Information Fusion 28 (2016) 1023

Contents lists available at ScienceDirect

Information Fusion
journal homepage: www.elsevier.com/locate/inffus

Information fusion for intuitionistic fuzzy decision making: An overview

Zeshui Xu , Na Zhao
School of Economics and Management, Southeast University, Nanjing, Jiangsu 211189, China

a r t i c l e

i n f o

Article history:
Received 30 April 2015
Received in revised form 27 June 2015
Accepted 4 July 2015
Available online 11 July 2015
Keywords:
Intuitionistic fuzzy decision making
Information fusion
Attribute weights
Aggregation
Ranking

a b s t r a c t
Intuitionistic fuzzy decision making is to nd the suitable method for ranking alternatives based on the
provided intuitionistic fuzzy information or some related attributes. To date, many studies have focused
on intuitionistic fuzzy decision making problems and various decision making methodologies and
approaches have been proposed. To provide a clear perspective on the information fusion for intuitionistic fuzzy decision making, this paper presents an overview on the existing intuitionistic fuzzy decision
making theories and methods from the perspective of information fusion, involving the determination of
attribute weights, the aggregation of intuitionistic fuzzy information and the ranking of alternatives.
Some potential challenges in future research are meanwhile pointed out. In addition, we provide a survey
of recent applications of the discussed theories and methods in various elds.
2015 Elsevier B.V. All rights reserved.

1. Introduction
Decision making is a fairly common activity in peoples daily
life, which can be seen as a process of ranking alternatives or
selecting the best one(s) from multiple alternatives based on the
provided decision information under the given environment. Due
to the complexity of the socioeconomic system, most decision
making problems involve multiple attributes/indices used to
reect the characteristics or performances of candidate alternatives, which we usually call multi-attribute decision making
(MADM) problems. To date, much attention has been paid to
MADM, and many fruitful research results have been achieved
[1]. There are three major families of methods for MADM: (i) The
utility theory based approaches, in which information fusion techniques are often adopted to synthesize the assessment values of
each alternative under different attributes into an overall value,
and then the alternatives are ranked by comparing the overall values. The weighted sum model based on some aggregation operators [2], the Analytic Hierarchy Process (AHP) method [3] and the
TOPSIS (Technique for Order Preference by Similarity to Ideal
Solution) method [4] are three classical methods in this family.
Besides, the method recently developed based upon penalty functions also belongs to this family [5,6]; (ii) The outranking
approaches, which are used to identify whether the considered
alternative is preferable, incomparable or indifferent to the others
over the attributes. There are two main outranking approaches,
which are the ELECTRE (Elimination et Choice Translating
Corresponding author. Tel.: +86 25 84483382.
E-mail addresses: xuzeshui@263.net (Z. Xu), zhaonawfxy@163.com (N. Zhao).
http://dx.doi.org/10.1016/j.inffus.2015.07.001
1566-2535/ 2015 Elsevier B.V. All rights reserved.

Reality) method [7] and the PROMETHEE (Preference Ranking

Organization Method for Enrichment Evaluation) method [8]; (iii)
The interactive approach, which allows the decision maker to
interact with the analyst or computer (i.e., manmachine
interactive decision making) to update his/her preferences [9].
With the increasing complexity and uncertainty of socioeconomic
environments and the sharp growth of the amount of knowledge
and information, more and more real-life decision making problems cannot be effectively resolved by a single decision maker
for a persons attention, knowledge and experience are limited.
Accordingly, it is needed to gather multiple decision makers with
different knowledge structures and experience to conduct a group
decision making (GDM). Two processes are necessarily implemented to solve GDM problems: the consensus process and the
selection process [10]. The former aims at reaching the maximum
degree of agreement among the decision makers opinions. The
latter encompasses two phases: the aggregation of individual
opinions into a collective opinion by using an information fusion
technique and the exploitation of the collective opinion for ranking
alternatives.
In decision making, uncertainty is ubiquitous since objective
things are uncertain and complex, and the managing and modelling of uncertain information are vital for the acquisition of desirable solutions. The fuzzy set [11] has been found to be a useful tool
to model peoples imprecise decision information, and lots of fuzzy
decision making methods have been put forward [10,12,13].
However, the fuzzy set only involves the membership degree, but
neglects the hesitation and the indeterminacy often involved in
decision making. For example, in a voting event, there is usually
abstention in addition to support and objection [14]. In order

Z. Xu, N. Zhao / Information Fusion 28 (2016) 1023

to fully reect the characteristics of afrmation, negation and hesitation of human cognitive performance, Atanassov [15] extended
the fuzzy set to introduce the intuitionistic fuzzy set (IFS), which
is characterized by a membership function, a non-membership
function and a hesitancy (indeterminacy) function.
As the IFS can express humans imprecise cognitions from the
aspects of afrmation, negation and hesitation, it has been widely
used to describe the imprecise, vague or uncertain preferences of
the decision makers in decision making process. Xu [16] dened
intuitionistic fuzzy numbers (IFNs) (also called intuitionistic fuzzy
values (IFVs) [17,18]) as the basic components of an IFS, and developed a decision making method to help a manufacturing company


to search the best global supplier. In this example, the IFN uij ; v ij
was adopted to evaluate the global supplier Ai , where uij indicates
the degree that the supplier Ai satises the attribute C j ; v ij indicates
the degree that the supplier Ai does not satisfy the attribute C j , and
1  uij  v ij indicates the hesitancy degree that the supplier Ai satises the attribute C j . Hernandez and Uddameri [19] applied the
IFNs to the evaluation of agriculture best management practices
in the coastal semi-arid area of South Texas. In this case study, they
used the IFN qi ; ri to evaluate the importance of the attribute C i ,
where qi and ri denote the degrees of membership and
non-membership of the attribute C i to the fuzzy concept importance, respectively. Wan and Dong [20] developed a novel mathematical programming method for hybrid multi-criteria group
decision making, in which the criteria values are expressed by
IFNs, interval-valued intuitionistic fuzzy numbers, trapezoidal
fuzzy numbers, linguistic variables, interval numbers and real
numbers. Liu et al. [21] constructed a partial binary tree DEA-DA
cyclic classication model for the decision makers in complex
multi-attribute large group interval-valued intuitionistic fuzzy
decision making problems. To depict the preferences of the decision makers for each pair of alternatives comprehensively, Xu
[22] gave a simple notion of intuitionistic fuzzy preference
relations (IFPRs) in which the basic elements are IFNs
uij ; v ij i; j 1; 2; . . . ; n and applied them to evaluate the agroecological regions in Hubei Province, China. Here, uij denotes the
certainty degree to which the agroecological region Ai is preferred
to the agroecological region Aj ; v ij denotes the certainty degree to
which Ai is non-preferred to Aj , and 1  uij  v ij is interpreted as
the hesitancy degree to which Ai is preferred to Aj . Paternain
et al. [23] presented a construction method of IFPRs from fuzzy
preference relations and proposed two novel decision making algorithms by generalizing the weighted voting strategy.
Up to now, a large number of studies have been done on the
intuitionistic fuzzy decision making and a variety of decision making methodologies and approaches have been proposed [14,24,25].
Since the fusion and processing of intuitionistic fuzzy information
are very critical for intuitionistic fuzzy decision making, in this
paper, we shall summarize and analyze the current approaches
to intuitionistic fuzzy decision making from the perspective of
information fusion, involving the determination of attribute
weights, the aggregation of intuitionistic fuzzy information and
the ranking of alternatives. Furthermore, we shall introduce in
depth their recent applications in different elds. Towards these
objectives, we set out the remainder of the paper as follows: In
Section 2, we briey introduce some relevant knowledge about
intuitionistic fuzzy decision making. Section 3 provides a review
of the approaches to derive attribute weights from the intuitionistic fuzzy decision information or from both the intuitionistic fuzzy
decision information and the partially known weight information.
In Section 4, we rst survey the orders of IFNs, which are vital for
the ordered aggregation of intuitionistic fuzzy information, and
then survey the aggregation operators of intuitionistic fuzzy information in different situations. In Section 5, we review four main

11

approaches for ranking alternatives within the intuitionistic fuzzy

decision contexts. Section 6 lists the recent applications of the discussed approaches in various elds. In the last section, we present
some conclusions.
2. Intuitionistic fuzzy decision making
In 1986, Atanassov [15] introduced the intuitionistic fuzzy set
(IFS) to comprehensively portray the uncertainty of human beings
when providing judgments over the objects.
Denition 1 [15]. Let X be a xed set, then an intuitionistic fuzzy
set (IFS) I on X is dened as:

I fhx; uI x; v I xij x 2 X g
where the functions uI : X ! 0; 1 and v I : X ! 0; 1 ascertain the
membership degree uI x and the non-membership degree v I x of
the element x 2 X to the set I, respectively, with the condition:
uI x v I x 6 1.
Usually, pI x 1  uI x  v I x is called the indeterminacy
degree or the hesitancy degree of x to I [17]. a ua ; v a is called
an intuitionistic fuzzy number (IFN) (also called intuitionistic fuzzy
value (IFV) [18]) whose physical interpretation can be presented as
follows: For instance, if ua ; v a 0:5; 0:2, then it can be interpreted as in a presidential election, the vote for a candidate is
50% in favor, 20% against, and 30% abstentions. From the perspective of a voting, the membership degree of an IFS can represent
the approval percentage, the non-membership degree can stand for
the rejection percentage, and the hesitancy degree, reecting the
percentage of voters who are not sure about whether the candidate
is competent as the president, can be regarded as abstention.
During the decision making process, the decision maker is usually required to provide his/her preferences for each pair of alternatives, and then construct a preference relation. Based on the IFNs,
Xu [22] dened an intuitionistic fuzzy preference relation (IFPR)



e e
on the set X fx1 ; x2 ; . . . ; xn g as A
a ij nn , where e
a ij uij ; v ij is
an IFN, for all i; j 1; 2; . . . ; n, and uij denotes the certainty degree
to which the object xi is preferred to the object xj ; v ij denotes the
certainty degree to which xi is non-preferred to xj , and 1  uij  v ij
is interpreted as the indeterminacy degree or hesitancy degree to
which xi is preferred to xj . Furthermore, uij and v ij satisfy the conditions: uji v ij ; v ji uij ; uii v ii 0:5, for all i; j 1; 2; . . . ; n.
It is obvious that the IFN is very useful in modeling the uncertainty and vagueness of objective things for it allocates to each element in a universe a membership degree, a non-membership
degree and a hesitancy degree. Accordingly, more and more
researchers have been applying IFNs to describe the imprecise or
uncertain decision information and dealing with the uncertainty
and vagueness in decision making under different situations.
According to the distinct forms of the decision information, the
current intuitionistic fuzzy decision making can be roughly divided
into two types: The rst type is based on the intuitionistic fuzzy
assessment information provided by the decision makers, which
is expressed by IFNs. The second type is based on the IFPRs provided by the decision makers through pair-wise comparisons of
alternatives. In this paper, we mainly focus our attention on the
rst type for we may refer to the review [26] for the second one.
Firstly, we pay attention to the most common intuitionistic
fuzzy MADM (IF-MADM) problems, which can be mathematically
described as follows: In a MADM problem, let A fA1 ; A2 ; . . . ; An g
be a discrete set of n alternatives, C fC 1 ; C 2 ; . . . ; C m g be the discussion universe containing m attributes, and w w1 ; w2 ; . . . ; wm T be
the weight vector of the attributes, where wj denotes the importance degree of the attribute C j . If the assessment of each

