
Citation: Yefu Zheng, Jun Xu, Hongzhang Chen. TOPSIS-based entropy measure for intuitionistic trapezoidal fuzzy sets and application to multi-attribute decision making[J]. Mathematical Biosciences and Engineering, 2020, 17(5): 5604-5617. doi: 10.3934/mbe.2020301
[1] | Fankang Bu, Jun He, Haorun Li, Qiang Fu . Interval-valued intuitionistic fuzzy MADM method based on TOPSIS and grey correlation analysis. Mathematical Biosciences and Engineering, 2020, 17(5): 5584-5603. doi: 10.3934/mbe.2020300 |
[2] | Haiping Ren, Laijun Luo . A novel distance of intuitionistic trapezoidal fuzzy numbers and its-based prospect theory algorithm in multi-attribute decision making model. Mathematical Biosciences and Engineering, 2020, 17(4): 2905-2922. doi: 10.3934/mbe.2020163 |
[3] | Yongfeng Yin, Routing Zhang, Qingran Su . Threat assessment of aerial targets based on improved GRA-TOPSIS method and three-way decisions. Mathematical Biosciences and Engineering, 2023, 20(7): 13250-13266. doi: 10.3934/mbe.2023591 |
[4] | Muhammad Akram, G. Muhiuddin, Gustavo Santos-García . An enhanced VIKOR method for multi-criteria group decision-making with complex Fermatean fuzzy sets. Mathematical Biosciences and Engineering, 2022, 19(7): 7201-7231. doi: 10.3934/mbe.2022340 |
[5] | Nian Zhang, Yifan Zhou, Qiang Pan, Guiwu Wei . Multi-attribute decision-making method with triangular fuzzy numbers based on regret theory and the catastrophe progression method. Mathematical Biosciences and Engineering, 2022, 19(12): 12013-12030. doi: 10.3934/mbe.2022559 |
[6] | Han Wu, Junwu Wang, Sen Liu, Tingyou Yang . Research on decision-making of emergency plan for waterlogging disaster in subway station project based on linguistic intuitionistic fuzzy set and TOPSIS. Mathematical Biosciences and Engineering, 2020, 17(5): 4825-4851. doi: 10.3934/mbe.2020263 |
[7] | Qiwen Wang, Guibao Song, Xiuxia Yang . Mixed-attitude three-way decision model for aerial targets: Threat assessment based on IF-VIKOR-GRA method. Mathematical Biosciences and Engineering, 2023, 20(12): 21514-21536. doi: 10.3934/mbe.2023952 |
[8] | Nian Zhang, Xue Yuan, Jin Liu, Guiwu Wei . Stochastic multiple attribute decision making with Pythagorean hesitant fuzzy set based on regret theory. Mathematical Biosciences and Engineering, 2023, 20(7): 12562-12578. doi: 10.3934/mbe.2023559 |
[9] | Yuting Zhu, Wenyu Zhang, Junjie Hou, Hainan Wang, Tingting Wang, Haining Wang . The large-scale group consensus multi-attribute decision-making method based on probabilistic dual hesitant fuzzy sets. Mathematical Biosciences and Engineering, 2024, 21(3): 3944-3966. doi: 10.3934/mbe.2024175 |
[10] | Sumera Naz, Muhammad Akram, Mohammed M. Ali Al-Shamiri, Mohammed M. Khalaf, Gohar Yousaf . A new MAGDM method with 2-tuple linguistic bipolar fuzzy Heronian mean operators. Mathematical Biosciences and Engineering, 2022, 19(4): 3843-3878. doi: 10.3934/mbe.2022177 |
Multi-attribute decision making (MADM) method has played an important role in operations research and modern decision science by effectively evaluating the alternative with multiple attributes. The evaluations of decision makers are always vague and imprecise due to the complexity of an actual decision-making environment. Si et al. [1] presented a novel method to compare the picture fuzzy numbers and applied it to solve decision making problems. Petrovic and Kankaras [2] developed a hybridized DEMATEL-AHP-TOPSIS for air traffic control radar position. Biswas et al. [3] proposed a multi-criteria decision making framework based on entropy measure to assess the mutual funds. Intuitionistic fuzzy (IF) sets (IFSs) proposed by Atanassov [4] can express the uncertainty and ambiguity of the information system quantitatively and intuitively. Subsequently, Atanassov and Gargov [5] introduced an interval-valued IFS (IVIFS) by using interval numbers to describe membership and non-membership functions. The IVIFS excellently expresses the imprecise preference for decision making. Thus far, IVIFS has received considerable attention in decision making [6,7,8,9] and entropy measure [10,11,12,13,14,15].
With the increasing uncertainties and complexities involved in the management and decision situation, the higher requirements are put forward to represent fuzzy information. As data analysis and processing theory, picture fuzzy set, and fuzzy neutrosophic set are an effective tool to deal with imprecise and inconsistent information, but their values are expressed as single values. In real decision-making, single values cannot accurately describe the reality, uncertainty, and distortion of things. Besides, modeling a continuous set by using IF numbers (IFNs) and interval-valued IFNs (IVIFNs) is difficult. Thus, as an extension of IFSs, intuitionistic trapezoidal fuzzy numbers (ITrFNs) introduced by Liu and Yuan [16], can express more uncertainty from different dimensions of decision information than IFNs and IVIFNs. ITrFN extends IFS's discourse universe from a discrete set to a continuous set [17] because its prominent characteristic is that trapezoidal fuzzy numbers describe the corresponding membership and non-membership degrees. Thus, ITrFNs not only can depict the fuzzy concept of 'good' or 'excellent' but also present the concept abundantly [16,17]. In recent years, the research and application of intuitionistic triangular fuzzy numbers (ITFNs), which are a particular case of ITrFNs, have attracted considerable attention from scholars, such as Wang [18]; Wei [19]; Gao et al. [20]; Yu and Xu [21]. The current achievements are mainly concentrated in two aspects: (1) The ranking method of ITFNs based score and accuracy functions, (2) the intuitionistic triangular fuzzy agammaegation operators. But there is no investigation on entropy measure and its application in intuitionistic triangular fuzzy MADM with attribute weight completely unknown. Therefore, the entropy measure and MADM method under ITrFNs, which are exciting yet relatively sophisticated, must be discussed.
