Access the full text.
Sign up today, get DeepDyve free for 14 days.
R. Krohling, T. Souza (2012)
Combining prospect theory and fuzzy numbers to multi-criteria decision makingExpert Syst. Appl., 39
C. Hwang, K. Yoon (1981)
Multiple Attribute Decision Making: Methods and Applications - A State-of-the-Art Survey, 186
(1966)
Electre: une methode pour guider le choix en presence de points de vue multiples, Note de travail 49. Direction Scientifique: SEMAMETRA International
C. Shannon (1948)
A mathematical theory of communicationBell Syst. Tech. J., 27
Shahzaib Ashraf, T. Mahmood, S. Abdullah, Q. Khan (2018)
Different Approaches to Multi-Criteria Group Decision Making Problems for Picture Fuzzy EnvironmentBulletin of the Brazilian Mathematical Society, New Series, 50
K. Kumar, Harish Garg (2018)
Connection number of set pair analysis based TOPSIS method on intuitionistic fuzzy sets and their application to decision makingApplied Intelligence, 48
A. Bonaert (1977)
Introduction to the theory of Fuzzy subsetsProceedings of the IEEE, 65
Harish Garg, Jaspreet Kaur (2018)
A Novel (R,S)-Norm Entropy Measure of Intuitionistic Fuzzy Sets and Its Applications in Multi-Attribute Decision-MakingMathematics
Vikas Arya, Satish Kumar (2020)
Knowledge measure and entropy: a complementary concept in fuzzy theoryGranular Computing, 6
N. Pal, S. Pal (1989)
Object-background segmentation using new definitions of entropy, 136
K. Ullah, Nasruddin Hassan, T. Mahmood, Naeem Jan, Mazlan Hassan (2019)
Evaluation of Investment Policy Based on Multi-Attribute Decision-Making Using Interval Valued T-Spherical Fuzzy Aggregation OperatorsSymmetry, 11
(1975)
Onmeasures of information and their characterization
Z. Fan, Xiao Zhang, Fa-Dong Chen, Yang Liu (2013)
Extended TODIM method for hybrid multiple attribute decision making problemsKnowl. Based Syst., 42
F. Boran, S. Genç, D. Akay (2011)
Personnel selection based on intuitionistic fuzzy setsHuman Factors and Ergonomics in Manufacturing & Service Industries, 21
D. Hong, Chang-Hwan Choi (2000)
Multicriteria fuzzy decision-making problems based on vague set theoryFuzzy Sets Syst., 114
Mehdi Ghorabaee, E. Zavadskas, M. Amiri, A. Esmaeili (2016)
Multi-criteria evaluation of green suppliers using an extended WASPAS method with interval type-2 fuzzy setsJournal of Cleaner Production, 137
Harish Garg, Jeonghwan Gwak, T. Mahmood, Zeeshan Ali (2020)
Power Aggregation Operators and VIKOR Methods for Complex q-Rung Orthopair Fuzzy Sets and Their ApplicationsMathematics
Harish Garg (2017)
Generalized Intuitionistic Fuzzy Entropy-Based Approach for Solving Multi-attribute Decision-Making Problems with Unknown Attribute WeightsProceedings of the National Academy of Sciences, India Section A: Physical Sciences, 89
Jin Shin, Sang Kim (2006)
A Mathematical Theory of Communication
Jiahang Yuan, Cunbin Li, Wenle Li, Ding Liu, Xiaopeng Li (2018)
Linguistic hesitant fuzzy multi-criterion decision-making for renewable energy: A case study in JilinJournal of Cleaner Production, 172
A. Rényi (1961)
On Measures of Entropy and Information
Da-Zeng Tian, Zhong-Tang Yang (2014)
An exponential entropy on intuitionistic fuzzy sets2014 International Conference on Machine Learning and Cybernetics, 1
L. Gomes, L. Rangel (2009)
An application of the TODIM method to the multicriteria rental evaluation of residential propertiesEur. J. Oper. Res., 193
W. Hung, Miin-Shen Yang (2006)
Fuzzy entropy on intuitionistic fuzzy setsInternational Journal of Intelligent Systems, 21
(1965)
Fuzzy sets, 8
Changzhong Wang, Mingwen Shao, Qiang He, Y. Qian, Yali Qi (2016)
Feature subset selection based on fuzzy neighborhood rough setsKnowl. Based Syst., 111
(2018)
Anovel (R, S)-normentropymeasureof intuitionistic fuzzy sets and its applications in multi-attribute decision-making
Jun Ye (2010)
Two effective measures of intuitionistic fuzzy entropyComputing, 87
(2008)
A note on information entropy measure for vague sets
(1981)
Multiple attribute decision making: methods and applications
K. Ullah, Harish Garg, T. Mahmood, Naeem Jan, Zeeshan Ali (2020)
Correlation coefficients for T-spherical fuzzy sets and their applications in clustering and multi-attribute decision makingSoft Computing, 24
(2013)
Extended TODIMmethod for hybrid multiple attribute decision making problems
E. Szmidt, J. Kacprzyk, Paweł Bujnowski (2014)
How to measure the amount of knowledge conveyed by Atanassov's intuitionistic fuzzy setsInf. Sci., 257
Muhammad Ali, Feng Feng, T. Mahmood, Imran Mahmood, Haider Faizan (2019)
A graphical method for ranking Atanassov's intuitionistic fuzzy values using the uncertainty index and entropyInternational Journal of Intelligent Systems, 34
M. Higashi, G. Klir (1982)
ON MEASURES OF FUZZINESS AND FUZZY COMPLEMENTSInternational Journal of General Systems, 8
Jian-san Zhao, Xiao-Yue You, Hu-chen Liu, Song-Man Wu (2017)
An Extended VIKOR Method Using Intuitionistic Fuzzy Sets and Combination Weights for Supplier SelectionSymmetry, 9
(1998)
Multicriteria optimization of civil engineering systems
Harish Garg, K. Kumar (2018)