12

Z. Xu, N. Zhao / Information Fusion 28 (2016) 1023

alternative Ai on each attribute C j can be expressed by an IFN



uij ; v ij , where uij indicates the degree that the alternative Ai satises the attribute C j , and v ij indicates the degree that the alternative
Ai does not satisfy the attribute C j , such that uij ; v ij 2 0; 1 and
uij v ij 6 1 for i 1; 2; . . . ; n; j 1; 2; . . . ; m. Then in this case, the
MADM problem can be referred to as an IF-MADM problem, which
can be concisely expressed by the following matrix:

Usually, we call the matrix D an intuitionistic fuzzy decision matrix,

which can be directly given by the decision maker according to
his/her judgments for the degree to which an alternative satises
an attribute and the degree to which the alternative does not satisfy
the attribute. It is worthwhile to point out that during the process of
decision making, if all attributes C j j 1; 2; . . . ; m are of the same
type, then the intuitionistic fuzzy decision matrix does not need
normalization, whereas if there are benet attributes (the bigger
the attribute values the better) and cost attributes (the smaller
the attribute values the better), then in this case, it is necessary to
 
normalize the intuitionistic fuzzy decision matrix D f
dij
into
nm
 
the intuitionistic fuzzy decision matrix M reij nm , where

(
reij

f
dij ; for benefit attribute C j
;
f
dij c ; for cost attribute C j

j 1; 2; . . . ; m



dij c is the complement of f
dij dened as f
dij c v ij ; uij [15].
where f
For convenience, we hereinafter suppose that all intuitionistic fuzzy
decision matrices are normalized without explicitly mentioning
them. In general, the weight information of attributes is in one of
the following cases:
(1) The weight information is completely unknown owing to
time pressure or the complexity and uncertainty of the considered problem [27].
(2) The weight information is partially known due to the inherent subjective nature of human thinking. In this case, the
known weight information H can be expressed by the following forms [28] for i j:


Form 1. A weak ranking: wi P wj ;


Form 2. A strict ranking: wi  wj P di di > 0;


Form 3. A ranking with multiples: wi P di wj 0 6 di 6 1;
Form 4. An interval form: fdi 6 wi 6 di ei g0 6 di 6 di
ei 6 1;


Form 5. A ranking of differences: wi  wj P wk  wl , for
j k l.
(3) The weight information is completely known. The weights
wj j 1; 2; . . . ; m of the attributes C j j 1; 2; . . . ; m are
assigned in advance by the decision maker according to
his/her expertise and experience, which is usually needed
to satisfy the normalization conditions: wj 2 0; 1 j 1;
P
2; . . . ; m and m
j1 wj 1.
In the IF-MADM problems, all the intuitionistic fuzzy assessment information is provided in the same time period or at the
same stage. Nevertheless, in many practical situations, such as
multi-period investment decision making, medical diagnosis, personnel dynamic examination, and military system efciency
dynamic evaluation, the assessment information is usually

provided in different time periods [14]. As a result, the intuitionistic fuzzy dynamic MADM (IF-DMADM) problems were introduced
[29], which can be depicted by the following notations:
(1) A; C and w are dened as above;


(2) t t1 ; t2 ; . . . ; tp is the set of p time periods, whose weight

 T
vector is wt wt 1 ; wt2 ; . . . ; w tp , where wtk 2
Pp
0; 1 k 1; 2; . . . ; p and k1 wt k 1;


(3) Mtl reij t l nm is the intuitionistic fuzzy decision matrix
provided in the time period tl l 1; 2; . . . ; p. Here,


reij tl uij t l ; v ij t l is an IFN, where uij t l indicates the
degree to which the alternative Ai satises the attribute C j
in the period tl , and v ij t l indicates the degree to which
the alternative Ai does not satisfy the attribute C j in the period tl , with uij t l ; v ij t l 2 0; 1 and uij t l v ij t l 6 1 for
i 1; 2; . . . ; n; j 1; 2; . . . ; m.
The aforementioned intuitionistic fuzzy decision making problems involve only one decision maker. However, the increasing complexity of socioeconomic environments makes it more and more
difcult for a single decision maker to consider all aspects of a problem and then make a reasonable decision [30]. Accordingly, it is
required to conduct a GDM. An intuitionistic fuzzy multi-attribute
group decision making (IF-MAGDM) problem encompasses a group
of decision makers fd1 ; d2 ; . . . ; ds g whose importance weight vector
P
is k k1 ; k2 ; . . . ; ks T with kk 2 0; 1k 1; 2; . . . ; s and sk1 ks 1,
and a collection of intuitionistic fuzzy decision matrices

n
o
f
k
k 1; 2; . . . ; s are
M1 ; M 2 ; . . . ; M s , where M k r ij
nm

provided by the decision makers dk k 1; 2; . . . ; s with



f
k
k
k
rij uij ; v ij i 1; 2; . . . ; n; j 1; 2; . . . ; m; k 1; 2; . . . ; s being
the intuitionistic fuzzy assessment values of the alternatives
Ai i 1; 2; . . . ; n under the attributes C j j 1; 2; . . . ; m.
3. Attribute weight derivation approaches
To solve the IF-MADM problems with completely unknown or
partially known weight information on attributes, the rst thing
we usually need to do is to determine the weights of attributes
according to the provided information on the decision problem
to be solved. Clearly, for the rst case, we can only resort to the
given intuitionistic fuzzy decision matrix to obtain the attribute
weights, whereas for the second case, we can also make use of
the known weight information. To date, different attribute weight
derivation methods have been put forward. In the following, let us
briey outline these methods.
3.1. Methods for deriving attribute weights only from intuitionistic
fuzzy decision matrix
3.1.1. The entropy-based method
This method is based on the entropy measures of IFSs, which
are used to measure the uncertainty associated with IFSs. In this
method, the basic principle of determining weights for attributes
is that the smaller the entropy value of the intuitionistic fuzzy
assessment information of alternatives under an attribute, the bigger the weight should be assigned to the attribute; otherwise, the
smaller the weight should be assigned to the attribute. In terms of
this principle, the attribute weights can be determined as follows
[31]:

Model 1

wj

1  Ej
P
;
m m
j1 Ej

j 1; 2; . . . ; m

13

Z. Xu, N. Zhao / Information Fusion 28 (2016) 1023



Ai ; uij ; v ij Ai 2 A

where Ej is the entropy value of the IFS Ij

 

P 
dened as Ej E Ij 1n ni1 1  uij  v ij for j 1; 2; . . . ; m.
Obviously, besides the entropy measure E, we may resort to
other entropy formulas in Model 1. (For the entropy measures of
IFSs, please refer to Ref. [32].) To observe the ranking outcomes
of the attribute weights generated by applying different entropy
measures of IFSs, Chen and Li [33] conducted a computational
experiment with simulation data, which illustrates that the rankings of the attribute weights are not only affected by the adopted
entropy measures for IFSs, but also affected by the number of attributes and the number of alternatives.
3.1.2. The deviation-based method
In this method, the basic principle of determining weights for
attributes is that the more similar the intuitionistic fuzzy assessment values of all alternatives with respect to an attribute, the
smaller the weight should be allocated to the attribute; otherwise,
the bigger the weight should be allocated to the attribute. Based on
this principle, the following nonlinear programming model was
established [27], whose objective is to maximize all the deviation
values of intuitionistic fuzzy assessment values of all alternatives
under all attributes:

Model 2

Max

m X
n X
X



d reij ; rf
kj wj

m
X
w2j 1;

wj P 0;



Pn P
e f
ki d r ij ; r kj
i1
wj q


Pm Pn P
e f 2
j1
i1
ki d r ij ; r kj

j1

wj P 0;

j1

j 1; 2; . . . ; m

j1

By linear equal weighted summation method [35], the above

model can be transformed into a single objective optimization
model:

Model 3

Min f w

m X
n


X
2
w2j d reij ; f
r j
j1 i1

m
X
wj 1;
s:t:

wj P 0;

j 1; 2; . . . ; m

j1

By constructing the Lagrange function, we get

wj 0
@

Pm

j1

Pn

1

d2

reij ; re
j

1
1
P


2
n
A
rj
reij ; f
i1 d

j 1; 2; . . . ; m



where d reij ; rf
is the distance between reij and rf
kj
kj dened as
 


 1 

 

d reij ; rf
kj 2 uij  ukj v ij  v kj .
By constructing the Lagrange function, we obtain

Max

j1
m
X
wj 1;
s:t:


 r
h
i
2
12 uij  1 v 2ij .
r
where for i 1; 2; . . . ; n; j 1; 2; . . . ; m, d reij ; f
j

j1

Model 4

Min f w

m
m
m

 X




X
X
2
2
2

f
f
f
;
;...;
w2j d rf
w2j d rf
w2j d rf
1j ; r j
2j ; r j
nj ; r j

i1

j1 i1 ki

s:t:

j 1; 2; . . . ; m. Considering that the smaller the distance between

each alternative and the IFPIS, the better the alternative, Xu [34]
established the following multiple-objective programming model:

3.1.4. The consistency-based method

This method is based on the consistency and inconsistency
indices of the decision maker, which are dened based on the preferences between alternatives given by the decision maker. The
main idea of this method is to maximize the consistency index of
the decision maker under the condition that the consistency index
should be greater than the inconsistency index. In light of this idea,
the following linear programming model was constructed [36]:

kki

k;i2X

8P
 
 

h



i
P
m
>
u2ij  u2kj v 2ij  v 2kj p2ij  p2kj 2 uij  ukj 2 v ij  v kj
>
k;i2X
j1 wj
>
>
>
 


 Pm
 



P
P
P
>
>
>
 m
 j1 v j k;i2X 2 uij  ukj 4 v ij  v kj P 2h;
>
j1 uj
k;i2X 4 uij  ukj 2 v ij  v kj
>
>






h
i
>




P
Pm
>
>
>
u2kj  u2ij v 2kj  v 2ij p2kj  p2ij 2 ukj  uij 2 v kj  v ij
>
j1 wj
k;i2X
>
< P
 


 Pm
 



P
P
 m
 j1 v j k;i2X 2 ukj  uij 4 v kj  v ij 2kki P 0; k; i 2 X;
s:t:
k;i2X 4 ukj  uij 2 v kj  v ij
j1 uj
>
>
> kki P 0; k; i 2 X;
>
>
>
>
> u ; v P 0; u v 6 w ; j 1; 2; . . . ; m;
>
j
j
j
j
j
>
>
>
> w P e; j 1; 2; . . . ; m;
>
j
>
>
>
: Pm w 1
j1

After normalization, we get



Pn P
d reij ; rf
kj


wj Pm i1
Pn ki
P
e f
ki d r ij ; r kj
j1
i1
In such a case, we have wj 2 0; 1 for j 1; 2; . . . ; m, and
Pm
j1 wj 1.
3.1.3. The ideal-solution-based method
This method is based on the intuitionistic fuzzy positive ideal



f
solution (IFPIS) A f
with f
r
r
for
1 ; r2 ; . . . ; rm
j 1; 0

where h > 0 and e > 0 are the given values, X

fk; ijAi  Ak ; k; i 1; 2; . . . ; ng is the pairwise comparison preference relations of alternatives given by the decision maker. Here,
Ai  Ak means that either the decision maker prefers the alternative Ak to Ai or the decision maker is indifferent between Ai and Ak .
3.1.5. The group consensus-based method
This method is suitable for the group decision making situations. The main idea of this method is to maximize the group consensus, based on which the following nonlinear optimization
model was developed [37]:

14

Z. Xu, N. Zhao / Information Fusion 28 (2016) 1023

0
!2
s X
m X
n
s
X
X
l
@ uk 
Min
k
u

l ij
ij

Model 5

k1 j1 i1

k
ij

s
X
l

kl v ij

l1

s
X
k
l
kl pij
pij 

!2

1<p62

and

m
X
s:t:
wj 1; wj P 0; j 1; 2; . . . ; m



CE reij ; relj

j1

where k k1 ; k2 ; . . . ; ks T is the importance weight vector of s deciP

sion makers with kk 2 0; 1 for k 1; 2; . . . ; s, and sk1 ks 1, and


f
k
k
k
for i 1; 2; . . . ; n; j 1; 2; . . . ; m, rij uij ; v ij is the intuitionistic

Ps Pn
k1

wj

Ps

k1



i1

Ps

k
u 
ij

l
ku
l1 l ij

v k 
ij

Ps
l1

kl v

l
ij

2 

Ps

pk 
ij

l1

p
lj

!p

p
ij

2
p
lj

p p
2

!p

p
ij

p p

v pij v plj
p
lj

2
!p )

1<p62

kl p

l
ij

2


2 
2 
2
P
P
P
k
l
k
l
k
l
uij  sl1 kl uij
v ij  sl1 kl v ij
pij  sl1 kl pij
i1

In this subsection, we introduce how to derive the attribute

weights according to both the intuitionistic fuzzy decision matrix
and the known weight information.
3.2.1. The entropy-valued method
This method is based on the intuitionistic fuzzy entropy measures, which are used to describe the fuzziness and intuitionism
of IFSs. Considering that the greater the entropy value, the more
uncertain the decision information, Wu and Zhang [38] developed
the following programming model, which aims at minimizing the
entropy values of the intuitionistic fuzzy assessment information
of all alternatives under all attributes:

Model 6

Min Ew

i1 j1

j1

 
where g is a small positive real number, and E reij is the
 
1
dened
by
E reij pij  ln 2
reij
entropy
value
of
h



i
u
v
uij ln u ijv v ij ln u ijv .
ij

ij

intuitionistic fuzzy assessment values of all alternatives under all

attributes, based on which the following linear programming
model was developed [40]:

Model 8

Max Dw

j1 i1 k1

s:t: w w1 ; w2 ; . . . ; wm T 2 H;
m
X
wj 1; wj P 0; j 1; 2; . . . ; m
j1



rf
for i; k 1; 2; . . . ; n,
kj ukj ; v kj


i k; j 1; 2; . . . ; m, and d reij ; rf
kj is the distance between reij and
 


 1 

 

f
rf
kj dened as d reij ; r
kj 2 uij  ukj v ij  v kj .
where

reij uij ; v ij

ij

j1 i1

n


1 X
CE reij ; relj
n  1 l1

and

j 1; 2; . . . ; m as the intuitionistic fuzzy positive ideal solution

(IFPIS) and the intuitionistic fuzzy negative ideal solution (IFNIS),
respectively. Then, based on the idea of TOPSIS method [4] that
the chosen alternative should have the shortest distance from
the IFPIS and the farthest distance from the IFNIS, the following
multi-objective optimization model was developed [41]:

Max qz1 w; qz2 w; . . . ; qzn w

Model 9

Obviously, Model 6 only considers the entropy information of

attributes. Xia and Xu [39] proposed the following model to determine the attribute weights by using the intuitionistic fuzzy
entropy and cross entropy measures:
m X
n
X

 
Max Ew
wj 1  E reij

m X
n X
n
X


d reij ; rf
kj wj

3.2.3. The closeness to ideal solution method




f
with f
r
r
Xu [41] suggested A f
1 ; r2 ; . . . ; rm
j 1; 0 for
  




f
j 1; 2; . . . ; m, and A f
for
r1 ; f
r2 ; . . . ; rf
r
with
m
j 0; 1

n X
m
X
 
wj E reij

s:t: w w1 ; w2 ; . . . ; wm T 2 H;
m
X
wj 1; wj P g; j 1; 2; . . . ; m

Model 7

p
ij

upij uplj

Pn

3.2. Methods for deriving attribute weights from partially known

weight information

ij

12

1p

3.2.2. The maximizing deviation method

This method is based on the distance measures of IFSs. The main
idea of this method is to maximize all the deviation values of

2 

( p
uij uplj

fuzzy assessment value of the alternative Ai under the attribute C j

provided by the decision maker dk .
By constructing the Lagrange function, we derive

j1

l1

!2 1
Aw2

l1

Pm

(
p 
p )
p 
1  uij 1  v ij
2  uij  v ij
;

2
2
1  21p

 
E reij 1

s:t: w w1 ; w2 ; . . . ; wm T 2 H;
m
X
wj 1; wj P 0; j 1; 2; . . . ; m
j1

 


where E reij is the entropy value of reij and CE reij ; relj is the cross
entropy between reij and relj , dened as:

s:t: w w1 ; w2 ; . . . ; wm T 2 H;
m
X
wj 1; wj P 0; j 1; 2; . . . ; m
j1

Here,

qzi w
where

dzi w; z w
dzi w; z w dzi w; z w

zi w

P
m

j1 wj uij ;

Pm

j1 wj

v ij

for

1
i 1; 2; . . . ; n; z w

0; 1 and z w 1; 0, and d is the distance measure between

 
 


 1 

 
 
 for
IFNs dened as d reij ; rf
kj 2 uij  ukj v ij  v kj pij  pkj




reij uij ; v ij and rf
kj ukj ; v kj .

15

Z. Xu, N. Zhao / Information Fusion 28 (2016) 1023

By the linear equal weighted summation method [35], Model 9

can be transformed into a single objective optimization model:

Model 10

Max

n
X



j1 wj uij v ij
Pm
1 j1 wj v ij

Pm

i1

s:t: w w1 ; w2 ; . . . ; wm 2 H;
m
X
wj 1; wj P 0; j 1; 2; . . . ; m
j1




A re1 ; re2 ; . . . ; rf
with
rej
m



 e

e
f
maxi uij ; mini v ij for j 1; 2; . . . ; m and A r 1 ; r 2 ; . . . ; rm with


rej mini uij ; maxi v ij for j 1; 2; . . . ; m as the IFPIS and the
Wei

[42]

considered

IFNIS, respectively. Based on the idea that the chosen alternative

should have the largest degree of grey relation from the IFPIS
and the smallest degree of grey relation from the IFNIS, the following single objective optimization model was constructed [42]:

Model 11

n X
m 

X
Min
nij  nij wj
i1 j1

s:t: w w1 ; w2 ; . . . ; wm T 2 H;
m
X
wj 1; wj P 0; j 1; 2; . . . ; m
j1

where




min16i6n min16j6m d reij ; rej 0:5max16i6n max16j6m d reij ; rej




nij
and
d reij ; rej 0:5max16i6n max16j6m d reij ; rej




min16i6n min16j6m d reij ; rej 0:5max16i6n max16j6m d reij ; rej





nij
d reij ; re 0:5max16i6n max16j6m d reij ; re
j

3.2.4. The maximizing overall score method

The main idea of this method is to maximize the overall score
value of each alternative, based on which the following optimization model was developed [16]:

Max si w

m
X
wj sij
j1

s:t: w w1 ; w2 ; . . . ; wm T 2 H;
m
X
wj 1; wj P 0; j 1; 2; . . . ; m
j1



where sij uij  v ij is the score of reij uij ; v ij .
By solving Model 12, we can get the optimal solution

T
i
i
i
i
corresponding to the alternative Ai . In
w w1 ; w2 ; . . . ; wm
general, in the process of determining the attribute weight vector,
it is required to consider all alternatives as a whole. Therefore, a
combined weight vector can be constructed [16]:

w x1 w1 x2
2 1
2
w1
w1
6 1
2
6w
w2
6 2
6
..
6 ..
4 .
.
1

wm
where,

wm

wm

w1

w2
..
.

wm

3
n
   w1
n 7
   w2 7
7
.. 7
..
7
. 5
.
n

   wm

and x x1 ; x2 ; . . . ; xn T is the normalized eigenvector of the


T 

 
matrix ST W
ST W ; S sij nm with sij uij  v ij is the score



matrix of the intuitionistic fuzzy decision matrix M uij ; v ij nm .
For the case in which the known weight information is
expressed by interval forms, i.e., for j 1; 2; . . . ; m; dj 6
wj 6 dj ej , where 0 6 dj 6 dj ej 6 1, Li [43] constructed the following linear programming model:

Max zi

m
X

bij wj

j1

8
>
< uij 6 bij 6 1  v ij ; i 1; 2; . . . ; n;
dj 6 wj 6 dj ej ; j 1; 2; . . . ; m
s:t:
>
: Pm w 1
j
j1

j 1; 2; . . . ; m

P
for each i 1; 2; . . . ; n. Here, Li [43] assumed that m
j1 dj 6 1 and
Pm
d

e
P
1
in
order
to
nd
the
weights
w
2

0;
1 satisfying
j
j
j
j1
Pm
dj 6 wj 6 dj ej and
w

1.
Then,
he
transformed
the above
j
j1
model into the following form:

Model 13

Max z
(
s:t:

n X
m
X



1  uij  v ij wj =n

i1 j1

dj 6 wj 6 dj ej ;
Pm
j1 wj 1

j 1; 2; . . . ; m

are the grey relational coefcients of each alternative from IFPIS and
IFNIS, respectively, d is the distance measure between IFNs dened
 


 1 



 

and
as d reij ; rf
for reij uij ; v ij
kj 2 uij  ukj v ij  v kj


rf

u
;
v
,
for
i

1;
2;
.
.
.
;
n;
j

1;
2;
.
.
.
;
m.
kj
kj
kj

Model 12

w1
6 1
6w
6 2
W 6
6 ..
4 .
1

w2 . . . xn wn
3
1
n 0
   w1
x1
7
n B
x2 C
   w2 7
C
7B
B . C Wx
7
C
..
.. 7B
.
.
. 5@ . A
n

   wm

xn

3.2.5. The interactive method

In many real decision making processes, a decision maker often
needs to interact with group members by providing and modifying
his/her preference information gradually. Xu [41] proposed the following interactive process, which is realized by revising the satisfactory degrees of alternatives until an optimum satisfactory
solution is achieved.
Firstly, based on Model 9, the following optimization model was
constructed [41]:

Max s
8
>
< qzi w P s; i 1; 2; . . . ; n
w w1 ; w2 ; . . . ; wm T 2 H
s:t:
>
: Pm
wj P 0; j 1; 2; . . . ; m
j1 wj 1;
where s mini qzi w, and qzi w is dened as Eq. (1). From the
above model, the original optimal weight vector w0

T
0
0
0
w1 ; w2 ; . . . ; wm
can be got, and then the corresponding satis  0 
faction degree q zi w
of the alternative Ai can be calculated
by Eq. (1). In the course of decision making, the decision maker
0

provides the lower bound si i 1; 2; . . . ; n of the satisfaction

degree of the alternative Ai i 1; 2; . . . ; n according to
 

q zi w0 i 1; 2; . . . ; n. Then, the following optimization model
was developed [41]:

Model 14

Max

n
X
si
i1

8
0
>
>
< qzi w P si P si ; i 1; 2; . . . ; n
s:t:
w w1 ; w2 ; . . . ; wm T 2 H
>
>
: Pm w 1; w P 0; j 1; 2; . . . ; m
j
j
j1

16

Z. Xu, N. Zhao / Information Fusion 28 (2016) 1023


T
t
t
t
from which the weight vector wt w1 ; w2 ; . . . ; wm
can be
  t 
derived, and then the corresponding satisfaction degree q zi w
of the alternative Ai can be calculated by Eq. (1). In this case, if
the decision maker is satised with the result, then we end the
interactive process and the nal weight vector is wt