Technique for order preference by similarity to an ideal solution (TOPSIS) [22] is a well-known method for MADM. The extended TOPSIS method for MADM problems with IFNs and IVIFNs using the connection numbers of set pair analysis theory was presented in [7] and [8], respectively. Garg and Kumar [6] proposed a TOPSIS approach based on a new exponential distance to handle MADM problems with IVIFN information. Subsequently, Garg and Kumar [9] applied the TOPSIS method to solve decision problems under a linguistic interval-valued IF (IVIF) environment. The present work is motivated by TOPSIS methods [6,7,8,9,22] and initially proposes an entropy measure of the intuitionistic trapezoidal fuzzy set (ITrFS) based on TOPSIS method and then provide an objective weighted approach. Accordingly, a MADM method with unknown weight information under an ITrFN environment is developed. The primary contributions of this study can be illuminated briefly as follows. (1) We newly define a Hamming distance measure of ITrFS and discuss its properties. (2) We propose entropy axioms and measure for ITrFS, which is the first report for entropy measure based on the idea of TOPSIS. (3) On this basis, we apply them to determine attribute weights in the ITrFN environment with unknown weight information and propose a method to address MADM problems with ITrFNs.
The remainder of this paper is organized as follows. Section 2 briefly introduces related basic concepts. Section 3 presents an entropy measure for ITrFSs. In section 4, an objective approach to determine attribute weights is developed, and a MADM method with ITrFNs is proposed. Section 5 provides a numerical example to illustrate the feasibility of the proposed method. Section 6 presents our conclusions.
Definition 1. [16]. A trapezoidal fuzzy number (TrFN) A is a fuzzy set in the set R of real numbers, with its membership function defined by
FA(x)={0, if x<a1,x−a1a2−a1, if a1≤x≤a2,1, if a2≤x≤a3,x−a4a3−a4, if a3≤x≤a4,0, if x>a4, | (1) |
where a1≤a2≤a3≤a4 , a1 and a4 present the lower limit and upper limit of A, respectively, [a2,a3] is the mode interval, which can be denoted as a four-tuple (a1,a2,a3,a4) .
Definition 2. Let X be a fixed set, μ A(x)=(tlA(x),tm1A(x),tm2A(x),thA(x)) and v A(x)=(flA(x),fm1A(x),fm2A(x),fhA(x)) are TrFNs defined on the unit interval [0, 1], then an intuitionistic trapezoidal fuzzy set (ITrFS) A over X is defined as A={(x,<μ A(x),v A(x)>)|x∈X} where the parameters μ A(x) and v A(x) indicate, respectively, the membership degree and non-membership degree of the element x in A , with the conditions 0≤thA(x)+fhA(x)≤1 . tlA(x) and thA(x) present the lower limit and upper limit of μ A(x) , [tm1A(x),tm2A(x)] is the most possible membership interval of μ A(x) . flA(x) and fhA(x) present the lower limit and upper limit of v A(x) , [fm1A(x),fm2A(x)] is the non-membership interval of v A(x) .
For convenience, we call α=<(tlA,tm1A,tm2A,thA),(flA,fm1A,fm2A,fhA)> an intuitionistic trapezoidal fuzzy number (ITrFN), where
tlA,tm1A,tm2A,thA∈[0,1],flA,fm1A,fm2A,fhA∈[0,1],thA+fhA∈[0,1]. | (2) |
It is clear that the largest and smallest ITFN are α+=<(1,1,1,1),(0,0,0,0)> and α−=<(0,0,0,0),(1,1,1,1)> , respectively. When tm1A=tm2A and fm1A=fm2A , an ITrFN reduces to an ITFN [16].
For example, the product quality attribute in online service trading selection example can be expressed in an ITrFN ((0.1, 0.2, 0.3, 0.4), (0.2, 0.3, 0.5, 0.6)), where 0.1 and 0.4 indicate the lower limit and upper limit of users' satisfactory degree, [0.2, 0.3] means the interval of most possible satisfactory degree; 0.2 and 0.6 denote the lower limit and upper limit of users' dissatisfactory degree, [0.3, 0.5] is the interval of most possible dissatisfactory degree.
Definition 3. [18] Let α1=<(tl1,tm11,tm21,th1),(fl1,fm11,fm21,fh1)> and α2=<(tl2,tm12,tm22,th2),(fl2,fm12, fm22,fh2)> be two ITrFNs and λ>0 , then the containment is:
α1⊆ α2 iff tl1≤tl2,tm11≤tm12,tm21≤tm22,th1≤th2,fl1≥fl2,fm11≥fm12,fm21≥fm22,fh1≥fh2. | (3) |
Some arithmetic operations between ITrFNs α1 and α2 are shown as below:
(1) α1+ α2=<(tl1+tl2−tl1tl2,tm11+tm12−tm11tm12,tm21+tm22−tm21tm22,th1+th2−th1th2),
(fl1fl2,fm11fm12,fm21fm22,fh1fh2)>; |
(2) λ α1=<(1−(1−tl1)λ,1−(1−tm11)λ,1−(1−tm21)λ,1−(1−th1)λ),
((fl1)λ,(fm11)λ,(fm21)λ,(fh1)λ)>; |
(3) α1c=<(fl1,fm11,fm21,fh1),(tl1,tm11,tm21,th1)>
Definition 4. Let α1=<(tl1,tm11,tm21,th1),(fl1,fm11,fm21,fh1)> and α2=<(tl2,tm12,tm22,th2),(fl2,fm12,fm22,fh2)> be two ITrFNs. The Hamming distance d( α1, α2) between the ITFNs α1 and α2 is defined as follows:
d( α1, α2)=18(|tl1−tl2|+|tm11−tm12|+|tm21−tm22|+|th1−th2|+ |
|fl1−fl2|+|fm11−fm12|+|fm21−fm22|+|fh1−fh2|) | (4) |
Theorem 1. The distance measure d( α1, α2) satisfies the following properties:
(i) 0≤d( α1, α2)≤1 .
(ii) d( α1, α2)=0 if and only if α1= α2 .
(iii) d( α1, α2)=d( α2, α1) .
(iv) If α3=<(tl3,tm13,tm23,th3),(fl3,fm13,fm23,fh3)> is an ITrFN and α1≤ α2≤ α3 , then d( α1, α3)≥d( α1, α2) and d( α1, α3)≥d( α2, α3) .