A novel exponential distance and its based TOPSIS method for interval-valued intuitionistic fuzzy sets using connection number of SPA theoryArtificial Intelligence Review, 53
S. De, R. Biswas, A. Roy (2000)
Some operations on intuitionistic fuzzy setsFuzzy Sets Syst., 114
(2020)
S.Multi-criteria decisionmaking problem for evaluating ERP systemusing entropy weighting approach and q-rung orthopair fuzzy TODIM approach
Harish Garg, R. Arora (2020)
TOPSIS method based on correlation coefficient for solving decision-making problems with intuitionistic fuzzy soft set information, 5
D. Boekee, J. Lubbe (1980)
The R-Norm Information MeasureInf. Control., 45
Zeshui Xu, Hui Hu (2010)
Projection Models for Intuitionistic Fuzzy Multiple Attribute Decision MakingInt. J. Inf. Technol. Decis. Mak., 9
(1967)
Quantification method classification process: concept of structural αentropy
Cuiping Wei, Z. Ren, Rosa Rodríguez (2015)
A Hesitant Fuzzy Linguistic TODIM Method Based on a Score FunctionInt. J. Comput. Intell. Syst., 8
P. Yu (1973)
A Class of Solutions for Group Decision ProblemsManagement Science, 19
R. Joshi, Satish Kumar (2017)
Parametric (R,S)-norm Entropy on Intuitionistic Fuzzy Sets with a New Approach in Multiple Attribute Decision MakingFuzzy Information and Engineering, 9
Shyi-Ming Chen, J. Tan (1994)
Handling multicriteria fuzzy decision-making problems based on vague set theoryFuzzy Sets and Systems, 67
B. Mareschal, J. Brans, P. Vincke (1984)
Prométhée: a new family of outranking methods in multicriteria analysisULB Institutional Repository
R. Verma, B. Sharma (2013)
Exponential entropy on intuitionistic fuzzy setsKybernetika, 49
(1972)
A definition of non-probabilistic entropy in the setting of fuzzy set theory
E. Szmidt, J. Kacprzyk (2001)
Entropy for intuitionistic fuzzy setsFuzzy Sets Syst., 118
R. Yager (1979)
ON THE MEASURE OF FUZZINESS AND NEGATION Part I: Membership in the Unit IntervalInternational Journal of General Systems, 5
(1975)
A study of generalized measures of information theory [Ph.D. thesis
Robert Lin (2014)
NOTE ON FUZZY SETSYugoslav Journal of Operations Research, 24
R. Krohling, André Pacheco, André Siviero (2013)
IF-TODIM: An intuitionistic fuzzy TODIM to multi-criteria decision makingKnowl. Based Syst., 53
(1984)
PROMETHEE: a new family of outrankingmethods inmulticriteria analysis. Multiple Criteria Decision Analysis: State of the Art Surveys, Editor: Brans JP
Rodolfo Lourenzutti, R. Krohling (2013)
A study of TODIM in a intuitionistic fuzzy and random environmentExpert Syst. Appl., 40
K. Atanassov (1986)
Intuitionistic fuzzy setsFuzzy Sets and Systems, 20
Multi - criteria decisionmaking problem for evaluating ERP systemusing entropy weighting approach and q - rung orthopair fuzzy TODIM approach
Harish Garg, Nikunj Agarwal, A. Tripathi (2017)
Generalized Intuitionistic Fuzzy Entropy Measure of Order α and Degree β and Its Applications to Multi-Criteria Decision Making ProblemInt. J. Fuzzy Syst. Appl., 6
(2007)
Intuitionistic fuzzy aggregations operators
Wenyi Zeng, Hongxing Li (2006)
Relationship between similarity measure and entropy of interval valued fuzzy setsFuzzy Sets Syst., 157
(1988)
Possible generalization of Boltzman–Gibbs statistics
(1966)
Electre: unemethode pour guider le choix en presence de points de vue multiples, Note de travail 49
S. Arimoto (1971)
Information-Theoretical Considerations on Estimation ProblemsInf. Control., 19
L. Zadeh (1968)
Probability measures of Fuzzy eventsJournal of Mathematical Analysis and Applications, 23
(1975)
A study of generalized measures of information theory
Qian-Sheng Zhang, Sheng-yi Jiang (2008)
A note on information entropy measures for vague sets and its applicationsInf. Sci., 178
(1980)
The R-norm information
Harish Garg, Gagandeep Kaur (2018)
Extended TOPSIS method for multi-criteria group decision-making problems under cubic intuitionistic fuzzy environmentScientia Iranica
P. Burillo, H. Bustince (1996)
Entropy on intuitionistic fuzzy sets and on interval-valued fuzzy setsFuzzy Sets Syst., 78
Vikas Arya, Satish Kumar (2020)
Fuzzy Entropy Measure with an Applications in Decision Making Under Bipolar Fuzzy Environment based on TOPSIS MethodInternational journal of information and management sciences, 31
Dinabandhu Bhandari, N. Pal (1993)
Some new information measures for fuzzy setsInf. Sci., 67
(2006)
Fuzzy entropyon intuitionistic fuzzy sets
FUZZY INFORMATION AND ENGINEERING 2019, VOL. 11, NO. 3, 357–384 https://doi.org/10.1080/16168658.2020.1840317 A Novel VIKOR-TODIM Approach Based on Havrda–Charvat–Tsallis Entropy of Intuitionistic Fuzzy Sets to Evaluate Management Information System Vikas Arya and Satish Kumar Department of Mathematics, Maharishi Markandeshwar (Deemed to be University), Mullana-Ambala, India ABSTRACT ARTICLE HISTORY Received 6 September 2019 The utilisation of Management information system has grabbed the Revised 20 August 2020 attention since last decades. The selection of management sources Accepted 14 October 2020 is a difficult assignment for decision-experts. Thus, the present article develops a new technique using entropy measures and a KEYWORDS novel extended VIKOR-TODIM model of