T
t
t
t
w1 ; w2 ; . . . ; wm ; otherwise, the decision maker needs to recont1

t P 1; i 1; 2; . . . ; n of the satisfaction
sider the lower bound si
degrees of the alternatives Ai i 1; 2; . . . ; n until the optimal solution is obtained, and let t : t 1 and return to Model 14.
To facilitate the reading and comparison, the abovementioned
attribute weight determination models are summarized in Table 1.
In the above, we have overviewed the existing main methods to
derive attribute weights from the intuitionistic fuzzy decision
information or from both the intuitionistic fuzzy decision information and the partially known weight information. However, different methods have different characteristics. For example, as we can
see, Models 1, 6 and 7 all involve the entropy measures of intuitionistic fuzzy information. The entropy measure used in Model
1 only captures the intuitionism, while the one used in Model 7
considers not only the intuitionism, but also the fuzziness of an
IFN. Furthermore, it is worth noting that when the IFN 0; 0 is
involved in Model 6, it is meaningless. Models 2, 3, 8, 9, 11 and
14 all involve the distance measures of intuitionistic fuzzy information. The distance measures used in Models 2, 3, 8 and 11 only
consider the membership degree information and the
non-membership degree information and omit the hesitancy
degree information of IFNs, whereas the ones used in Models 9

and 14 consider all the three aspects. Although some comparison

results of various attribute weight determination approaches have
been presented in Table 1, different approaches have different
advantages and drawbacks. The existing studies do not make sufcient comparisons among these approaches. Therefore, in future
research, it is necessary to conduct a detailed comparative study
to adequately evaluate and compare the advantages and drawbacks of different attribute weight determination approaches or
to develop some software tools to carry out the evaluation and
comparison in order to provide the decision making practitioners
some references to nd out which ones are more suitable for the
problem to be solved.
4. Intuitionistic fuzzy aggregation techniques
After determining the weights of attributes, the next thing we
usually do in solving the IF-MADM problems is to select a suitable
information fusion tool to incorporate the intuitionistic fuzzy decision information and the attribute weight information. Considering
that the orders of IFNs play an important role in the ordered aggregation of intuitionistic fuzzy information, we rst conduct a survey
on the orders of IFNs.
4.1. Orders of intuitionistic fuzzy information
The orders of IFNs can be roughly divided into two categories.
The rst category is dened just by a function from which we
can calculate a value for any IFN and then compare the IFNs

Table 1
Different attribute weight determination models for intuitionistic fuzzy decision making.
Based-information

Models

Papers

Based-techniques

Optimization models

Characteristics

Intuitionistic fuzzy
decision information

Model 1

Ye [31]

Entropy measure

Model 2

Xu [27]

Distance measure

A nonlinear programming model

Model 3

Xu [34]

Distance measure

A nonlinear programming model

Model 4

Li et al. [36]

Distance measure

A linear programming model

Model 5

Xu and Cai [37]

Simple additive
weighting method

A nonlinear optimization model

Different entropy measures produce

different rankings of attribute weights
Focusing on the divergence between
alternatives
Only considering the IFPIS, but neglecting
the IFNIS
Considering the decision makers pairwise
comparison preference for alternatives
Suitable for group decision making

Model 6

Wu and Zhang [38]

Entropy measure

A linear programming model

Model 7

Xia and Xu [39]

Entropy and cross

entropy measures

A linear programming model

Model 8

Wei [40]

Distance measure

A nonlinear programming model

Model 10

Xu [41]

Distance measure

A fractional programming model

Model 11

Wei [42]

Grey relational
coefcient

A linear programming model

Model 12

Xu [16]

Score function

A linear programming model

Model 13

Li [43]

Simple additive
weighting method

A linear programming model

Model 14

Xu [41]

Distance measure,
interactive mechanism

A nonlinear programming model

Intuitionistic fuzzy
decision information,
partially known
weight information

Note: denotes that there is no optimization models involved in the discussed model.

Focusing on the entropy information of

attributes
Considering both the entropy information
of attributes and the divergence between
alternatives
Generating a loss of information when
zero is involved in the process of
aggregation
Considering the distance of each
alternative to the IFPIS and the IFNIS
simultaneously
Considering the degrees of grey relation of
each alternative from the IFPIS and the
IFNIS simultaneously
The intuitionistic fuzzy decision matrix is
transformed into a real-valued decision
matrix, which may lose some original
information
The known weight information is
expressed by interval forms; The
intuitionistic fuzzy decision matrix is
transformed into an interval-valued
decision matrix, which may lose some
original information
Allowing the decision maker to modify the
lower bounds of satisfaction degrees of
alternatives

Z. Xu, N. Zhao / Information Fusion 28 (2016) 1023

according to the values. Chen and Tan [44] rstly developed a score
function for the IFN a ua ; v a : sa ua  v a . Obviously, the
score of a is directly related to the deviation between ua and v a ,
and the bigger the score of a, the larger the IFN a is. However,
Chen and Tans score function cannot efciently compare the
IFNs with the same score. Therefore, a series of modied score
functions have been subsequently put forward. Chen [45] provided
a comparative analysis of different score functions.
The second category is dened by means of two functions. The
most famous one of this category is Xu and Yager [17]s order
dened by adding the accuracy function [46], shown as follows:


Denition 2 [17]. Let ai uai ; v ai ; i 1; 2 be two IFNs,
sai uai  v ai be the score of ai , and hai uai v ai be the
accuracy degree of ai , then
(1) If sa1 < sa2 , then a1 is smaller than a2 , denoted by
a1 XY a2 .
(2) If sa1 sa2 , then

If ha1 ha2 , then a1 and a2 represent the same

information, denoted by a1 XY a2 .

If ha1 < ha2 , then a1 is smaller than a2 , denoted by

a1 XY a2 .

If ha1 > ha2 , then a1 is greater than a2 , denoted by

a2 XY a1 .
From Denition 2, we see that for two IFNs, if the score function
cannot differentiate them, the accuracy function is considered as
the second criterion, i.e., the more accuracy an IFN, the larger the
IFN is. Moreover, let XY XY [ XY , then we have

a1 XY a2 () ua1  v a1 < ua2  v a2



_ ua1  v a1 ua2  v a2 ^ ua1 v a1 6 ua2 v a2



where ai uai ; v ai ; i 1; 2.
Bustince et al. [47] claimed that the order XY is an admissible
order, i.e., XY is a linear order and for any two IFNs a1 and a2 , if
a1 62 a2 , then it holds that a1  XY a2 . Here,
the partial order 62 is
 

a1 62 a2 () ua1 6 ua2 ^ v a1 P v a2 ,
where
dened
as


ai uai ; v ai ; i 1; 2. Furthermore, motivated by the lexicographical order of points, Bustince et al. [47] presented the following two
admissible orders:

a1 Lex1 a2 () ua1 < ua2 _ ua1 ua2 ^ v a1 P v a2


 

a1 Lex2 a2 () v a1 > v a2 _ v a1 v a2 ^ ua1 6 ua2



where ai uai ; v ai ; i 1; 2. Additionally, Bustince et al. [47] and
Miguel et al. [48] investigated the methods of constructing admissible orders in terms of two aggregation functions. The important
roles of admissible orders in different aspects, such as in the construction of the ordered weighted aggregation operators [49,50]
and in the selection of alternatives in decision making [50,51], were
also investigated.
4.2. Aggregation operators for intuitionistic fuzzy information
At present, a lot of studies have focused on the aggregation of
intuitionistic fuzzy information, and a variety of aggregation operators have been developed. Since the operational laws of IFNs are
very vital for the denition of aggregation operators for intuitionistic fuzzy information, we here rst introduce the existing main
operational laws of IFNs.




Given three IFNs a ua ; v a ; a1 ua1 ; v a1 and a2 ua2 ; v a2 ,
some basic operations developed by Xu and Yager [17,18] are presented as follows:

17



(1) a1 ^ a2 minfua1 ; ua2 g; maxfv a1 ; v a2 g ,


(2) a1 _ a2 maxfua1 ; ua2 g; minfv a1 ; v a2 g ,


(3) a1  a2 ua1 ua2  ua1 ua2 ; v a1 v a2 ,


(4) ka 1  1  ua k ; v ka ; k > 0,


(5) a1 a2 ua1 ua2 ; v a1 v a2  v a1 v a2 ,


(6) ak uka ; 1  1  v a k ; k > 0.
Later, based on Archimedean t-norm and t-conorm [52],
Beliakov et al. [53] dened the generalized sum and the generalized multiplication as:


(7) a1  a2 Sua1 ; ua2 ; Tv a1 ; v a2


1
h hua1 hua2 ; g 1 gv a1 gv a2 ,


1
(8) ka h khua ; g 1 kgv a ; k > 0,
where the function T is an Archimedean t-norm and the function S
is an Archimedean t-conorm. Klir and Yuan [52] pointed out that an
additive generator of a continuous Archimedean t-norm is a strictly
decreasing function g : 0; 1 ! 0; 1, such that g1 0. Here,
ht g1  t.
Based on the operational laws (5) and (6), Xia et al. [54] dened
the following generalized forms:


(9) a1 a2 Tua1 ; ua2 ; Sv a1 ; v a2


1
g 1 gua1 gua2 ; h hv a1 hv a2 ,


1
(10) ak g 1 kgua ; h khv a ; k > 0.
It can be noted that let gt  logt, then the operational laws
(7)(10) become (3)(6) accordingly.
Xia and Xu [55] found that the above operations are only based
on the original information, and thus cannot reduce the uncertainty of the aggregated IFNs. Hence, they introduced some point
aggregation operators to reduce the uncertainty degree of IFNs,
which is necessary in some situations [56]:
(11)
(12)
(13)
(14)
(15)
(16)
(17)
(18)

Dja a ua ja pa ; v a 1  ja pa ,
F ja ;ka a ua ja pa ; v a ka pa , where ja ka 6 1,
Hja ;ka a ja ua ; v a ka pa ,
Hja ;ka a ja ua ; v a ka 1  ja ua  v a ,
J ja ;ka a ua ja pa ; ka v a ,
J ja ;ka a ua ja 1  ua  ka v a ; ka v a ,
P ja ;ka a maxja ; ua ; minka ; v a , where ja ka 6 1,
Q ja ;ka a minja ; ua ; maxka ; v a , where ja ka 6 1,

where ja ; ka 2 0; 1.




Let a1 ua1 ; v a1 with ua1 0 and a2 ua2 ; v a2 with ua2 0,
then by the operation (5), we can get ua1 a2 0, which illustrates
that ua2 is not accounted for at all. He et al. [57] noted that it is
an undesirable feature for an averaging operation, and dened
some new operational laws by considering the interactions
between the membership functions and the non-membership
functions of different IFSs:

 
 

1  v a1  1  v a2  1  ua1 v a1



 

 1  ua2 v a2 ; 1  1  v a1  1  v a2 ,


(20) ak 1  v a k  1  ua v a k ; 1  1  v a k ; k > 0.