Proof. It is easy to see that the proposed similarity measure d( α1, α2) meets the third property of Theorem 1. We only need to prove (i), (ii) and (iv).
For (i),
By Eq (2), we have
0≤|tl1−tl2|≤1 , 0≤|tm11−tm12|≤1 , 0≤|tm21−tm22|≤1 , 0≤|th1−th2|≤1 , 0≤|fl1−fl2|≤1 , 0≤|fm11−fm12|≤1 , 0≤|fm21−fm22|≤1 , 0≤|fh1−fh2|≤1 .
It is easy to see that
0≤18(|tl1−tl2|+|tm11−tm12|+|tm21−tm22|+|th1−th2|+ |
|fl1−fl2|+|fm11−fm12|+|fm21−fm22|+|fh1−fh2|)≤1 |
0≤12max(|tl1−tl2|,|tm1−tm2|,|th1−th2|,|fl1−fl2|,|fm1−fm2|,|fh1−fh2|)≤12 |
Thus the inequality: 0≤d( α1, α2)≤1 is established.
For (ii),
When d( α1, α2)=1 , if and only if
18(|tl1−tl2|+|tm11−tm12|+|tm21−tm22|+|th1−th2|+|fl1−fl2|+ |
|fm11−fm12|+|fm21−fm22|+|fh1−fh2|)=0 |
Apparently, it's easy to derive
|tl1−tl2|=0 , |tm11−tm12|=0 , tm21−tm22|=0 , |th1−th2|=0 , |fl1−fl2|=0 ,
|fm11−fm12|=0 , |fm21−fm22|=0 , |fh1−fh2|=0 .
Thus we get tl1=tl2 , tm11=tm12 , tm21=tm22 , th1=th2 , fl1=fl2 , fm11=fm12 , fm21=fm22 , fh1−fh2 . And then α1= α2 .
For (iv),
Since
tl1≤tl2≤tl3 , tm11≤tm12≤tm13 , tm21≤tm22≤tm23 , th1≤th2≤th3 , fl1≥fl2≥fl3 , fm11≥fm12≥fm13 , fm21≥fm22≥fm23 , fh1≥fh2≥fh3 ,
We get
|tl1−tl2|≤|tl1−tl3| , |tm11−tm12|≤|tm11−tm13| , |tm21−tm22|≤|tm21−tm23| ,
|th1−th2|≤|th1−th3| , |fl1−fl2|≤|fl1−fl3| , |fm11−fm12|≤|fm11−fm13| ,
|fm21−fm22|≤|fm21−fm23| , |fh1−fh2|≤|fh1−fh3| .
Based on the above inequalities, it's easy to derive
|tl1−tl2|+|tm11−tm12|+|tm21−tm22|+|th1−th2|+|fl1−fl2|+ |
|fm11−fm12|+|fm21−fm22|+|fh1−fh2|≤|tl1−tl3|+|tm11−tm13|+|tm21−tm23|+|th1−th3|+|fl1−fl3|+ |
|fm11−fm13|+|fm21−fm23|+|fh1−fh3| |
and
max(|tl1−tl2|,|tm1−tm2|,|th1−th2|,|fl1−fl2|,|fm1−fm2|,|fh1−fh2|)≤ |
max(|tl1−tl3|,|tm1−tm3|,|th1−th3|,|fl1−fl3|,|fm1−fm3|,|fh1−fh3|) .
Thus, d( α1, α3)≥d( α1, α2) . By the same way, it is proved that d( α1, α3)≥d( α2, α3) .
For example, consider α1=<(0.3,0.4,0.5,0.6),(0.0,0.1,0.2,0.3)> and α2=<(0.5,0.5,0.6,0.6),(0.0,0.1,0.2,0.3)> , α3=<(0.5,0.5,0.7,0.7),(0.0,0.1,0.2,0.3)> are three ITrFNs in [0, 1]. According to Definition 3, we have α1< α2< α3 . By Definition 4, we know d( α1, α2)=0.05 , d( α1, α3)=0.075 , d( α2, α3)=0.025 . Obvious, d( α1, α3)>d( α1, α2) and d( α1, α3)≥d( α2, α3) .
Entropy measure is worthy of investigation in IF environment. It is widely used in the field of decision-making. Burillo and Bustince [10] discussed the entropy on IFSs and interval-valued. Szmidt and Kacprzyk [11] proposed an entropy measure from a geometric point of view. Chen and Li [12] conducted a comparative analysis on determining objective weights with intuitionistic fuzzy entropy measures. Joshi and Kumara [15] discussed the parametric (R, S)-norm IF entropy and applied it to MADM. Some researchers have recently used distance measures to derive fuzzy entropy by extending De Luca's axioms [14]. Liu [23] proposed some entropy measures for fuzzy sets (FSs) based on distances. Zhang and Zhang et al. [24] discussed the entropy of interval-valued FSs based on distance and its relationship with a similarity measure. Zhang and Xing et al. [13] introduced the relationship among distance measures, inclusion measures and fuzzy entropy of IVIFSs. To address the completely unknown attribute weights in MADM problems, Garg [25] proposed some IF Hamacher agammaegation operators based on entropy function to agammaegate the attribute values. Later, Garg [26] developed a generalized IF entropy for IVIFS and applied it to solve MADM problems. This section combines the entropy concept in [13] and TOPSIS method to develop a novel axiomatical definition of entropy measure for ITrFS.
Definition 5. A real-valued function E:ITrFS(X)→[0,1] is called an entropy on ITrFS(X) if it satisfies the following properties:
(EP1) E(A)=0 iff A is a crisp set;
(EP2) E(A)=1 iff d(A,A+)=d(A,A−) for all A∈ITrFS(X) , where d(A,A+) is a distance from A to A+ , and d(A,A−) is a distance from A to A− ;
(EP3) E(A)=E(Ac) for all A∈ITrFS(X) ;
(EP4) For all A,B∈ITrFS(X) , if |d(A,A−)d(A,A−)+d(A,A+)−12|≥|d(B,B−)d(B,B−)+d(B,B+)−12| , then E(A)≤E(B) , where d(B,B+) is a distance from B to B+ , and d(B,B−) is a distance from B to B− .