Intuitionistic Fuzzy Sets. A Intuitionistic fuzzy set; new parametric entropy measure based on Havrda–Charvat–Tsallis Havrda–Charvat–Tsallis entropy for a probability distribution is proposed and discussed entropy; hamming distance; modified VIKOR-TODIM their particular cases. Some properties of this measure has been model proved mathematically. A comparison of proposed entropy measure with several existing entropy measures is shown to demonstrate the MSC consistency of the proposed approach. In this paper, we proposed 94A15; 94A17; 26D15; 68R10; an integrated approach based on VIKOR and TODIM and entropy- 68R05 weighted method for solving multi-criteria decision-making prob- lems under intuitionistic fuzzy environment and the criteria weights are partial known. Firstly, the problem with multi-criteria decision- making is designed and the steps, principles of the proposed VIKOR- TODIM method are presented. Finally, to verify the applicability and feasibility of the proposed approach, a problem of management information system in a teaching hospital is presented and compared the results with other existing methods. The present work demon- strated, that the proposed VIKOR-TODIM is effective to select and evaluate management information system in any organisation. 1. Introduction Shannon [1] introduced a measure of uncertainty in information theory and proved vari- ous properties for continuous and discrete probability distributions. Zadeh [2] pioneered the theory of Fuzzy sets (FSs) to extend the probability theory, depicted by crisp member- ship degrees whose range lies between 0 and 1. The theory of FSs play a significant role in decision-making under uncertainty. After this, Zadeh [3] suggested the concept of fuzzy entropy in 1968. Following Zadeh’s work, DeLuca and Termini [4] proposed a probabilistic entropy measure for FSs. They likewise advanced some aphoristic properties for the fuzzy entropy measure, according to which fuzzy entropy can be defined. Yager [5] proposed an entropy measure for FSs based on the distance between an FS and its complementation CONTACT Vikas Arya vikasarya1988@gmail.com © 2021 The Author(s). Published by Taylor & Francis Group on behalf of the Fuzzy Information and Engineering Branch of the Operations Research Society of China & Operations Research Society of Guangdong Province. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/ licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. 358 V. ARYA AND S. KUMAR while Higashi and Klir [6] stretched out Yager’s concept to a more general kind of fuzzy complementation. In fuzzy set, negative membership value of a member is complement of its membership degree from one, but practically it is not true. The concept of FSs has undergone further extensions and generalisations [7,8]. Motivated by FSs, Atanassov [9] proposed the notion of intuitionistic fuzzy sets (IFSs) by adding one more element in it and is termed as ‘hesitancy degree’. The prominent concept of intuitionistic fuzzy (IF) set theory is that it assigns to every element a membership function (ξ ∈ [0, 1]) and a non- membership function (ν ∈ [0, 1]) satisfies a condition that the sum of membership (ξ ) and non-membership (ν) functions is less than or equal to 1. Due to its flexibility and ability to tackle real-life problems, IFSs have been widely used in many disciplines like as image pro- cessing, robotics, algebraic structures, pattern recognition, agriculture areas, irrigation and engineering fields, etc. Therefore, the proposed article focuses under intuitionistic fuzzy theory. Kaufmann [10] introduced an entropy measure for measuring the degree of a FS by a metric distance between its membership and non-membership function of its closest crisp set. Firstly, Burillo and Bustince [11] introduced the entropy on IFSs can be utilised to assess intuitionism of an IFS. De Luca and Termini’s axioms reformulated by Szmidt and Kacprzyk [12] in 2001 and characterised an entropy measure for IFSs with an alternative approach dependent on geometric translation of IFS and the closest crisp set. Hung and Yang [13] suggested their axiomatic characterisations of entropy on IFSs by utilising the theory of probability. Verma and Sharma [14] presented an exponential entropy measure for IFSs by extending the exponential fuzzy entropy as defined by Pal and Pal [15]. Ye [16] defined two entropy measures for IFSs using triangular function. Ali et al. [17] proposed a graphical method by employing uncertainty index and IF entropy. Garg et al. [18] proposed a gener- alised IF entropy measure of order α and degree β using logarithmic function. Therefore, it has been observed that the entropy measures are playing a dominate role in the decision- making problems. But it has been concluded from the above mentioned studies that all IF entropy measures only consider the derivation of membership and non-membership degree, not consider the effect of hesitancy degree of the IF set. In fact, the degree of uncertainty of an IFS involves both fuzziness and intuitionism. The fuzziness is associated by deviation of membership degree and