^ a2
(19) a1

Considering that in real MADM, the attribute weights may be

expressed by IFNs, Li and He [58] dened the following new operational laws:

18

Z. Xu, N. Zhao / Information Fusion 28 (2016) 1023



(21) ka 1  1  ua r ; v a1q ,


q
r
(22) ak u1
a ; 1  1  v a ,
where k r; q is an IFN.
Based upon the above-mentioned operations on IFNs, a large
number of aggregation operators have been developed for fusing
the intuitionistic fuzzy information. From the viewpoint of the relationships between the aggregated IFNs, the existing aggregation
operators for intuitionistic fuzzy information are mainly classied
into four categories: the ones for aggregating the independent
IFNs, the ones for aggregating the correlative IFNs, the ones for
aggregating the IFNs when the attributes are in different priority
levels, and the ones for aggregating the IFNs in different time periods. For the convenience of reading and understanding, we summarize the existing main aggregation operators for intuitionistic fuzzy
information in Table 2.
It can be seen from Table 2 that a variety of aggregation operators
have been designed for fusing the intuitionistic fuzzy information in
different situations. In the following, we discuss their characteristics
and relationships. Among the aggregation operators for the independent IFNs, the IFWA, IFWG, IFOWA, IFOWG, IFHWA, IFHWG, GIFWA,
GIFOWA, GIFHWA, ATS-IFWA and ATS-IFWG operators possess the
desirable properties of idempotency, boundedness, and monotonicity, while the IFHA, IFHG and GIFHA operators do not satisfy some
basic properties such as idempotency and boundedness. Moreover,
from the ATS-IFWA and ATS-IFWG operators, some known operators
can be derived, such as the IFWA, IFWG, EIFWA and EIFWG operators [69]. The GIFPWA, GIFPOWA and GIFPHA operators can reduce
the uncertainty of the aggregated IFNs, which is necessary in some
intuitionistic fuzzy decision situations [56]. Additionally, the
IFWGIA, IFOWGIA and IFHGIA operators take into account of the
interactions between the non-membership degrees and the membership degrees of different IFNs.

Among the aggregation operators for the correlative IFNs, the

IFCA and IFCG operators consider not only the importance of the
aggregated IFNs or their ordered positions but also the correlations
of the IFNs or their ordered positions. The weighting vectors of the
IFPWA and IFPWG operators depend upon the aggregated IFNs and
allow the IFNs to support and reinforce each other. Furthermore,
although the WIFBM, GIFWBGM and IFWGBM are all developed
based on the Bonferroni mean [81] and capable of capturing the
expressed interrelationship of the aggregated IFNs, the WIFBM
can only model the situation where there exist correlations
between any two aggregated IFNs, but not the situation where
there exist connections among any three aggregated IFNs, and it
cannot reduce to the IFBM [75] when the weights of the aggregated
IFNs are the same. The GIFWBGM overcomes the drawbacks of the
WIFBM, while it just considers the whole geometric correlation of
any three aggregated IFNs and cannot reect the interrelationship
between the individual IFN and other IFNs. The IFWGBM avoids all
these shortcomings. In addition, it should be noted that the
Bonferroni mean or the Heronian mean based intuitionistic fuzzy
aggregation operators focus on the aggregated IFNs, while the
Choquet integral or the power average based intuitionistic fuzzy
aggregation operators focus on changing the weighting vector of
the aggregation operator. The correlations of the aggregated IFNs
in the Choquet integral or the Bonferroni mean or the Heronian
mean based the intuitionistic fuzzy aggregation operators are
determined subjectively by the decision makers, while the power
average based the intuitionistic fuzzy aggregation operators
determine the weighting vector by means of the aggregated IFNs
objectively.
Although there are several series of aggregation operators for
fusing the intuitionistic fuzzy decision information when the attributes are in different priority levels and each can well model the
prioritization relationship among the attributes by means of the
priority weights determined by the satisfaction of the attributes

Table 2
Various aggregation operators for intuitionistic fuzzy information.
Relationships between the
aggregated IFNs

Operators

Based-operations

Based-operators

Similar studies

Independent

IFWA operator [18], IFOWA operator [18], IFHA

operator [18], IFHWA operator [59]
GIFWA operator [60], GIFOWA operator [60],
GIFHA operator [60], GIFHWA operator [59]
IFWG operator [17], IFOWG operator [17], IFHG
operator [17], IFHWG operator [59]
ATS-IFWA operator [54]
ATS-IFWG operator [54]
GIFPWA operator [55], GIFPOWA operator [55],
GIFPHA operator [55]
IFWGIA operator [57], IFOWGIA operator [57],
IFHGIA operator [57]

(3), (4)

WA operator [61]

Yager [66]

(3), (4), (6)

WG operator [62]

Li [67]

(5), (6)

OWA operator [63]

Zeng and Su [68]

(7), (8)
(9), (10)
(11)(18)

OWG operator [64]

HWA operator [2]
GWA operator [60]

Wang and Liu [69]

Su et al. [70]
Yang and Chen [71]

(19), (20)

GOWA operator [65]

IFCA operator [72]

IFCG operator [73]
IFPWA operator [74]
IFPWG operator [74]
WIFBM [75], GIFWBGM [76], IFWGBM [77]
IFGWHM operator [78]

(3)
(5)
(3), (4)
(5), (6)
(3)(6)
(3)(6)

Choquet integral [79]

IFPOWA operator [88]

IFPWA operator [89]
IFPWG operator [89]
GIFPWG operator [90]
PIFA operator [91]
IFPRI-OR operator [58]
IFPRI-AND operator [58]

(3), (4)
(3), (4)
(5), (6)
(9), (10)
(1), (2), (3), (5)
(21)
(22)

POWA operator [92]

PA operator [93]

Prioritized or operator [93]

Prioritized and operator [93]

DIFWA operator [29]

DIFWG [94]

(3), (4)
(5), (6)

IFWA operator [18]

IFWG operator [17]

Correlative

At different priority levels

At different time periods

Power average operator [80]

Bonferroni mean [81]
Heronian mean [82]

Wu et al. [83]
Tan and Chen [84]
Zhou et al. [85]
Zhang [86]
Beliakov and James [87]

Z. Xu, N. Zhao / Information Fusion 28 (2016) 1023

in all higher priority levels. However, the generated priority

weights in the IFPOWA, IFPWA, IFPWG and GIFPWG operators
are all real numbers, while those in the PIFA, IFPRI-AND and
IFPRI-OR operators are all expressed by IFNs, which can reect
the intuitionistic fuzzy characteristic of the IF-MADM problems.
Moreover, in the IFPOWA, IFPWA, IFPWG and GIFPWG operators,
the given intuitionistic fuzzy decision information is not fully utilized during the process of inducing the priority weights and just
transformed into single real numbers by the score function, which
leads to the distortion and loss of original decision information,
and may further result in the unreasonable decision results.
In the above, we have provided an overview on the aggregation
techniques for intuitionistic fuzzy information, based on which we
now point out several critical comments and new directions for
future research:
(1) Various operational laws for IFNs have been dened, based
on which a large number of aggregation operators have been
subsequently developed for intuitionistic fuzzy information.
However, there is usually not a specic interpretation when
dening a new operational law for IFNs. Some of them are
provided just for the denitions of some intuitionistic fuzzy
aggregation operators subsequently. Hence, it is expected to
give some interpretations, such as some geometrical interpretations, to explain the meaning of each operation on IFNs.
(2) A great number of aggregation operators have been developed for aggregating the rst three types of intuitionistic
fuzzy information, however, only several aggregation operators have been designed to aggregate the intuitionistic fuzzy
information in different time periods, which are very
common in real-life decision making problems. More new
aggregation operators should be developed. Furthermore,
since different operators in the same category have different
advantages and disadvantages, it is very hard for the decision making practitioners to select the suitable aggregation
operator in practical decision making.
(3) In decision making, it is sometimes needed to select, from a
given set of aggregation operators, the one that provides the
output least dissimilar to all the inputs [5,6]. In such cases,
we may consider applying the concept of penalty function
to the intuitionistic fuzzy setting to assist the decision
makers to select the most suitable aggregation operator.
(4) Many aggregation operators are dened just by generalizing
some classical aggregation operators and may be not dened
from the practical application point of view. Accordingly, it
is required to conduct a clear justication analysis of the real
usefulness.

19





for j 1; 2; . . . ; m, and A re1 ; re2 ; . . . ; rf
with rej mini uij ;
m
maxi v ij for j 1; 2; . . . ; m, as the IFPIS and the IFNIS, respectively.



f
r
r
However, Li [43] suggested A f
with f
1 ; r2 ; . . . ; rm
j 1; 0
  




f
for j 1; 2; . . . ; m, and A f
r1 ; f
r 2 ; . . . ; rf
r

0; 1 for
with
m
j
j 1; 2; . . . ; m, as the IFPIS and the IFNIS, respectively. In the
TOPSIS based method, the best alternative should simultaneously
have the shortest distance from the IFPIS and the farthest distance
from the IFNIS, which states that this method is suitable for the cautious decision makers who want to have a decision which not only
makes as much prot as possible, but also avoids as much risk as possible [98]. To cope with the situations where the decision makers
want to have maximum prot and do not care for the decision risk
too much, the VIKOR method was extended to the intuitionistic
fuzzy context[99,100] for solving the problems with conicting
and noncommensurable attributes by mutual concessions. In the
TOPSIS-based and VIKOR-based methods, the weight vector of attributes is known a priori, and the IFPIS and the IFNIS are derived from
the intuitionistic fuzzy decision matrix directly. Nonetheless, sometimes, the weight vector of attributes is unknown a priori. Hence,
some researchers have extended the LINMAP method to the intuitionistic fuzzy environment and developed the IF-LINMAP method
[36,101], in which the IFPIS and the weights of attributes are
estimated by using a linear programming model, and the best compromise alternative is the feasible one with the shortest distance to
the IFPIS.
5.2. The outranking-based approach
The second category is the outranking based method, which is
used to identify which alternative is preferable, incomparable or
indifferent by a pairwise comparison of alternatives over each
attribute. Currently, the existing outranking based methods for
intuitionistic fuzzy decision making are mainly designed on the
basis of the classical outranking methods, such as the ELECTRE
(Elimination et Choice Translating Reality) method [7] and the
PROMETHEE (Preference Ranking Organization Method for
Enrichment Evaluation) method [8]. For example, Wu and Chen
[102] and Vahdani et al. [103] extended the ELECTRE method [7]
to solve the IF-MADM and IF-MAGDM problems in which the attribute weights are expressed by real numbers or IFNs, respectively.
The extended ELECTRE methods analyze the outranking relations
among alternatives by dening the concordance and discordance
indices. Liao and Xu [104] put forward the intuitionistic fuzzy
PROMETHEE method to solve the IF-MADM problem with intuitionistic fuzzy weights, which derives a partial or complete ranking of alternatives based upon the positive outranking ow, the
negative outranking ow and the net outranking ow.