Remark 1. A new axiomatical definition of distance-based entropy for ITrFS is proposed in Definition 4 based on the idea of TOPSIS. Given a set type, we can define the entropy for the corresponding ITrFSs by using different distance measures between two ITrFSs. The properties in Definition 4 imply the following realities:
(EP1) Crisp sets are not fuzzy;
(EP2) If d(A,A+)=d(A,A−) , then A is the fuzziest set;
(EP3) The fuzziness of a generalized set is equal to that of its complement;
(EP4) An ITrFS is fuzzier when its relative closeness is nearly 0.5.
Theorem 2. Let d be the distance of ITrFS(X) . Then, for any A∈ITrFS(X) ,
E(A)=1−2|d(A,A−)d(A,A−)+d(A,A+)−12| | (5) |
is entropy of F(X) based on TOPSIS.
Proof. We can prove that E(A) meets properties (EP1)-(EP4).
EP1: If A is crisp set, that is, A(X)=<(1,1,1,1),(0,0,0,0)> or A(X)=<(0,0,0,0),(1,1,1,1)> , by using Eq. (5), then we have d(A,A−)d(A,A−)+d(A,A+)=1 or d(A,A−)d(A,A−)+d(A,A+)=0 . Thus, E(A)=1−2|1−12|=0 or E(A)=1−2|0−12|=0 .
EP2: If E(A)=1 , then we have d(A,A−)d(A,A−)+d(A,A+)=12⟺d(A,A−)=d(A,A+) .
EP3: Given d(Ac,A−)=d(A,A+) and d(Ac,A+)=d(A,A−) , then |d(Ac,A−)d(Ac,A−)+d(Ac,A+)−12|=
|12−d(A,A−)d(A,A−)+d(A,A+)| . Thus, E(A)=E(Ac) .
EP4: If |d(A,A−)d(A,A−)+d(A,A+)−12|≥|d(B,B−)d(B,B−)+d(B,B+)−12| , then E(A)≤E(B) can be easily derived.
Remark 2. Consider the distance measure dIVIFN(⋅,⋅) of IVIFNs, for an ITrFN A=<(tlA,tm1A,tm2A,thA),(flA,fm1A,fm2A,fhA)> , if tlA=tm1A , tm2A=thA , flA=fm1A and fm2A=fhA , then A is degenerated to an IVIFN AIVIFN=<[tlA,thA],[flA,fhA]> , the largest IVIFN is A+IVIFN=<[1,1],[0,0]> , the smallest IVIFN is A−IVIFN=<[0,0],[1,1]> , and Eq. (4) is degenerated to the distance measure of IVIFNs dIVIFN( α1, α2)=14(|tl1−tl2|+|th1−th2|+|fl1−fl2|+|fh1−fh2|) . According to Eq (5), the entropy of IVIFN AIVIFN can be calculated as
E( AIVIFN)=1−2|dIVIFN( AIVIFN,A−IVIFN)dIVIFN( AIVIFN,A−IVIFN)+d( AIVIFN,A+IVIFN)−12| | (6) |
Obviously, E( AIVIFN) satisfies properties (EP1)-(EP4). Thus the proposed entropy measure is a generalization of IVIFS.
In this section, we provide a method to address ITrFN MADM problems unknown attribute weight by using the proposed entropy measure.
For the MADM problem, the final decision should be derived from the assessments of all feasible alternatives on multiple attributes. For convenience, some symbols are introduced to characterize the MADM problem as follows.
(1) The set of alternatives is Si(i∈M={1,2,⋯,m}) .
(2) The set of attributes is Aj(j∈N={1,2,⋯,n}) . The attribute weight vector is denoted by w=(w1,w2,⋯,wn) , where wj represents the weight of Aj such that wj∈[0,1] (j∈N) and ∑nj=1wj=1 .
(3) The assessments of alternatives Si on attributes Aj are ITrFNs αij=<(tlij,tm1ij,tm2ij,thij),(flij,fm1ij,fm2ij,fhij)> .
(4) An ITrFN MADM problem can be described by an ITrFN decision matrix D=( αij)m×n .
Attribute weights depend on the certainty and reliability of the assessments given by the decision maker. The objective weight is smaller when the evaluation value is more uncertain. The fuzziness and uncertainty of attribute values can be measured by the fuzzy entropy. According to the entropy-weighting method [9,13,26], we employ the proposed IF entropy measure to determine the weights of the attributes. The decision matrix D=( αij)m×n can be turned into an IF entropy matrix Γ=(Eij)m×n , where
Eij=1−2|d( αij,α−)d( αij,α−)+d( αij,α+)−12|. | (7) |
α−=<(0,0,0,0),(1,1,1,1)> and α+=<(1,1,1,1),(0,0,0,0) are the negative ideal solution (NIS) and positive ideal solution (PIS), respectively.
Then, the normalized entropy matrix H=(hij)m×n is obtained as follows:
hij=Eijmax{Ei1,Ei2,⋯,Ein}. | (8) |
The objective attribute weights are determined by
wj=1−∑mi=1Eij1−∑nj=1∑mi=1Eiji={1,2,⋯,m},j={1,2,⋯,n}. | (9) |
Evidently, attribute weight wj is inversely proportional to the summation of the entropy values of attribute Aj . In other words, if the values of the attribute are vaguer and more unreliable, then we assign a lower weight; otherwise, a higher weight is attached.
This section extends TOPSIS to agammaegate ITrFNs and rank alternatives. Suppose that PIS and NIS are R+=(α+1,α+2,⋯,α+m) and R−=(α−1,α−2,⋯,α−m) , respectively, where α+j=<(1,1,1,1),(0,0,0,0)> and α−j=<(0,0,0,0),(1,1,1,1)> for benefit attributes and α+j=<(0,0,0,0),(1,1,1,1)> and α−j=<(1,1,1,1),(0,0,0,0)> for cost attributes. In the decision matrix D=( αij)m×n , the separation measures from alternative Si to PIS α+ and NIS α− can be defined as follows.
Definition 6. The weighted positive separation measure between alternative Si and PIS is defined as follows:
G+i=∑nj=1wjd( αij,α+j), | (10) |
where d( αij,α+j) is the distance from αij to α+j , and wj is the attribute weight of attribute Aj .