non-membership degree, while the intuitionism is dominated with hesitancy degree. To enrich the abovementioned shortcomings, some new entropy measures for IFSs are proposed, such as Garg and Kaur [19] proposed an IF entropy measure with two parameters and Joshi and Kumar [20] proposed a parametric entropy under IF environment and applied in multiple attribute decision-making. There- fore, encouraged from the characteristic of fuzziness and intuitionism of IFS, the present paper presents a new parametric intuitionistic fuzzy entropy measure to measure the fuzzi- ness of a set. Recently, many authors have explored the concept of intuitionistic fuzzy entropies from their points of view [17–21] and applied them in different areas like sup- plier selection, real life situations, medical diagnosis, education, image processing pattern recognition, failure mode. The multi-criteria decision-making (MCDM) method is an important branch of decision- making. A problem that contains a set of fixed criteria and a set of alternatives is named as multi-criterion decision-making problem. Decision-making theory is characterised as a process of selecting the best alternatives among finite feasible alternatives according to the preference given by various decision-experts. Its theory is a very important branch, which is FUZZY INFORMATION AND ENGINEERING 359 mostly used in human activities. Because decision-making problem is frequently produced from a complicated environment, evaluated information is usually fuzzy. Under different environments such as IFS [9], T-spherical FSs [22], interval valued T-spherical FSs [23], q- rung orthopair FSs [24,25], various researchers presented various kinds of algorithms based on aggregation operator and entropy measures to solve the decision-making problems [18,21,23,26–28]. TODIM (an acronym in Portuguese for Interactive and Multi-Criteria Decision Mak- ing) is the most implemented technique for decision-making problems which was introduced by Hwang and Yoon [29] especially in economics, medical sciences, social sciences, engineering etc. Many scholars have contributed in fuzzy TODIM models [30,31], intuitionistic fuzzy TODIM models [32,33], the multi-hesitant fuzzy linguistic TODIM approach [34,35] and the intuitionistic linguistic TODIM method [34]. In a MCDM prob- lem, we want to select the foremost reasonable alternative from a finite set of alternatives and the goal is to find the best possible alternative satisfying a specific set of criteria. Fre- quently, the criterion is thus conflicting and equivalent that it turns out to be very difficult to take a final decision,for example purchasing a cellphone or buying a car etc. are some familiar real world applications of MCDM problems. However, in the solution of MCDM problems criteria weights play a vital role. It is desirable to find an alternative perform- ing closest to an ideal alternative. VIKOR (VlseKriterijumska Optimizacija I Kompromisno Resenje) method initially proposed by Opricovic [36] has been widely used in solving MCDM problems with incommensurable and different criteria. This is because of the way that the VIKOR method provides a compromise solution(s). You and Liu [37] presented an extended VIKOR approach under the IFS environment. Garg et al. [24] presented a VIKOR method using power aggregation operators and applied to solve the decision-making problems. However, TOPSIS (technique for order preference with respect to the similarity to the ideal solution) method, introduced by Hwang and Yoon [29] is a well-known MCDM method. The aim of this method is to choose the best alternative whose distance from its positive ideal solution is shortest. After their invention/existence, numerous attempts have been done by the authors to apply the TOPSIS method under the fuzzy and IFS environment. For instance, Boran et al. [38] presented the TOPSIS method to solve the problem of Personnel selec- tion. Kumar and Garg [39,40] worked on connection number of the set pair analysis (SPA) theory and presented TOPSIS method to solving the expert problems. Arya and Kumar [7] developed TOPSIS method to solve decision problems under bipolar fuzzy environment. Recently, Garg and Kaur [27] developed the distance and weighted distance measures to solve the problems under cubic intuitionistic fuzzy set environment. Apart from this, some authors have proposed many theories and tools to solve MCDM problems, for example, the ELECTRE method [41], PROMETHEE [42], WASPAS method [43], Choquet integral [44] etc., each having its own benefits and downsides. The objectives of this paper are outlined as follows: (1) Intuitionistic fuzzy numbers (IFNs) demonstrate more and more execution in providing reliable, vague and inexact assessment information. (2) Taking into account that entropy is an important measure and used for calculating weights for solving decision-making issues, this article, in general, aims to define new entropy measure based on Havdra–Charvat–Tsallis entropy in the IFS context to handle complicated problems of choice. 