5. Approaches to ranking alternatives

In this section, we review the main approaches for ranking
alternatives under the intuitionistic fuzzy decision scenarios,
which are concluded into the following four categories.
5.1. The ideal solution-based approach
The rst category is to rank the alternatives by means of the idea
solution, which is mainly developed on the basis of some
well-known ranking methods, such as the TOPSIS (Technique for
Order Preference by Similarity to Ideal Solution) method [4], the
LINMAP (Linear programming technique for Multidimensional
Analysis of Preference) method [95] and the VIKOR
(VlseKriter-ijumska Optimizacija I Kompromisno Resenje) method
[96]. In order to extend the TOPSIS method successfully, Chen and





Tsao [97] dened A re1 ; re2 ; . . . ; rf
with rej maxi uij ; mini v ij
m

5.3. The interactive approach

The main idea of this method is that it allows the decision
maker to interact with group members (or analysts) in the process
of decision making by revising or supplementing the original information on the decision making problem. Recently, Su et al. [105]
introduced an interactive model for intuitionistic fuzzy dynamic
multi-attribute group decision making (IF-DMAGDM) based on
the DIFWA operator [29] and the intuitionistic fuzzy TOPSIS
method [97], in which the assessments of the decision makers
for alternatives are allowed to be revised when the consensus of
the group is not satisfactory. Xu [41] developed a satisfaction
degree based intuitionistic fuzzy interactive method, which permits the decision maker to modify the lower bounds of the satisfaction degrees of alternatives if he/she is not satised with the
yielded decision results. Later, Xu and Xia [106] presented another

20

Z. Xu, N. Zhao / Information Fusion 28 (2016) 1023

interactive method for IF-MADM by eliminating any dominated

alternatives, which updates the decision makers preferences for
weight information gradually until the optimal alternative is
found.
5.4. The psychological behavior-based approach
The last category takes account of the decision makers psychological behavior in the decision making process, which thinks that

the psychological characteristics of the decision maker have an

important inuence on the nal decision results. Chen [107,108]
related the decision-makers optimistic and pessimistic attitudes
to the IF-MADM by developing the optimistic and pessimistic point
operators and establishing a unipolar bivariate model, respectively.
Chen and Yang [109] reected the risk attitude of the decision
maker by means of a new class of decision functions, which are
developed based on the mean-variance type utility functions in
the nancial risk management. Recently, Krohling et al. [110]

Table 3
Different alternative ranking approaches for intuitionistic fuzzy decision making.
Categories

Methods

Papers

Based-techniques

IFPIS

IFNIS

Application
scopes

Ideal solution- based approach

TOPSIS-based method

Chen and Tsao [97]

Boran et al. [112]
Li [43]
Xu [113]
Wei [42]
Li [101]
Li et al. [36]
Vahdani et al. [99]
Park et al. [100]
Xu [34]
Xu and Hu [114]
Ye [31]
Wang and Zhang [115]

Distance measure
Distance measure
Distance measure
Similarity measure
Grey relational coefcient
Distance measure
Distance measure
Distance measure
Distance measure
Distance measure
Projection model
Correlation coefcient
Evidential reasoning algorithms, distance
measure

A
A
A
A
A
unknown
unknown
A
A
A
A
A
A

A
A
A
A
A

A
A

A

MADM
MAGDM
MADM
MADM
MADM
MADM
MAGDM
MADM
DMADM
MADM
MADM
MADM
MADM

Wu and Chen [102]

Distance measure

MADM

Vahdani et al. [103]

Liao and Xu [104]

Distance measure
Distance measure

MAGDM
MADM

Interactive approach

Su et al. [105]
Xu [41]

Revising decision information

Modifying the lower bounds of satisfaction
degrees of alternatives
Updating the preferences for attribute weights

A
A

A
A

DMAGDM
MADM

MADM

Psychological behavior-based
approach

Chen [107]
Chen [108]
Chen and Yang [109]

Optimistic and pessimistic point operators

Unipolar bivariate model
Decision functions developed by meanvariance type utility functions
TODIM method [111]

MADM
MADM
MADM

MADM

LINMAP-based
method
VIKOR-based method
Other methods

Outranking-based approach

ELECTRE-based
method
PROMETHEE-based
method

Xu and Xia [106]

Krohling et al. [110]

Note: denotes that the listed information is not involved in the discussed method.

Table 4
Applications in different elds.
Fields

Applications

Papers

Based-means

Logistic management

Supplier selection & evaluation

Boran et al. [112]

Ashayeri et al. [117]
Chai et al. [118]
Devi and Yadav [120]
Boran [121]
Liu et al. [122]
Zhang et al. [123]
Chang et al. [124]

TOPSIS method [4]

IFCA operator [72]
SIR method [119]
ELECTRE-based method [102]
TOPSIS-based method [97]
IFCA operator [72]
Comparison method of IFSs [123]
Arithmetic operations [15]

Location selection
Strategy selection
Job-Shop scheduling
Failure analysis
Resource management

Resource service selection

Water resources management
Sustainable energy management
Human resources management

Tao et al. [125]

Hernandez and Uddameri [19]
Boran et al. [126]
Zhang and Liu [127]
Boran et al. [129]

IFWA operator [18]

Simple additive weighting method [43]
TOPSIS-based method [97]
GRA [128], Entropy-based method [31]
IFWA operator [18], TOPSIS-based method [97]

Evaluations

Web quality
Intellectual capital
Value evaluation
Network security evaluation
Performance evaluation
Teaching evaluation
Information system evaluation

Wang [130]
Zhao [132]
Zhang et al. [133]
Li et al. [134]
Zhang [135]
Fan [136]
Cui and Xu [137]

MaxMin-Max composition of IFSs [131]

IFWA operator [18]
IFWA and IFHA operators [18]
IFWA operator [18]
Entropy-based method [31]
IFEWA operator [69]
IFCG operator [73]

Predictions

Air quality prediction

Sentiment prediction

Yue et al. [138]

Wang et al. [139]

IFWA operator [18]

I-IFOWA operator [140]

Z. Xu, N. Zhao / Information Fusion 28 (2016) 1023

generalized the TODIM (an acronym in Portuguese of interactive

and multi-criteria decision making) method [111] to the intuitionistic fuzzy context and proposed the intuitionistic fuzzy TODIM
method, which represents the decision makers behavioral characteristics such as the loss aversion by the prospect value function.
To facilitate the reading and comparison, we summarize some
representative studies on the four ranking methods in Table 3.
From Table 3, it can be observed that lots of approaches have
been developed for ranking alternatives within the intuitionistic
fuzzy decision contexts, and each approach has its own characteristic. As a result, it is difcult for the decision making practitioners
to select the suitable approach in real decision making. Therefore,
it is required to make sufcient comparisons among these
approaches to present the advantages and disadvantages of each
approach fully in order to render some references for the decision
making practitioners to nd out the most suitable approach.
Moreover, it can be also seen from Table 3 that many approaches
apply the distance measures of IFSs into the process of ranking
alternatives. As there are numerous distance measures of IFSs
[116], it is necessary to conduct an investigation on the issues that
in each of these ranking approaches, whether all distance measures
of IFSs are applicable and which ones perform better in ranking
alternatives.
6. Recent applications
Up to now, we have overviewed the main processes in the intuitionistic fuzzy decision making: deriving attribute weight vector,
aggregating intuitionistic fuzzy decision information and ranking
candidate alternatives, and surveyed some relevant solution
approaches in depth. According to some representative application
papers, in this section, we review recent applications of these
approaches in different elds, including logistic management,
resource management, evaluation, and prediction, etc. To improve
readability, we summarize these applications in Table 4.
From Table 4, it can be seen that compared with the huge number of theoretical studies, there are relatively few works concerning the applications of the intuitionistic fuzzy decision making
approaches. At present, a few decision making approaches have
been adopted to solve practical problems and most of them have
not been applied in practice. The applications of more theoretical
approaches to real-word problems should be investigated.
Furthermore, researchers have mainly applied the existing intuitionistic fuzzy decision making approaches to the elds of logistic
management, resource management, evaluation, and prediction.
Applications in other practical elds should also be explored.
7. Conclusions
Intuitionistic fuzzy decision making, which addresses the problems of ranking alternatives based on the given intuitionistic fuzzy
decision information or several concerned attributes, has received
great attention in recent years, and various decision making
methodologies and approaches have been put forward. In this
paper, we have reviewed the existing main approaches to intuitionistic fuzzy decision making from the perspective of information fusion. Firstly, we have reviewed and analyzed the current
approaches for deriving attribute weights from the intuitionistic
fuzzy decision information or from both the intuitionistic fuzzy
decision information and the partially known weight information.
Then, we have surveyed the orders of IFNs, which play an important role in the ordered aggregation of intuitionistic fuzzy information, and overviewed the existing main aggregation operators for
intuitionistic fuzzy information in different situations. Moreover,
the main approaches for ranking alternatives based upon the

21

intuitionistic fuzzy decision information have been summarized

and analyzed, which can be divided into four categories: the ideal
solution-based approach, the outranking-based approach, the
interactive approach and the psychological behavior-based
approach. Finally, to show the applicability of the discussed
approaches, we have outlined their recent applications in different
elds.
Acknowledgements
The work was supported by the National Natural Science
Foundation of China (No. 61273209), the Fundamental Research
Funds for the Central Universities (No. KYLX_0207) and the
Scientic Research Foundation of Graduate School of Southeast
University (No. YBJJ1527).
References
[1] Z.S. Xu, Uncertain Multi-attribute Decision Making, Tsinghua University
Press, Beijing, 2004.
[2] Z.S. Xu, Q.L. Da, An overview of operators for aggregating information, Int. J.
Intell. Syst. 18 (2003) 953969.
[3] T.L. Saaty, A scaling method for priorities in hierarchical structures, J. Math.
Psychol. 15 (1977) 234281.
[4] C.L. Hwang, K. Yoon, Multiple Attribute Decision Making: Methods and
Applications, Springer-Verlag, New York, 1981.
[5] H. Bustince, E. Barrenechea, T. Calvo, S. James, G. Beliakov, Consensus in
multi-expert decision making problems using penalty functions dened over
a Cartesian product of lattices, Inform. Fusion 17 (2014) 5664.
[6] H. Bustince, A. Jurio, A. Pradera, R. Mesiar, G. Beliakov, Generalization of the
weighted voting method using penalty functions constructed via faithful
restricted dissimilarity functions, Eur. J. Oper. Res. 225 (2013) 472478.
[7] B. Roy, Classement et choix en prsence de points de vue multiples, Revue
franaise dautomatique, dinformatique et de recherche oprationnelle, Rech.
Oper. 2 (1968) 5775.
[8] J.P. Brans, Lingenierie de la decision, llaboration dinstruments daidea la
decision. Colloque sur lAidea la Decision, Faculte des Sciences de
lAdministration, Universite Laval, 1982.
[9] A.A. Salo, Interactive decision aiding for group decision support, Eur. J. Oper.
Res. 84 (1995) 134149.
[10] I. Palomares, F.J. Estrella, L. Martnez, F. Herrera, Consensus under a fuzzy
context: taxonomy, analysis framework AFRYCA and experimental case of
study, Inform. Fusion 20 (2014) 252271.
[11] L.A. Zadeh, Fuzzy sets, Inform. Control 8 (1965) 338353.
[12] E. Herrera-Viedma, F.J. Cabrerizo, J. Kacprzyk, W. Pedrycz, A review of soft
consensus models in a fuzzy environment, Inform. Fusion 17 (2014) 413.
[13] X. Chen, H. Zhang, Y. Dong, The fusion process with heterogeneous preference
structures in group decision making: a survey, Inform. Fusion 24 (2015) 72
83.
[14] Z.S. Xu, X.Q. Cai, Intuitionistic Fuzzy Information Aggregation, Springer, New
York, 2012.
[15] K.T. Atanassov, Intuitionistic fuzzy sets, Fuzzy Sets Syst. 20 (1986) 8796.
[16] Z.S. Xu, Multi-person multi-attribute decision making models under
intuitionistic fuzzy environment, Fuzzy Optim. Decis. Making 6 (2007)
221236.
[17] Z.S. Xu, R.R. Yager, Some geometric aggregation operators based on
intuitionistic fuzzy sets, Int. J. Gen. Syst. 35 (2006) 417433.
[18] Z.S. Xu, Intuitionistic fuzzy aggregation operators, IEEE Trans. Fuzzy Syst. 15
(2007) 11791187.
[19] E.A. Hernandez, V. Uddameri, Selecting agricultural best management
practices for water conservation and quality improvements using
Atanassovs intuitionistic fuzzy sets, Water Resour. Manage. 24 (2010)
45894612.
[20] S.P. Wan, J.Y. Dong, Interval-valued intuitionistic fuzzy mathematical
programming method for hybrid multi-criteria group decision making with
interval-valued intuitionistic fuzzy truth degrees, Inform. Fusion 26 (2015)
4965.
[21] B. Liu, Y. Shen, X. Chen, Y. Chen, X. Wang, A partial binary tree DEA-DA cyclic
classication model for decision makers in complex multi-attribute largegroup interval-valued intuitionistic fuzzy decision-making problems, Inform.
Fusion 18 (2014) 119130.
[22] Z.S. Xu, Intuitionistic preference relations and their application in group
decision making, Inform. Sci. 177 (2007) 23632379.
[23] D. Paternain, A. Jurio, E. Barrenechea, H. Bustince, B. Bedregal, E. Szmidt, An
alternative to fuzzy methods in decision-making problems, Expert Syst. Appl.
39 (2012) 77297735.
[24] K. Atanassov, Intuitionistic Fuzzy Sets: Theory and Applications, PhysicaVerlag, Germany, 1999.
[25] Z.S. Xu, Intuitionistic Preference Modeling and Interactive Decision Making,
Springer, New York, 2014.