Definition 7. The weighted positive separation measure between alternative Si and NIS is defined as follows:
G−i=∑nj=1wjd( αij,α−j), | (11) |
where d( αij,α−j) is the distance from αij to α−j , and wj is the attribute weight of Aj .
Then, a closeness coefficient to the PIS and NIS for each alternative is calculated as follows:
RCi=G−iG−i+G+i. | (12) |
Evidently, alternative Si is better when RCi is larger.
This section presents a procedure for solving MADM problems with unknown attribute weights under an ITrFN environment; it can be summarized in the following steps:
Step 1. Provide the decision matrix D=( αij)m×n .
Step 2. Calculate IF entropy matrix using Eq (7).
Step 3. Determine the weight vector of attributes by Eq (9).
Step 4. Identify the PIS and NIS and compute the separation measures from each alternative to PIS and NIS using Eqs (10) and (11), respectively.
Step 5. Construct the closeness coefficient of alternatives according to Eq (12).
Step 6. Rank the alternatives according to the closeness coefficient and select the best one.
The detailed decision process of the proposed method is shown in Figure 1.
Online service trading generally transpires between autonomous parties in an environment where the buyer often has insufficient information about the seller and goods. Many scholars believe that trust is a prerequisite for successful trading. Therefore, buyers must be able to identify the most trustworthy seller. Suppose that a consumer desires to select a reliable seller. After preliminary screening, four candidate sellers S1 , S2 , S3 and S4 remain to be further evaluated. Based on detailed seller ratings, the consumer assesses the four candidate sellers according to five trust factors, namely, product quality (A1), service attitude (A2), website usability (A3), response time (A4) and shipping speed (A5). The first three attributes are benefit attributes, whereas the last two are cost attributes. The decision maker provides the lower and upper limits and the most possible intervals to describe these attributes. The candidate sellers' ratings concerning the attributes can be represented as ITrFNs by using statistical methods, as shown in Table 1.
A1 | A2 | A3 | A4 | A5 | |
S1 | < (0.3, 0.4, 0.5, 0.5), (0.2, 0.3, 0.3, 0.4) > | < (0.1, 0.4, 0.5, 0.6), (0.1, 0.2, 0.3, 0.3) > | < (0.4, 0.4, 0.5, 0.6), (0.1, 0.2, 0.3, 0.4 > | < (0.3, 0.4, 0.5, 0.5), (0.1, 0.2, 0.3, 0.5) > | < (0.2, 0.4, 0.5, 0.5), (0.3, 0.3, 0.5, 0.5) > |
S2 | < (0.1, 0.2, 0.3, 0.4), (0.2, 0.3, 0.4, 0.5) > | < (0.1, 0.2, 0.3, 0.3), (0.1, 0.3, 0.4, 0.5) > | < (0.2, 0.3, 0.4, 0.4), (0.2, 0.3, 0.4, 0.6 > | < (0.1, 0.2, 0.3, 0.3), (0.2, 0.3, 0.4, 0.6) > | < (0.2, 0.2, 0.3, 0.4), (0.3, 0.5, 0.5, 0.5) > |
S3 | < (0.3, 0.4, 0.5, 0.5), (0.1, 0.2, 0.3, 0.5) > | < (0.2, 0.3, 0.5, 0.6), (0.1, 0.2, 0.3, 0.4) > | < (0.0, 0.2, 0.3, 0.3), (0.1, 0.3, 0.4, 0.5) > | < (0.2, 0.3, 0.4, 0.4), (0.2, 0.2, 0.4, 0.4) > | < (0.3, 0.3, 0.4, 0.4), (0.1, 0.2, 0.5, 0.5) > |
S4 | < (0.1, 0.2, 0.4, 0.5), (0.1, 0.3, 0.4, 0.5) > | < (0.0, 0.1, 0.2, 0.3), (0.2, 0.3, 0.5, 0.6) > | < (0.0, 0.1, 0.3, 0.4), (0.2, 0.3, 0.4, 0.6) > | < (0.1, 0.2, 0.4, 0.4), (0.2, 0.2, 0.3, 0.4) > | < (0.1, 0.3, 0.4, 0.4), (0.1, 0.3, 0.4, 0.4) > |
Step 1. Form a decision matrix that is listed in Table 1.
Step 2. Since A1, A2, A3 are benefit attributes and A4, A5 are cost attributes, we have the following PIS and NIS:
R+=(<(1,1,1,1),(0,0,0,0)>,<(1,1,1,1),(0,0,0,0)>,<(1,1,1,1),(0,0,0,0)>,<(0,0,0,0),(1,1,1,1)>,<(0,0,0,0),(1,1,1,1)>) |
R−=(<(0,0,0,0),(1,1,1,1)>,<(0,0,0,0),(1,1,1,1)>,<(0,0,0,0),(1,1,1,1)>,<(1,1,1,1),(0,0,0,0)>,<(1,1,1,1),(0,0,0,0)>) |
Using Eq (7), the decision matrix turns into IF entropy matrix as follows
Γ=(0.8750.8250.7750.8251.0000.9000.9000.9500.8500.8250.8500.8500.8750.9750.9750.9750.7500.8251.0001.000). |
Step 3. Utilizing Eq (9), the attribute weight vector is determined as w=(0.20,0.20,0.19,0.20,0.21)T .
Step 4. By Eqs (10) and (11), The positive and negative weighted separation are obtained as G+=(0.465,0.492,0.487,0.545) and G−=(0.535,0.508,0.513,0.455) .
Step 5. Using Eq (12), the closeness coefficients of each seller are calculated as RC=(0.535,0.508,0.513,0.455) .
Step 6. Since RC1>RC3>RC2>RC4 , the best seller is S1 .
Given different attribute weights will produce various decision results, this section carries out a sensitivity analysis on attribute weights to observe whether different attribute weights will lead to a different ranking of four trustworthy sellers. After expert discussion, the weight of product quality, service attitude and website usability are correct. I analyze the seller's ranking, in the case that the weights meet w4+w5=0.41 . When 0≤w4≤0.24 , the ranking of four trustworthy sellers is S1>S3>S2>S4 . If w4=0.25 , their ranking is S1>S3=S2>S4 . When 0.26≤w4≤0.35 , their ranking is S1>S2>S3>S4 . When 0.36≤w4≤0.41 , their ranking is S2>S1>S3>S4 . The above results reveal the importance of attribute weights in decision-making.