360 V. ARYA AND S. KUMAR (3) Weights calculated by the entropy method are extraordinary and these objective weights make the decision results more exact and precise when applied to real-life MCDM based on the IFS environment. (4) The judgement of a perfect alternative in an IFS environment is a laborious MCDM prob- lem. In existing techniques, optimal alternative is characterised by single model which promote to do work in IF environment with combined TODIM and VIKOR method. (5) The proposed VIKOR-TODIM model overcome the restrictions of existing models. The main contributions of this article are: (1) The idea of entropy measures is extended to IFNs and properties of this measure are discussed. (2) As the random weights may lead to wrong interpretation to the final ordering of the alternatives, thus computation of weight vector objectively is more important than subjectively. Hence, the presented approach is more suitable than existing approaches. (3) A MCDM problem is discussed using proposed entropy measure and an algorithm is developed with intuitionistic VIKOR-TODIM to tackle complex realistic problems with IF data. (4) Finally, the proposed IF information measure has been studied in MCDM problems for evaluating university teaching hospital to make a newly information managing sys- tem and discuss a comparative study to show the effectiveness and rationality of the proposed approach with other existing decision approaches. The remainder of this communication is classified as follows. In Section 2, we quickly audit some essential concepts about FSs and IFSs. In Section 3, we propose a new entropy measure of IFSs and discuss their properties with the proof of its validity. Section 4 pro- vides an example to examine and conform to the performance of the proposed entropy measure and compare it with the existing entropy measures to demonstrate its validity. In Section 5, we propose a new MCDM method based on the proposed entropy measure and the proposed modified VIKOR-TODIM method. Section 6 describes the analysis of the result associated with modified VIKOR-TODIM approach. A sensitivity analysis and comparative analysis has been described in Sections 6.2 and 6.3, respectively. Finally, the concluding remarks are presented in Section 7. 2. Preliminaries This section recalls the vital concept and characterisations of FSs and IFSs. Definition 2.1: An FS [2] M in a finite universe of discourse K is defined as M ={(l , ξ (l )) : i M i ˜ ˜ l ∈ K} where ξ : K → [0, 1] is membership function. The number ξ (l ) denotes the i M M i degree of membership of l ∈ K to M. Definition 2.2: An IFS [9] M in a finite universe of discourse K is defined as M = ˜ ˜ {(l , I(ξ (l ), ν (l )))l ∈ K} where ξ , ν : K → [0, 1] shows the membership and non- i M i M i i M M membership functions. The numbers ξ (l ), ν (l ) denote the degrees of membership M i M i FUZZY INFORMATION AND ENGINEERING 361 and non-membership of l ∈ K to M, correspondingly, under the requirement 0 ≤ ξ (l ) + i M i ν (l ) ≤ 1, for all l ∈ K. M i i For each l in K, the degree of hesitancy or intuitionistic index is given by φ (l ) = i M i 1 − ξ (l ) − ν (l ). Obviously, when φ (l ) = 0, that is ν (l ) = 1 − ξ (l ) for all l ∈ K,IFS M i M i M i M i M i i M alters into an ordinary FS. For simplicity, IFN is denoted by σ = (ξ , ν ) which fulfills ξ , ν ∈ [0, 1] and ξ + ν ≤ 1. σ σ σ σ σ σ 2.1. Geometrical Representiation of IFSs A convenient geometrical representation of IFS is illustrated in Figure 1 with triangle POQ. In an IFS M, each member l has a membership degree ξ (l ) , negative membership degree i M i ν (l ) and a hesitation degree φ (l ) with the condition ξ (l ) + ν (l ) + φ (l ) = 1and M i M i M i M i M i ξ (l ), ν (l ), φ (l ) ∈ [0, 1]. In triangle POQ, P(1, 0, 0) and Q(0, 1, 0) points are known as crisp M i M i M i elements, for example, any point M(1, 0, 0) on membership axis denotes fully belonging- ness / membership degrees with ξ = 1and ν, φ = 0. Analogously, point O(0,0,1) denotes hesitant part while the segment PQ considers the elements belong to the FS as ξ + ν = 1 and φ = 0. For example, consider an IFS M ={(l , 0.3, 0), (l , 0.4, 0.5)} in Figure 1 in which 1 2 the point (l , 0.4, 0) represents an element with hesitation parameter 0.7 while, the point (l , 0.5, 0.3) represents a degree of hesitation is equal to 0.1. Definition 2.3 ([45]): Let σ = I(ξ , ν ), σ = I(ξ , ν ) and σ = I(ξ , ν ) be three IFNs. σ σ 1 σ σ 2 σ σ 1 1 3 3 Then the following laws are satisfied with IFNs: (i) σ ⊕ σ = (ξ + ξ − ξ ξ , ν ν ); 1 2 σ σ σ σ σ σ 1 2 1 2 1 2 (ii) σ ⊗ σ = (ξ ξ , ν + ν − ξ ν , ); 1 2 σ σ σ σ σ σ 1 2 1 2 1 2 λ λ λ (iii) σ = (ξ ,1 − (1 − ν ) )for λ > 0. λ λ (iv) λ σ = (1 − (1 − ξ ) , ν )for λ > 0, Figure 1. Geometrical representation of IFS. 362 V. ARYA AND S. KUMAR In order to rank the IFNs, Chen and Tan [46] and Hong and Choi [47] further gave a score and an accuracy function to compare the numeric size of the IFNs : Definition 2.4: Consider an IFN g = I(ξ , ν ). Then H(g) and s(g) are called the accuracy g g and score function of f are defined as: H(g) = ξ + ν , s(g) = ξ − ν,(1) g g g g Based on score functions of IFNs, the ranking laws of IFNs can be illustrated as below: Definition 2.5: Let g = I(ξ , ν ) and g = I(ξ , ν ) be two IFNs. s (g )(i = 1, 2) be the 1 g g 2 g g i 1 1 2 2 score function values of g and g , respectively and H(g )(i = 1, 2) be the accuracy degree 1 2 i of g and g , then: 1 2 • for s(g )< s(g ),wehave g < g ; 1 2 1 2 • for s(g ) = s(g ),wehave 1 2 (i) if H(g )< H(g ), then g < g . 