22

Z. Xu, N. Zhao / Information Fusion 28 (2016) 1023

[26] Z.S. Xu, H.C. Liao, A survey of approaches to decision making with
intuitionistic fuzzy preference relations, Knowl.-Based Syst. 80 (2015) 131
142.
[27] Z.S. Xu, A deviation-based approach to intuitionistic fuzzy multiple attribute
group decision making, Group Decis. Negot. 19 (2010) 5776.
[28] K.S. Park, Mathematical programming models for characterizing dominance
and potential optimality when multicriteria alternative values and weights
are simultaneously incomplete, IEEE Trans. Syst., Man Cybernet., A: Syst.
Hum. 34 (2004) 601614.
[29] Z.S. Xu, R.R. Yager, Dynamic intuitionistic fuzzy multi-attribute decision
making, Int. J. Approx. Reason. 48 (2008) 246262.
[30] S.H. Kim, B.S. Ahn, Interactive group decision making procedure under
incomplete information, Eur. J. Oper. Res. 116 (1999) 498507.
[31] J. Ye, Fuzzy decision-making method based on the weighted correlation
coefcient under intuitionistic fuzzy environment, Eur. J. Oper. Res. 205
(2010) 202204.
[32] E. Szmidt, J. Kacprzyk, Entropy for intuitionistic fuzzy sets, Fuzzy Sets Syst.
118 (2001) 467477.
[33] T.Y. Chen, C.H. Li, Determining objective weights with intuitionistic fuzzy
entropy measures: a comparative analysis, Inform. Sci. 180 (2010) 4207
4222.
[34] Z.S. Xu, Models for multiple attribute decision making with intuitionistic
fuzzy information, Int. J. Uncert., Fuzz. Knowl.-Based Syst. 15 (2007) 285297.
[35] S. French, R. Hartley, L.C. Thomas, D.J. White, Multi-objective Decision
Making, Academic Press, New York, 1983.
[36] D.F. Li, G.H. Chen, Z.G. Huang, Linear programming method for multiattribute
group decision making using IF sets, Inform. Sci. 180 (2010) 15911609.
[37] Z.S. Xu, X.Q. Cai, Nonlinear optimization models for multiple attribute group
decision making with intuitionistic fuzzy information, Int. J. Intell. Syst. 25
(2010) 489513.
[38] J.Z. Wu, Q. Zhang, Multicriteria decision making method based on
intuitionistic fuzzy weighted entropy, Expert Syst. Appl. 38 (2011) 916922.
[39] M.M. Xia, Z.S. Xu, Entropy/cross entropy-based group decision making under
intuitionistic fuzzy environment, Inform. Fusion 13 (2012) 3147.
[40] G.W. Wei, Maximizing deviation method for multiple attribute decision
making in intuitionistic fuzzy setting, Knowl.-Based Syst. 21 (2008) 833836.
[41] Z.S. Xu, Intuitionistic fuzzy multiattribute decision making: an interactive
method, IEEE Trans. Fuzzy Syst. 20 (2012) 514525.
[42] G.W. Wei, GRA method for multiple attribute decision making with
incomplete weight information in intuitionistic fuzzy setting, Knowl.-Based
Syst. 23 (2010) 243247.
[43] D.F. Li, Multiattribute decision making models and methods using
intuitionistic fuzzy sets, J. Comput. Syst. Sci. 70 (2005) 7385.
[44] S.M. Chen, J.M. Tan, Handling multicriteria fuzzy decision-making problems
based on vague set theory, Fuzzy Sets Syst. 67 (1994) 163172.
[45] T.Y. Chen, A comparative analysis of score functions for multiple criteria
decision making in intuitionistic fuzzy settings, Inform. Sci. 181 (2011) 3652
3676.
[46] D.H. Hong, C.H. Choi, Multicriteria fuzzy decision-making problems based on
vague set theory, Fuzzy Sets Syst. 114 (2000) 103113.
[47] H. Bustince, J. Fernandez, A. Kolesrov, R. Mesiar, Generation of linear orders
for intervals by means of aggregation functions, Fuzzy Sets Syst. 220 (2013)
6977.
[48] L.D. Miguel, H. Bustince, J. Fernandez, E. Indurin, A. Kolesrov, R. Mesiar,
Construction of admissible linear orders for interval-valued Atanassov
intuitionistic fuzzy sets with an application to decision making, Inform.
Fusion 27 (2015) 189197.
[49] H. Bustince, M. Galar, B. Bedregal, A. Kolesarova, R. Mesiar, A new approach to
interval-valued Choquet Integrals and the problem of ordering in intervalvalued fuzzy set applications, IEEE Trans. Fuzzy Syst. 21 (2013) 11501162.
[50] E. Barrenechea, J. Fernandez, M. Pagola, F. Chiclana, H. Bustince, Construction
of interval-valued fuzzy preference relations from ignorance functions and
fuzzy preference relations. Application to decision making, Knowl.-Based
Syst. 58 (2014) 3344.
[51] U. Bentkowska, H. Bustince, A. Jurio, M. Pagola, B. Pekala, Decision making
with an interval-valued fuzzy preference relation and admissible orders, Appl.
Soft Comput., in press (2015), http://dx.doi.org/10.1016/j.asoc.2015.03.012.
[52] G.J. Klir, B. Yuan, Fuzzy Sets and Fuzzy Logic: Theory and Applications,
Prentice Hall, USA, 1995.
[53] G. Beliakov, H. Bustince, D. Goswami, U. Mukherjee, N.R. Pal, On averaging
operators for Atanassovs intuitionistic fuzzy sets, Inform. Sci. 181 (2011)
11161124.
[54] M.M. Xia, Z.S. Xu, B. Zhu, Some issues on intuitionistic fuzzy aggregation
operators based on Archimedean t-conorm and t-norm, Knowl.-Based Syst.
31 (2012) 7888.
[55] M.M. Xia, Z.S. Xu, Generalized point operators for aggregating intuitionistic
fuzzy information, Int. J. Intell. Syst. 25 (2010) 10611080.
[56] H.W. Liu, G.J. Wang, Multi-criteria decision-making methods based on
intuitionistic fuzzy sets, Eur. J. Oper. Res. 179 (2007) 220233.
[57] Y. He, H. Chen, L. Zhou, J. Liu, Z. Tao, Intuitionistic fuzzy geometric interaction
averaging operators and their application to multi-criteria decision making,
Inform. Sci. 259 (2014) 142159.
[58] B. Li, W. He, Intuitionistic fuzzy PRI-AND and PRI-OR aggregation operators,
Inform. Fusion 14 (2013) 450459.
[59] H.C. Liao, Z.S. Xu, Intuitionistic fuzzy hybrid weighted aggregation operators,
Int. J. Intell. Syst. 29 (2014) 971993.

[60] H. Zhao, Z.S. Xu, M.F. Ni, S.S. Liu, Generalized aggregation operators for
intuitionistic fuzzy sets, Int. J. Intell. Syst. 25 (2010) 130.
[61] J.C. Harsanyi, Cardinal welfare, individualistic ethics, and interpersonal
comparisons of utility, J. Polit. Econ. 63 (1955) 309321.
[62] Z.S. Xu, On consistency of the weighted geometric mean complex judgement
matrix in AHP, Eur. J. Oper. Res. 126 (2000) 683687.
[63] R.R. Yager, On ordered weighted averaging aggregation operators in
multi-criteria decision making, IEEE Trans. Syst., Man Cybernet. 18 (1988)
183190.
[64] Z.S. Xu, Q.L. Da, The ordered weighted geometric averaging operators, Int. J.
Intell. Syst. 17 (2002) 709716.
[65] R.R. Yager, Generalized OWA aggregation operators, Fuzzy Optim. Decis.
Making 3 (2004) 93107.
[66] R.R. Yager, OWA aggregation of intuitionistic fuzzy sets, Int. J. Gen. Syst. 38
(2009) 617641.
[67] D.F. Li, Multiattribute decision making method based on generalized OWA
operators with intuitionistic fuzzy sets, Expert Syst. Appl. 37 (2010)
86738678.
[68] S. Zeng, W. Su, Intuitionistic fuzzy ordered weighted distance operator,
Knowl.-Based Syst. 24 (2011) 12241232.
[69] W. Wang, X. Liu, Intuitionistic fuzzy geometric aggregation operators based
on Einstein operations, Int. J. Intell. Syst. 26 (2011) 10491075.
[70] Z.X. Su, G.P. Xia, M.Y. Chen, L. Wang, Induced generalized intuitionistic fuzzy
OWA operator for multi-attribute group decision making, Expert Syst. Appl.
39 (2012) 19021910.
[71] W. Yang, Z. Chen, The quasi-arithmetic intuitionistic fuzzy OWA operators,
Knowl.-Based Syst. 27 (2012) 219233.
[72] C. Tan, X. Chen, Intuitionistic fuzzy Choquet integral operator for multicriteria decision making, Expert Syst. Appl. 37 (2010) 149157.
[73] Z.S. Xu, Choquet integrals of weighted intuitionistic fuzzy information,
Inform. Sci. 180 (2010) 726736.
[74] Z.S. Xu, Approaches to multiple attribute group decision making based on
intuitionistic fuzzy power aggregation operators, Knowl.-Based Syst. 24
(2011) 749760.
[75] Z.S. Xu, R.R. Yager, Intuitionistic fuzzy Bonferroni means, IEEE Trans. Syst.,
Man Cybernet., B: Cybernet. 41 (2011) 568578.
[76] M.M. Xia, Z.S. Xu, B. Zhu, Generalized intuitionistic fuzzy Bonferroni means,
Int. J. Intell. Syst. 27 (2012) 2347.
[77] W. Zhou, J.M. He, Intuitionistic fuzzy geometric Bonferroni means and their
application in multicriteria decision making, Int. J. Intell. Syst. 27 (2012)
9951019.
[78] D.J. Yu, Intuitionistic fuzzy geometric Heronian mean aggregation operators,
Appl. Soft Comput. 13 (2013) 12351246.
[79] G. Choquet, Theory of capacities, Ann. linst. Fourier 5 (1953) 8587.
[80] R.R. Yager, The power average operator, IEEE Trans. Syst., Man Cybernet., A:
Syst. Hum. 31 (2001) 724731.
[81] C. Bonferroni, Sulle medie multiple di potenze, B. Unione Mat. Ital. 5 (1950)
267270.
[82] G. Beliakov, A. Pradera, T. Calvo, Aggregation Functions: A Guide for
Practitioners, Springer, Berlin, 2007.
[83] J. Wu, F. Chen, C. Nie, Q. Zhang, Intuitionistic fuzzy-valued Choquet integral
and its application in multicriteria decision making, Inform. Sci. 222 (2013)
509527.
[84] C. Tan, X. Chen, Induced intuitionistic fuzzy Choquet integral operator for
multicriteria decision making, Int. J. Intell. Syst. 26 (2011) 659686.
[85] L. Zhou, H. Chen, J. Liu, Generalized power aggregation operators and
their applications in group decision making, Comput. Ind. Eng. 62 (2012)
989999.
[86] Z.M. Zhang, Generalized Atanassovs intuitionistic fuzzy power geometric
operators and their application to multiple attribute group decision making,
Inform. Fusion 14 (2013) 460486.
[87] G. Beliakov, S. James, On extending generalized Bonferroni means to
Atanassov orthopairs in decision making contexts, Fuzzy Sets Syst. 211
(2013) 8498.
[88] Y.J. Xu, T. Sun, D.F. Li, Intuitionistic fuzzy prioritized OWA operator and its
application in multi-criteria decision-making problem, Control Decis. 26
(2011) 129132.
[89] D. Yu, Intuitionistic fuzzy prioritized operators and their application in multicriteria group decision making, Technol. Econ. Dev. Econ. 19 (2013) 121.
[90] D. Yu, Group decision making based on generalized intuitionistic fuzzy
prioritized geometric operator, Int. J. Intell. Syst. 27 (2012) 635661.
[91] X.H. Yu, Z.S. Xu, Prioritized intuitionistic fuzzy aggregation operators, Inform.
Fusion 14 (2013) 108116.
[92] R.R. Yager, Prioritized OWA aggregation, Fuzzy Optim. Decis. Making 8 (2009)
245262.
[93] R.R. Yager, Prioritized aggregation operators, Int. J. Approx. Reason. 48 (2008)
263274.
[94] G. Wei, Some geometric aggregation functions and their application to
dynamic multiple attribute decision making in the intuitionistic fuzzy
setting, Int. J. Uncert., Fuzz. Knowl.-Based Syst. 17 (2009) 179196.
[95] V. Srinivasan, A.D. Shocker, Linear programming techniques for
multidimensional analysis of preferences, Psychometrika 38 (1973) 337369.
[96] S. Opricovic, Multicriteria optimization of civil engineering systems, Fac. Civ.
Eng. 2 (1998) 521.
[97] T.Y. Chen, C.Y. Tsao, The interval-valued fuzzy TOPSIS method and
experimental analysis, Fuzzy Sets Syst. 159 (2008) 14101428.

