mathematical and Computer
Modeling, 42:991998.
[42] Yang, T., Chen, M., and Hung, C. (2007). Multiple at-
tribute decision-making methods for the dynamic operator al-
location problem. Mathematics and Computers in Simulation,
73:285299.
[43] Yang, T. and Hung, C. (2007). Multiple-attribute decision
making methods for plant layout design problem. Robotics
and Computer-Integrated Manufacturing, 23:126137.
[44] Zimmermann, H. (1991). Fuzzy Set Theory and its Applica-
tions. Kluwer Academic Publishers, Dordrecht, London, 2nd
edition.
IJMSEM email for subscription: info@msem.org.uk

Вам также может понравиться