This section performs a comparison with the MADM method based on the generalized IF entropy developed by Garg [26]. We use the proposed method for solving the supplier selection problem given in [26] by appropriate modifications, given that the attribute ratings are in the form of IVIFNs. Specifically, when tlA=tm1A , tm2A=thA , flA=fm1A and fm2A=fhA , the ITrFN α is reduced to an IVIFN, α+=<[1,1],[0,0]> , α−=<[0,0],[1,1]> ; and the distance measure in Eq (4) is reduced to d( α1, α2)=14(|tl1−tl2|+|th1−th2|+|fl1−fl2|+|fh1−fh2|) . According to Eq (6), we have the entropy of each IVIFN. Using the proposed decision procedure for MADM, the ranking order of suppliers is as follows: A4≻A5≻A3≻A2≻A1 , which is the same as that obtained by the method in [26]. Hence, the proposed method is suitable for MADM problems with unknown attribute weight under an IVIF environment. The method in [26] cannot address decision problems with ITrFNs. Moreover, the proposed method is superior in terms of using generalized ITrFNs in comparison with the IVIFNs employed in [26]. The proposed method also has shortcomings. For example, it is not suitable for MADM problems with incomplete weight attribute information under ITrFSs environment. To solve this problem, we can define the cross-entropy of ITrFSs by learning from the cross-entropy of IFSs [27]. Then, the programming models can be constructed based on the cross-entropy of ITrFSs to obtain attribute weights.
This study presented a TOPSIS-based entropy method to solve MADM problems with ITrFNs and unknown attribute weight information. We applied ITrFNs for MADM problems to address the imprecise and vague decision data in the actual MADM environment. We developed a distance measure for ITrFNs and discussed its properties. We put forward a TOPSIS-based entropy measure for ITrFNs, in which the entropy axioms for ITrFNs are easy to understand and compute because they only require identifying the largest and least values. Further, we provided an objective attribute weight method by using the proposed entropy measure. Then, by combining TOPSIS and entropy-weighted approach, a MADM method was proposed to select the best alternative. Finally, an online trustworthy service evaluation example indicated that the proposed MADM method is practical and useful. Our future research will cover the following three aspects. (1) We will construct additional entropy measures of ITrFNs and study the relationship between the entropy and similarity measure of ITrFNs. (2) We will extend the proposed method to a decision environment with linguistic interval-valued Atanassov IFSs [28]. (3) The proposed method will be used for large group decision-making problems [29] by integrating the evaluation information into ITrFNs. (4) The proposed method will be applied to the evaluation of text classification [30] and financial risk analysis [31] in ITrFSs environment.
This work was supported by the National Natural Science Foundation of China (Nos. 61602219 and 71662014), the Science and Technology Project of Jiangxi Province Education Department of China (No. GJJ181482) and the Natural Science Fund Project of Jiangxi science and Technology Department (No. 20202BABL202027)
The authors declare there is no conflict of interest.
[1] | A. Si, S. Das, S. Kar, An approach to rank picture fuzzy numbers for decision making problems, Decis. Making Appl. Manage. Eng., 2 (2019), 54-64. |
[2] | I. Petrovic, M. Kankaras, A hybridized IT2FS-DEMATEL-AHP-TOPSIS multi-criteria decision making approach: Case study of selection and evaluation of criteria for determination of air traffic control radar position, Decis. Making Appl. Manage. Eng., 3 (2020), 146-164. |
[3] | S. Biswas, G. Bandyopadhyay, B. Guha, & M. Bhattacharjee, An ensemble approach for portfolio selection in a multi-criteria decision making framework, Decis. Making Appl. Manage. Eng., 2 (2019), 138-158. |
[4] | K. T. Atanassov, Intuitionistic fuzzy sets, Fuzzy Sets Syst., 20 (1986), 87-96. |
[5] | K. Atanassov, G. Gargov, Interval valued intuitionistic fuzzy sets, Fuzzy Sets Syst., 31 (1989), 343-349. |
[6] | H. Garg, K. Kumar, A novel exponential distance and its based TOPSIS method for interval-valued intuitionistic fuzzy sets using connection number of SPA theory, Artif. Intell. Rev., 1 (2020), 595-624. |
[7] | K. Kumar, H. Garg, Connection number of set pair analysis based TOPSIS method on intuitionistic fuzzy sets and their application to decision making, Appl. Intell., 48 (2018), 2112-2119. |
[8] | K. Kumar, H. Garg, TOPSIS method based on the connection number of set pair analysis under interval-valued intuitionistic fuzzy set environment, Comput. Appl. Math., 37 (2018), 1319-1329. |
[9] | H. Garg, K. Kumar, An extended technique for order preference by similarity to ideal solution group decision-making method with linguistic interval-valued intuitionistic fuzzy information, J. Multi. Crite. Decis. Anal., 26 (2019), 16-26. |