Z. Xu, N. Zhao / Information Fusion 28 (2016) 1023

[98] S. Opricovic, G.H. Tzeng, Compromise solution by MCDM methods: a
comparative analysis of VIKOR and TOPSIS, Eur. J. Oper. Res. 156 (2004)
445455.
[99] B. Vahdani, H. Hadipour, J.S. Sadaghiani, M. Amiri, Extension of VIKOR method
based on interval-valued fuzzy sets, Int. J. Adv. Manuf. Technol. 47 (2010)
12311239.
[100] J.H. Park, H.J. Cho, Y.C. Kwun, Extension of the VIKOR method to dynamic
intuitionistic fuzzy multiple attribute decision making, Comput. Math. Appl.
65 (2013) 731744.
[101] D.F. Li, Extension of the LINMAP for multiattribute decision making under
Atanassovs intuitionistic fuzzy environment, Fuzzy Optim. Decis. Making 7
(2008) 1734.
[102] M.C. Wu, T.Y. Chen, The ELECTRE multicriteria analysis approach based on
Atanassovs intuitionistic fuzzy sets, Expert Syst. Appl. 38 (2011) 12318
12327.
[103] B. Vahdani, S.M. Mousavi, R. Tavakkoli-Moghaddam, H. Hashemi, A new
design of the elimination and choice translating reality method for multicriteria group decision-making in an intuitionistic fuzzy environment, Appl.
Math. Model. 37 (2013) 17811799.
[104] H.C. Liao, Z.S. Xu, Multi-criteria decision making with intuitionistic fuzzy
PROMETHEE, J. Intell. Fuzzy Syst. 27 (2014) 17031717.
[105] Z.X. Su, M.Y. Chen, G.P. Xia, L. Wang, An interactive method for dynamic
intuitionistic fuzzy multi-attribute group decision making, Expert Syst. Appl.
38 (2011) 1528615295.
[106] Z.S. Xu, M.M. Xia, Identifying and eliminating dominated alternatives in
multi-attribute decision making with intuitionistic fuzzy information, Appl.
Soft Comput. 12 (2012) 14511456.
[107] T.Y. Chen, An outcome-oriented approach to multicriteria decision analysis
with intuitionistic fuzzy optimistic/pessimistic operators, Expert Syst. Appl.
37 (2010) 77627774.
[108] T.Y. Chen, Bivariate models of optimism and pessimism in multi-criteria
decision-making based on intuitionistic fuzzy sets, Inform. Sci. 181 (2011)
21392165.
[109] Z. Chen, W. Yang, A new multiple criteria decision making method based on
intuitionistic fuzzy information, Expert Syst. Appl. 39 (2012) 43284334.
[110] R.A. Krohling, A.G. Pacheco, A.L. Siviero, IF-TODIM: an intuitionistic fuzzy
TODIM to multi-criteria decision making, Knowl.-Based Syst. 53 (2013)
142146.
[111] L. Gomes, M. Lima, TODIM: basics and application to multicriteria ranking of
projects with environmental impacts, Found. Comput. Decis. Sci. 16 (1992)
113127.
[112] F.E. Boran, S. Gen, M. Kurt, D. Akay, A multi-criteria intuitionistic fuzzy
group decision making for supplier selection with TOPSIS method, Expert
Syst. Appl. 36 (2009) 1136311368.
[113] Z.S. Xu, Some similarity measures of intuitionistic fuzzy sets and their
applications to multiple attribute decision making, Fuzzy Optim. Decis.
Making 6 (2007) 109121.
[114] Z.S. Xu, H. Hu, Projection models for intuitionistic fuzzy multiple attribute
decision making, Int. J. Inform. Tech. Decis. 9 (2010) 267280.
[115] J.Q. Wang, H.Y. Zhang, Multicriteria decision-making approach based on
Atanassovs intuitionistic fuzzy sets with incomplete certain information on
weights, IEEE Trans. Fuzzy Syst. 21 (2013) 510515.
[116] Z. Xu, J. Chen, An overview of distance and similarity measures of
intuitionistic fuzzy sets, Int. J. Uncert., Fuzz. Knowl.-Based Syst. 16 (2008)
529555.
[117] J. Ashayeri, G. Tuzkaya, U.R. Tuzkaya, Supply chain partners and conguration
selection: an intuitionistic fuzzy Choquet integral operator based approach,
Expert Syst. Appl. 39 (2012) 36423649.
[118] J.Y. Chai, J.N. Liu, Z.S. Xu, A new rule-based SIR approach to supplier selection
under intuitionistic fuzzy environments, Int. J. Uncert., Fuzz. Knowl.-Based
Syst. 20 (2012) 451471.

23

[119] X. Xu, The SIR method: a superiority and inferiority ranking method for
multiple criteria decision making, Eur. J. Oper. Res. 131 (2001) 587602.
[120] K. Devi, S.P. Yadav, A multicriteria intuitionistic fuzzy group decision making
for plant location selection with ELECTRE method, Int. J. Adv. Manuf. Technol.
66 (2013) 12191229.
[121] F.E. Boran, An integrated intuitionistic fuzzy multi criteria decision making
method for facility location selection, Math. Comput. Appl. 16 (2011)
487496.
[122] W. Liu, Z. Lin, F. Wen, G. Ledwich, Intuitionistic fuzzy Choquet integral
operator-based approach for black-start decision-making, IET Gener.,
Transm. Distrib. 6 (2012) 378386.
[123] X. Zhang, Y. Deng, F.T. Chan, P. Xu, S. Mahadevan, Y. Hu, IFSJSP: a novel
methodology for the job-shop scheduling problem based on intuitionistic
fuzzy sets, Int. J. Prod. Res. 51 (2013) 51005119.
[124] K.H. Chang, C.H. Cheng, Y.C. Chang, Reprioritization of failures in a silane
supply system using an intuitionistic fuzzy set ranking technique, Soft
Comput. 14 (2010) 285298.
[125] F. Tao, D. Zhao, L. Zhang, Resource service optimal-selection based on
intuitionistic fuzzy set and non-functionality QoS in manufacturing grid
system, Knowl. Inform. Syst. 25 (2010) 185208.
[126] F. Boran, K. Boran, T. Menlik, The evaluation of renewable energy
technologies for electricity generation in Turkey using intuitionistic fuzzy
TOPSIS, Energy Source, B 7 (2012) 8190.
[127] S.F. Zhang, S.Y. Liu, A GRA-based intuitionistic fuzzy multi-criteria group
decision making method for personnel selection, Expert Syst. Appl. 38 (2011)
1140111405.
[128] J.L. Deng, Introduction to grey system theory, J. Grey Syst. UK 1 (1989)
124.
[129] F.E. Boran, S. Gen, D. Akay, Personnel selection based on intuitionistic fuzzy
sets, Hum. Factor. Ergon. Man. 21 (2011) 493503.
[130] P. Wang, QoS-aware web services selection with intuitionistic fuzzy set
under consumers vague perception, Expert Syst. Appl. 36 (2009) 44604466.
[131] S.K. De, R. Biswas, A.R. Roy, An application of intuitionistic fuzzy sets in
medical diagnosis, Fuzzy Sets Syst. 117 (2001) 209213.
[132] X. Zhao, Evaluation model for intellectual capital with intuitionistic fuzzy
information, in: IEEE International Conference on Industrial Mechatronics
and Automation, 2009, pp. 398401.
[133] D. Zhang, X. Zeng, H. Chen, W. He, Research on the evaluation models of
customer value of brokers in the circumstances of electronic commerce with
intuitionistic fuzzy information, Adv. Inform. Sci. Serv. Sci. 3 (2011) 7681.
[134] Y. Li, J. Yin, G. Wu, An approach to evaluating the computer network security
with intuitionistic fuzzy information, Adv. Inform. Sci. Serv. Sci. 3 (2011)
195200.
[135] Y. Zhang, Research on the pre-evaluation methods on performance balance of
computer with intuitionistic fuzzy information, Adv. Inform. Sci. Serv. Sci. 4
(2012) 161167.
[136] H. Fan, Research on the physical education teaching effectiveness of the
higher colleges and universities with intuitionistic fuzzy information, J.
Converg. Inform. Technol. 7 (2012) 463469.
[137] S.Y. Cui, A.F. Xu, An approach to evaluating the enterprise management
information system with intuitionistic fuzzy information, Int. J. Adv. Comput.
Technol. 4 (2012) 4955.
[138] Z. Yue, Y. Jia, C. Zhu, Prediction of air quality during 2010 asian games in
Guangzhou, in: IEEE International Conference on Bioinformatics and
Biomedical Engineering, 2009, pp. 15.
[139] H. Wang, G. Qian, X.Q. Feng, Predicting consumer sentiments using online
sequential extreme learning machine and intuitionistic fuzzy sets, Neural
Comput. Appl. 22 (2013) 479489.
[140] G. Wei, Induced intuitionistic fuzzy ordered weighted averaging operator and
its application to multiple attribute group decision making, in: Rough Sets
and Knowledge Technology, Springer, 2008, pp. 124131.