1 2 1 2 (ii) if H(g ) = H(g ), then g = g . 1 2 1 2 Definition 2.6: Suppose g = I(ξ , ν ) and g = I(ξ , ν ) be two IFNs. Then the Ham- 1 g g 2 g g 1 1 2 2 ming distance measures of f and f proposed by Szmidt et al. [48] is computed as follows: 1 2 d (g , g ) = (|ξ − ξ |) + (|ν − ν |) + (|φ − φ |) (2) h 1 2 g g g g g g 1 2 1 2 1 2 To illustrate the Hamming distance measure ,we can use an example as follows: Take any two IFNs like as: g = (0.6, 0.2) and g = (0.4, 0.5), then the hamming degrees of f and f 1 2 1 2 are φ (l) = 1 − 0.6 − 0.2 = 0.2, φ (l) = 1 − 0.4 − 0.5 = 0.1 g g 1 2 then the Hamming distance between g and g is : 1 2 d (g , g ) = |0.6 − 0.4| + |0.2 − 0.5| + |0.2 − 0.1| = 0.3 H 1 2 Definition 2.7 ([9]): For any two IFSs M and N defined in the universe set K given by ˜ ˜ M ={l , ξ (l ), ν (l )|l ∈ K} and N ={l , ξ (l ), ν (l )|l ∈ K};(3) i M i M i i i N i N i i then properties and operations on IFSs are briefly defined as follows: (1) M ⊆ N ⇔∀ l ∈ K, ξ (l ) ≤ ξ (l ), ν (l ) ≥ ν (l ) for ξ (l ) ≤ ν (l ) OR ξ (l ) ≥ ξ (l ), i M i N i M i N i N i N i M i N i ν (l ) ≤ ν (l ) for ξ (l ) ≥ ν (l ) ; M i N i N i N i 1. M = N ⇔∀ l ∈ K, M ⊆ N and N ⊆ M; (2) co M = M ={l , ν (l ), ξ (l )|l ∈ K}; i M i M i i (3) M ∩ N ={ξ (l ) ∧ ξ (l ) and ν (l ) ∨ ξ (l )|l ∈ K}; M i N i M i N i i (4) M ∪ N ={ξ (l ) ∨ ξ (l ) and ν (l ) ∧ ξ (l )|l ∈ K}. M i N i M i N i i FUZZY INFORMATION AND ENGINEERING 363 Definition 2.8 ([12]): A mapping ϒ defined on IFS(K) is an entropy if it holds the following four requirements: (IFP1). ϒ (M) = 0 ⇔ϒ is a crisp set, i.e. for all l ∈ K, ξ (l ) = 0, ν (l ) = 1; or ξ (l ) = e e i M i M i M i 1, ν (l ) = 0. M i (IFP2). ϒ (M) = 1, that is, attains maximum value ⇔ξ (l ) = ν (l ) = φ (l ) = , for all e M i M i M i l ∈ K. 1. (IFP3).] ϒ (M) ≤ ϒ (N) ⇔M ⊆ N ,i.e. ξ ≤ ξ and ν ≤ ν for max (ξ , ν ) ≤ and e e M N M N N N ξ ≥ ξ and ν ≥ ν for min (ξ , ν ) ≥ . M N M N N M (IFP4). ϒ (M) = ϒ (M ) for all M ∈ IFS(K). e e ˜ ˜ Throughout this paper, FS(K) and IFS(K) will, respectively, denote the set of all fuzzy set and the set of all intuitionistic fuzzy set on set K. It is realised that Renyi [49], Tsallis [50], Pal and Pal [15], Boekee and Vander Lubbe [51], Taneja [52] measures have been summed up based on the Shannon entropy [1]. Now we can construct a new parametric information measure based on the Havrda–Charvat–Tsallis entropy in the following section. 3. A New Parametric Measure for IFSs 3.1. Background For a finite set of probability distribution defined by ={K = (l , l , ... , l ) : l ≥ r 1 2 r i 0, l = 1}, r ≥ 2 , the entropy given by i=1 −1 V (K) = l − l (4) i i −1 ( − ) i=1 for some K ∈ and > 0(= 1). Particular Cases: −1 ˜ ˜ (1) V (K) = V (K). (2) If → 1,(4) recovers the entropy studied by Shannon [1], i.e. lim V (K) = →1 − (l )log(l ). i i i=1 −1 1 1 1− 1− (3) If l = l = ··· = l = , V (K) = [r − r ]. 1 2 r −1 r − (4) If = 2, (4) becomes =2 2 V (L) = ( l − l ) , i=1 which alters into an entropy for probability distribution K = (l , l , ... , l ). 1 2 r −1 (5) Setting = σ , (4) becomes Sharma and Taneja entropy [52]. −1 1 (, =σ) σ i.e. V (K) = l − l . − σ i=1 364 V. ARYA AND S. KUMAR −1 1 −1 ˜ ˜ ˜ (6) V (K) = [exp((1 − )V (K)) − exp((1 − )V (K))], represents a close −1 Renyi Renyi −1 relation between proposed and Renyi [49] entropy, where V (K) = log −1 Renyi i=1 1− −1 l . −1 −1 ˜ ˜ ˜ (7) V (K) = [1 − ( − 1)V (K) − [1 − ( − 1)V (K)]], denotes a relation −1 − Tsallis Tsallis −1 −1 1 r between Tsallis [50] and proposed entropy, where V (K) = [1 − l ]. −1 i=1 i Tsallis −1 −1 −1 −1 1 1− 1− ˜ ˜ (8) V (K) = [1 − (2 − 1)V ] − [1 − (2 − 1)V ], where V (K) = −1 − H−Ch H−Ch H−Ch −1 [1 − l ]. It shows a relation between Havrda and Charvat [53], Daro- −1 i=1 i 1− 2 −1 cozy [54] and proposed entropy. −1 −1 −1 ˜ ˜ ˜ (9) V (K) = [(1 − V (K)) − (1 − V (K)) ], where V −1 −1 −1 -norm -norm -norm (−1) − ( −1) −1 −1 −1 r r ˜ ˜ (K) = [1 − ( l ) ]and H (K) = [1 − ( l ) ] are entropy -norm −1 i=1 i −1 i=1 i −1 measures studied by Arimoto [55] and Boekee–Vander Lubbe [51]. Hence, the pro- posed entropy has a close relationship with Arimoto [55] and Boekee–Vander Lubbe [51] entropy. ˜ ˜ Theorem 3.1: The parametric entropy V (K), K ∈ , fulfills the following properties: (1) Non-negative: V (K) ≥ 0 for all > 0(= 1). (2) Expansible: V (l , l , ... , l ,0) = V (l , l , ... , l ). 1 2 r 1 2 r (3) Symmetry: V (l , l , ... , l ) is a symmetric function of (l , l , ... , l ). 1 2 r 1 2 r (4) Decisive: V (0, 1) = 0 = V (1, 0). (5) Maximility: −1 1 1 1 1 1− 1− V (l , l , ... , l ) ≤ V ( , , ... , ) = [r − r ]. 1 2 r −1 r r r (6) Concavity: ˜ ˜ ˜ ˜ V (tK + (1 − t)K ) ≥ V (K ) + (1 − t)V (K ). 1 2 1 2 (7) Continuity: V (l , l , ... , l ) is continuous, where l ≥ 0 for all i = 1, 2, ... , r and > 0. 