[10] | P. Burillo, H. Bustince, Entropy on intuitionistic fuzzy sets and on interval-valued fuzzy sets, Fuzzy Sets Syst., 118 (1996), 305-316. |
[11] | E. Szmidt, J. Kacprzyk, Entropy for intuitionistic fuzzy sets, Fuzzy Sets Syst., 118 (2001), 467-477. |
[12] | T. Y. Chen, C. H. Li, Determing objective weights with intuitionistic fuzzy entropy measures: A comparative analysis, Inf. Sci., 180 (2010), 4207-4222. |
[13] | Q. S. Zhang, H. Y. Xing, F. C. Liu, J. Ye, P. Tang, Some new entropy measures for interval-valued intuitionistic fuzzy sets based on distances and their relationships with similarity and inclusion measures, Inf. Sci., 283 (2014), 55-69. |
[14] | A. De Luca, S. Termini, A definition of a nonprobabilistic entropy in the setting of fuzzy sets theory, Inf. Control, 20 (1972), 301-312. |
[15] | R. Joshi, S. Kumar, Parametric (R, S)-norm entropy on intuitionistic fuzzy sets with a new approach in multiple attribute decision making, Fuzzy Inf. Eng., 9 (2019), 181-203. |
[16] | F. Liu, X. H. Yuan, Fuzzy number intuitionistic fuzzy set, Fuzzy Syst. Math., 21 (2007), 88-91. |
[17] | M. H. Shu, C. H. Cheng, J. R. Dong, Using intuitionistic fuzzy sets for fault-tree analysis on printed circuit board assembly, Microelectron. Reliab., 46 (2006), 2139-2148. |
[18] | X. F. Wang, Fuzzy number intuitionistic fuzzy geometric aggregation operators and their application to decision making, Control Decis., 10 (2008), 104-111. |
[19] | G. W. Wei, Induced fuzzy number intuitionistic fuzzy ordered weighted averaging (I-FIFOWA) operator and its application to group decision making, Chin. Jou. Manage., 7 (2010), 903-908. |
[20] | Y. Gao, D. Q. Zhou, C. C. Liu, L. Zhang, Triangular fuzzy number intuitionistic fuzzy aggregation operators and their application based on interaction, Syst. Eng. Theory Prac., 32 (2012), 1964-1972. |
[21] | Shan, Yu, Z. S. Xu, Aggregation and decision making using intuitionistic multiplicative triangular fuzzy information, J. Syst. Sci. Syst. Eng., 23 (2014), 20-38. |
[22] | C. L. Hwang, K. Yoon, Multiple attributes decision making methods and applications, Berlin Heidelberg: Springer, (1981). |
[23] | X. C. Liu, Entropy, distance measure and similarity measure of fuzzy sets and their relations, Fuzzy Sets Syst., 52 (1992), 305-318. |
[24] | H. Y. Zhang, W. X. Zhang, C. L. Mei, Entropy of interval-valued fuzzy sets based on distance and its relationship with similarity measure, Knowledge-Based Syst., 22 (2009), 449-454. |
[25] | H. Garg, Intuitionistic fuzzy Hamacher aggregation operators with entropy weight and their applications to multi-criteria decision-making problems, Iran. J. Sci. Technol.-Trans. Electr. Eng., 43 (2019), 597-613. |
[26] | H. Garg, Generalized intuitionistic fuzzy entropy-based approach for solving multi-attribute decision-making problems with unknown attribute weights, Proc. Natl. Acad. Sci. ndia Sect. A, 89 (2019), 129-139. |
[27] | X. W. Qi, C. Y. Liang, J. L. Zhang, Generalized cross-entropy based group decision making with unknown expert and attribute weights under interval-valued intuitionistic fuzzy environment, Comput. Ind. Eng., 79 (2015), 52-64. |
[28] | H. Garg, K. Kumar, Linguistic interval-valued Atanassov intuitionistic fuzzy sets and their applications to group decision-making problems, IEEE Trans. Fuzzy Syst., 27 (2019), 2302-2311. |
[29] | H. Zhang, G. Kou, Y. Peng, Soft consensus cost models for group decision making and economic interpretations, Eur. J. Oper. Res., 277 (2019), 964-980. |
[30] | G. Kou, P. Yang, Y. Peng, F. Xiao, Y. Chen, F. E. Alsaadi, Evaluation of feature selection methods for text classification with small datasets using multiple criteria decision-making methods, Appl. Soft Comput., 86 (2020), 105836. |
[31] | G. Kou, Y. Peng, G. Wang, Evaluation of clustering algorithms for financial risk analysis using MCDM methods, Inf. Sci., 275 (2014), 1-12. |
1. | Xin Kang, Xiangjun Xu, Zaoli Yang, Evaluation and selection of green suppliers for papermaking enterprises using the interval basic probability assignment-based intuitionistic fuzzy set, 2022, 8, 2199-4536, 4187, 10.1007/s40747-022-00691-z | |
2. | Junli Cao, Fangfang Xu, Qiangyi Li, Entropy-Based Fuzzy TOPSIS Method for Investment Decision Optimization of Large-Scale Projects, 2022, 2022, 1687-5273, 1, 10.1155/2022/4381293 | |
3. | Nian Zhang, Yifan Zhou, Qiang Pan, Guiwu Wei, Multi-attribute decision-making method with triangular fuzzy numbers based on regret theory and the catastrophe progression method, 2022, 19, 1551-0018, 12013, 10.3934/mbe.2022559 | |
4. | Yanjie Fan, Ying Zhou, Huili Liu, Xiaomei Luo, Ting Xu, Yu Sun, Tingting Yang, Linlin Chen, Xuefan Gu, Yongguo Yu, Improving variant prioritization in exome analysis by entropy‐weighted ensemble of multiple tools, 2023, 103, 0009-9163, 190, 10.1111/cge.14257 | |
5. | Arunodaya Raj Mishra, Pratibha Rani, Multi-criteria healthcare waste disposal location selection based on Fermatean fuzzy WASPAS method, 2021, 7, 2199-4536, 2469, 10.1007/s40747-021-00407-9 | |
6. | Muhammet Deveci, Raghunathan Krishankumar, Ilgin Gokasar, Rumeysa Tuna Deveci, Prioritization of healthcare systems during pandemics using Cronbach’s measure based fuzzy WASPAS approach, 2022, 0254-5330, 10.1007/s10479-022-04714-3 | |
7. | Xin Kang, Xiangjun Xu, Fei Yuan, Alessandro Rasulo, Investment Decision of New Energy Vehicle Enterprises Using the Interval Basic Probability Assignment-Based Intuitionistic Fuzzy Set, 2021, 2021, 1563-5147, 1, 10.1155/2021/5585461 | |
8. | Pingqing Liu, Hongjun Wang, Baoquan Ning, Guiwu Wei, MAGDM method based on generalized hesitant fuzzy TODIM and cumulative prospect theory and application to recruitment of university researchers, 2023, 45, 10641246, 1863, 10.3233/JIFS-224437 | |
9. | Osman Taylan, Bulent Guloglu, Abdulaziz S. Alkabaa, Salih Sarp, Hassan M. Alidrisi, Ahmad H. Milyani, Hisham Alidrisi, Mohammed Balubaid, AI based fuzzy MCDM models: Comparison and evaluation of dissimilar outcomes, an application to enhance pilot recruitment process, 2024, 41, 0266-4720, 10.1111/exsy.13590 | |
10. | Raghunathan Krishankumar, K.S. Ravichandran, Manish Aggarwal, Dragan Pamucar, An improved entropy function for the intuitionistic fuzzy sets with application to cloud vendor selection, 2023, 7, 27726622, 100262, 10.1016/j.dajour.2023.100262 | |
11. | Vakkas Uluçay, Irfan Deli, Vikor method based on the entropy measure for generalized trapezoidal hesitant fuzzy numbers and its application, 2023, 1432-7643, 10.1007/s00500-023-09257-8 | |
12. | Anjali Patel, Subhankar Jana, Juthika Mahanta, Intuitionistic fuzzy EM-SWARA-TOPSIS approach based on new distance measure to assess the medical waste treatment techniques, 2023, 144, 15684946, 110521, 10.1016/j.asoc.2023.110521 | |
13. | Weidong Wu, Chengyu Xie, Shuai Geng, Honglin Lu, Juan Yao, Intuitionistic fuzzy-based entropy weight method–TOPSIS for multi-attribute group decision-making in drilling fluid waste treatment technology selection, 2023, 195, 0167-6369, 10.1007/s10661-023-11724-6 | |
14. | Anjali Patel, Subhankar Jana, Juthika Mahanta, Construction of similarity measure for intuitionistic fuzzy sets and its application in face recognition and software quality evaluation, 2024, 237, 09574174, 121491, 10.1016/j.eswa.2023.121491 | |
15. | Jiaxin Feng, Zhaoyu Guo, Peijun Qian, Wenya Wang, Hehua Hu, Xia Zhang, JingBo Xue, Yinlong Li, Chunli Cao, Yuwan Hao, Shizhu Li, Uwem Friday Ekpo, Field validation of different intervention modes for the potential transmission risk of schistosomiasis during post-transmission interruption phase, 2023, 17, 1935-2735, e0011739, 10.1371/journal.pntd.0011739 |
A1 | A2 | A3 | A4 | A5 | |
S1 | < (0.3, 0.4, 0.5, 0.5), (0.2, 0.3, 0.3, 0.4) > | < (0.1, 0.4, 0.5, 0.6), (0.1, 0.2, 0.3, 0.3) > | < (0.4, 0.4, 0.5, 0.6), (0.1, 0.2, 0.3, 0.4 > | < (0.3, 0.4, 0.5, 0.5), (0.1, 0.2, 0.3, 0.5) > | < (0.2, 0.4, 0.5, 0.5), (0.3, 0.3, 0.5, 0.5) > |
S2 | < (0.1, 0.2, 0.3, 0.4), (0.2, 0.3, 0.4, 0.5) > | < (0.1, 0.2, 0.3, 0.3), (0.1, 0.3, 0.4, 0.5) > | < (0.2, 0.3, 0.4, 0.4), (0.2, 0.3, 0.4, 0.6 > | < (0.1, 0.2, 0.3, 0.3), (0.2, 0.3, 0.4, 0.6) > | < (0.2, 0.2, 0.3, 0.4), (0.3, 0.5, 0.5, 0.5) > |
S3 | < (0.3, 0.4, 0.5, 0.5), (0.1, 0.2, 0.3, 0.5) > | < (0.2, 0.3, 0.5, 0.6), (0.1, 0.2, 0.3, 0.4) > | < (0.0, 0.2, 0.3, 0.3), (0.1, 0.3, 0.4, 0.5) > | < (0.2, 0.3, 0.4, 0.4), (0.2, 0.2, 0.4, 0.4) > | < (0.3, 0.3, 0.4, 0.4), (0.1, 0.2, 0.5, 0.5) > |
S4 | < (0.1, 0.2, 0.4, 0.5), (0.1, 0.3, 0.4, 0.5) > | < (0.0, 0.1, 0.2, 0.3), (0.2, 0.3, 0.5, 0.6) > | < (0.0, 0.1, 0.3, 0.4), (0.2, 0.3, 0.4, 0.6) > | < (0.1, 0.2, 0.4, 0.4), (0.2, 0.2, 0.3, 0.4) > | < (0.1, 0.3, 0.4, 0.4), (0.1, 0.3, 0.4, 0.4) > |
A1 | A2 | A3 | A4 | A5 | |
S1 | < (0.3, 0.4, 0.5, 0.5), (0.2, 0.3, 0.3, 0.4) > | < (0.1, 0.4, 0.5, 0.6), (0.1, 0.2, 0.3, 0.3) > | < (0.4, 0.4, 0.5, 0.6), (0.1, 0.2, 0.3, 0.4 > | < (0.3, 0.4, 0.5, 0.5), (0.1, 0.2, 0.3, 0.5) > | < (0.2, 0.4, 0.5, 0.5), (0.3, 0.3, 0.5, 0.5) > |
S2 | < (0.1, 0.2, 0.3, 0.4), (0.2, 0.3, 0.4, 0.5) > | < (0.1, 0.2, 0.3, 0.3), (0.1, 0.3, 0.4, 0.5) > | < (0.2, 0.3, 0.4, 0.4), (0.2, 0.3, 0.4, 0.6 > | < (0.1, 0.2, 0.3, 0.3), (0.2, 0.3, 0.4, 0.6) > | < (0.2, 0.2, 0.3, 0.4), (0.3, 0.5, 0.5, 0.5) > |
S3 | < (0.3, 0.4, 0.5, 0.5), (0.1, 0.2, 0.3, 0.5) > | < (0.2, 0.3, 0.5, 0.6), (0.1, 0.2, 0.3, 0.4) > | < (0.0, 0.2, 0.3, 0.3), (0.1, 0.3, 0.4, 0.5) > | < (0.2, 0.3, 0.4, 0.4), (0.2, 0.2, 0.4, 0.4) > | < (0.3, 0.3, 0.4, 0.4), (0.1, 0.2, 0.5, 0.5) > |
S4 | < (0.1, 0.2, 0.4, 0.5), (0.1, 0.3, 0.4, 0.5) > | < (0.0, 0.1, 0.2, 0.3), (0.2, 0.3, 0.5, 0.6) > | < (0.0, 0.1, 0.3, 0.4), (0.2, 0.3, 0.4, 0.6) > | < (0.1, 0.2, 0.4, 0.4), (0.2, 0.2, 0.3, 0.4) > | < (0.1, 0.3, 0.4, 0.4), (0.1, 0.3, 0.4, 0.4) > |