1 2 r i Proof: Proof of the above theorem is trivial and easy to calculation. DeLuca and Termini [4] suggested fuzzy entropy for a fuzzy set M analogously to Shannon entropy as L (M) =− [ξ (l ) log (ξ (l )) + (1 − ξ (l )) log (1 − ξ (l ))], (5) DLT M i 2 M i M i 2 M i i=1 ˜ ˜ where M ∈ FS(K) and l ∈ K. Analogous to Equation (5), Bhandari and Pal [56] extended the idea of Renyi [49] entropy from probabilistic setting to fuzzy to develop a new entropy FUZZY INFORMATION AND ENGINEERING 365 information measure given by L (M) = log [ξ (l ) + (1 − ξ (l )) ]. (6) BP 2 M i M i r(1 − ) i=1 The well known idea of Bhandari and Pal [56] was further generalised to IFSs by Hung and Yang [13] to introduce a new intuitionistic fuzzy entropy given by L (M) = log [ξ (l ) + ν (l ) + φ (l ) ]; where ∈ (0, 1).(7) H&Y 2 M i M i M i r(1 − ) i=1 Keeping these concepts in mind, we present a new information measure for IFSs by extending the idea of Hung and Yang [13] in the next subsection. 3.2. Definition For any M ∈ IFSs, we define 1 −1 −1 −1 V (M) = ξ (l ) + ν (l ) + φ (l ) M i M i M i −1 r( − ) i=1 − ξ (l ) + ν (l ) + φ (l );(8) M i M i M i where > 0(= 1). Limiting Cases: (1) If = 1, then Equation (8) alters an intuitionistic fuzzy entropy, which was studied by Hung and Yang [13]: H =− (ξ (l ) log (ξ (l )) + ν (l ) log (ξ (l )) + φ (l ) log (φ (l )) (9) HY M i M i M i M i M i M i 2 2 2 i=1 (2) If φ (l ) = 0, then Equation (8) becomes the fuzzy entropy studied by Arya and Kumar M i [7]: 1 −1 −1 i.e. V (M) = ξ (l ) + ν (l ) + − ξ (l ) + ν (l ) M i M i M i M i −1 r( − ) i=1 where > 0(= 1) Keeping above said in mind, in the next subsection, we will prove the existence of proposed entropy measure by satisfying the requirements in Definition 2.8. 3.3. Justification Before proving the efficiency of the proposed entropy measure, we will investigate a property needed in the proof of justification. 366 V. ARYA AND S. KUMAR Property 1: Under the condition IFP4, we have 1 1 1 ξ (l ) − + ν (l ) − + φ (l ) − M i M i M i 3 3 3 1 1 1 ≥ ξ (l ) − + ν (l ) − + φ (l ) − (10) N i N i N i 3 3 3 and 2 2 2 1 1 1 ξ (l ) − + ν (l ) − + φ (l ) − M i M i M i 3 3 3 2 2 2 1 1 1 ≥ ξ (l ) − + ν (l ) − + φ (l ) − (11) N i N i N i 3 3 3 Proof: If ξ (l ) ≤ ξ (l ) and ν (l ) ≤ ν (l ) with max{ξ (l ), ξ (l )}≤ , then ξ (l ) ≤ M i N i M i N i N i N i M i 1 1 1 ξ (l ) ≤ ; ν (l ) ≤ ν (l ) ≤ and φ (l ) ≥ φ (l ) ≥ which shows that (10) and (11) hold. N i M i N i M i N i 3 3 3 Similarly, if ξ (l ) ≥ ξ (l ) and ν (l ) ≥ ν (l ) with min{ξ (l ), ν (l )}≥ , then (10) and (11) M i N i M i N i N i N i hold. Theorem 3.2: Measure V (M) is a valid entropy measure for IFSs. Proof: IFP1: Let M is a crisp set having either ξ (l ) = 1, and ν (l ) = φ (l ) = 0or ν (l ) = M i M i M i M i −1 −1 1, and ξ (l ) = φ (l ) = 0or φ (l ) = 1and ξ (l ) = ν (l ) = 0. ⇒ (ξ (l ) + ν (l ) + M i M i M i M i M i M i M i −1 φ (l ) ) − (ξ (l ) + ν (l ) + φ (l ) ) = 0. Since > 0( = 1), V (M) = 0. M i M i M i M i Conversely, if V (M) = 0, we have −1 −1 −1 (ξ (l ) + ν (l ) + φ (l ) ) − (ξ (l ) + ν (l ) + φ (l ) ) = 0. M i M i M i M i M i M i Since > 0( = 1), this is possible only in the following cases: either ξ (l ) = 1, and ν (l ) = φ (l ) = 0, M i M i M i or ν (l ) = 1, and ξ (l ) = φ (l ) = 0, M i M i M i or φ (l ) = 1, and ξ (l ) = ν (l ) = 0. M i M i M i From the above cases, V (M) = 0if M is crisp . This proves (IFP1). IFP2: Since ξ (l ) + ν (l ) + φ (l ) = 1, to obtain the maximim value of IF-entropy M i M i M i V (M),wewrite g(ξ , ν , φ ) = ξ (l ) + ν (l ) + φ (l ) − 1 and taking λ as the Lagrange’s M N M M i M i M i multipliers, we construct the Lagrange’s function as: G(ξ , ν , φ ) = V (ξ , ν , φ ) + λg(ξ , ν , φ ). (12) M N M M N M M N M To obtain the maximality of, V (M), differentiating Equation (12) partially with respect to ξ , ν , φ and λ and put them equal to zero,we obtain ξ (l ) = ν (l ) = φ (l ) = .Itcan M N M M i M i M i be easy to calculate that all the first-order derivative becomes zero if and only if ξ (l ) = M i 1 1 ν (l ) = φ (l ) = . The stationary point of V (M) is ξ (l ) = ν (l ) = φ (l ) = .Now,we M i M i M i M i M i 3 3 prove V (M) is a concave function of M ∈ IFS(K). For this, we compute Hessian of V (M) at the stationary point. FUZZY INFORMATION AND ENGINEERING 367 Definition 3.1 (Hessian): The Hessian matrix of a function (ξ , ξ , ξ ) of three variables 1 2 3 is given by ⎡ ⎤ 2 2 2 ∂ ∂ ∂ ⎢ ⎥ ∂ξ ∂ξ ∂ξ ∂ξ ∂ξ 2 1 3 1 ⎢ ⎥ ⎢ ⎥ 2 2 2 ∂ ∂ ∂ ⎢ ⎥ HEN( ) = , (13) ⎢ ⎥ ⎢ ∂ξ ∂ξ ∂ξ ∂ξ ⎥ ∂ξ 1 2 3 2 ⎢ ⎥ 2 2 2 ⎣ ⎦ ∂ ∂ ∂ ∂ξ ∂ξ ∂ξ ∂ξ 1 3 2 3 ∂ξ The function is called strictly convex at a point in its domain if HEN( ) is positive defi- nite and strictly concave if HEN( ) is negative definite and. The Hessian of V (M) is given by ⎡ ⎤ −10 0 ⎣ ⎦ HEN(V (M)) = 0 −10 (14) −1 r( − ) 00 −1 (2−) −1 −1 which is negative definite for all > 0(= 1), where p = ( − 1)3 − ( − −1 (2− ) 1)3 .Thus, V (M) is strictly concave function for all > 0(= 1) having its maximum value at stationary point ξ (l ) = ν (l ) = φ (l ) = . M i M i M i IFP3: Since, V (M) be concave function of M ∈ IFS(K), with maximum value at stationary point,then if max{ξ (l ), ν (l )}≤ ,then ξ (l ) ≤ ξ (l ) and ν (l ) ≤ ν (l ) implies φ (l ) ≥ M i M i M i M i M i M i M i φ (l ) ≥ . Therefore, using property (1), we see that V (M) agrees with condition IFP3. N i Similarly, if min{ξ (l ), ν (l )}≥ , then ξ (l ) ≤ ξ (l ) and ν (l ) ≥ ν (l ). Therefore, M i M i M i N i M i N i using property (1), we see that V (M) agrees with condition IFP3. IFP4: For an IFS, V (M) = V (M ), which is clear from the definition. ˜ ˜ Theorem 3.3: For any two M, N ∈ IFS(K), such that for all l ∈ K either M ⊆ Nor N ⊆ M; then, V (M ∪ N) + V (M ∩ N) = V (M) + V (N). (15) ˜ ˜ ˜ Proof: To prove Theorem 3.3, we divide universal set K into two sets say K and K such 1 2 that ˜ ˜ ˜ ˜ K ={l ∈ K : M ⊆ N},and K ={l ∈ K : M ⊇ N} (16) 1 i 2 i ξ (l ) ≤ ξ (l ), ν (l ) ≥ ν (l ) ∀ l ∈ K (17) M i N i M i N i i 1 ξ (l ) ≥ ν (l ), ξ (l ) ≤ ν (l ) ∀ l ∈ K (18) M i N i M i N i i 2 1 −1 −1 −1 Now, V (M ∪ N) = {(ξ (l ) + ν (l ) + φ (l ) ) (M∪N) i (M∪N) i (M∪N) i −1 r( − ) i=1 − ξ (l ) + ν (l ) + φ (l ) } (19) (M∪N) i (M∪N) i (M∪N) i 1 −1 −1 −1 = ξ (l ) + ν (l ) + φ (l ) N i N i N i −1 r( − ) 1 368 V. ARYA AND S. KUMAR − ξ (l ) + ν (l ) + φ (l ) N i N i N i 1 −1 −1 −1 + ξ (l ) + ν (l ) + φ (l ) M i M i M i −1 r( − ) − ξ (l ) + ν (l ) + φ (l ) (20) M i M i M i Similarly, we get 1 −1 −1 −1 V (M ∩ N) = ξ (l ) + ν (l ) + φ (l ) N i N i N i −1 r( − ) − ξ (l ) + ν (l ) + φ (l ) N i N i N i 1 −1 −1 −1 + ξ (l ) + ν (l ) + φ (l ) M i M i M i −1 r( − ) − ξ (l ) + ν (l ) + φ (l ) (21) M i M i M i Now, adding Equations (20) and (21), we have V (M ∪ N) + V (M ∩ N) = V (M) + V (N) (22) Corollary 3.1: For any intuitionistic fuzzy set M and M (known as complement of M), we have c c c V (M ∪ M ) + V (M ∩ M ) = V (M) + V (M ). (23) 4. Numerical Examples In this section, we demonstrate the performance of the proposed measure V (M) will be validated based on the following examples. To illustrate the effectiveness and perfor- mance of the proposed measure for IFSs,some existing fuzzy measures will be adopted for comparison. Therefore, firstly we recall some widely used fuzzy measures for IFSs. The entropy measure proposed by Zeng and Li [57] is appeared as below: V (M) = 1 − |ξ (l ) − ν (l )| ZL M i M i i=1 The entropy measure proposed by Burillo and Bustince [11] is appeared as below: V (M) = |1 − ξ (l ) − ν (l )| BB M i M i i=1 The entropy measure proposed by Szmidt et al. [48] is appeared as below: 1 min(ξ (l ), ν (l )) + φ (l ) M i M i M i V (M) = SK r max(ξ (l ), ν (l )) + φ (l ) M i M i M i i=1 FUZZY INFORMATION AND ENGINEERING 369 The entropy measure proposed by Hung and Yang [13] is appeared as below: 2 2 2 V (M) = [1 − (ξ (l )) − (ν (l )) + (φ (l )) ] HC M i M i M i i=1 The entropy measure proposed by Zhang and Jiang’s [58] is appeared as below: 1 min(ξ (l ), ν (l )) M i M i V (M) = ZJ r max(ξ (l ), ν (l )) M i M i i=1 The entropy measure proposed by Garg and Kaur [19] is appeared as below: ⎡ ⎤ 2−β 2−β 1− 2−β log ξ (l ) + ν (l ) (ξ (l ) + ν (l )) 2 − β 2 i i M i M i ⎢ ⎥ ,β M M V (M) = × ⎣ ⎦ GK i=1 r(2 − − β) 1− 2−β +2 (1 − ξ (l ), ν (l )) M i M i where > 0, β ∈ [0, 1], + β = 2. The entropy measure proposed by Garg [21] is appeared as below: 1 1 S R R × S R,S S S S R R R V (M) = ξ (l ) + ν (l ) + φ (l ) − ξ (l ) + ν (l ) + φ (l ) i i i i i i G M M M M M M r(R − S) i=1 either R > 1, 0 < S <1or0 < R < 1, S > 1. Example 4.1: Let us consider a IFS M in K ={3, 4, 5, 6, 7} is defined as M ={(3, 0.1, 0.8), (4, 0.3, 0.5), (5, 0.5, 0.4), (6, 0.9, 0.0), (7, 1.0, 0.0)}. Then the modifier for the IFS M ={(l, ξ (l), ν (l))|l ∈ K} M M in K is given by r r r M ={(l, (ξ (l)) ), (1 − ν (l)) |l ∈ K}. (24) M M Based on the operations De et al. [59] in Equation (24), we have: M ={(3, 0.316, 0.552), (4, 0.547, 0.292), (5, 0.707, 0.225), (6, 0.948, 0.0), (7, 1.0, 0.0)}, M ={(3, 0.01, 0.960), (4, 0.090, 0.750), (5, 0.250, 0.640), (6, 0.810, 0.0), (7, 1.0, 0.0)}, M ={(3, 0.001, 0.992), (4, 0.027, 0.875), (5, 0.125, 0.784), (6, 0.729, 0.0), (7, 1.0, 0.0)}, M ={(3, 0001, 0.998), (4, 0.008, 0.938), (5, 0.063, 0.870), (6, 0.659, 0.0), (7, 1.0, 0.0)}. By considering the classification of linguistics variables which are shown in Table 1,we justify the performance of proposed measure. 370 V. ARYA AND S. KUMAR Table 1. Classification of linguistic variables. IFSs Linguistic variables M Medium Large (ML) M Large (L) M Very Large (VL) M Absolutely Large (AL) M Very Very Large (VVL) Table 2. IFVs with different entropy measures. 2 2 3 4 Entropies M M M M M 1 1 1 1 V 0.4291 0.4400 0.2160 0.1364 0.1082 ZL V 0.0683 0.0800 0.0760 0.0752 0.0801 BB V 0.3571 0.4074 0.1677 0.1102 0.0.949 SK V 0.3354 0.3281 0.2890 0.2603 0.2397 HY V 0.2486 0.2261 0.2117 0.1949 0.1457 ZJ ,β V 0.5254 0.5168 0.4763 0.4276 0.3146 GK R,S V 0.4789 0.4354 0.3654 0.3350 0.2711 =2 V 0.5160 0.5110 0.4122 0.3247 0.2801 2 4 Intuitively, from M to M , the losss of information hidden in them become less. The 1 1 entropy conveyed by them increasing. So the following relation holds for good perfor- mance: 2 2 3 4 V(M )> V(M )> V(M )> V(M )> V(M ). (25) 1 1 1 1 To make a comparison, eight entropy measures V (M ), V (M ), V (M ), V (M ), ZL 1 BB 1 SK 1 HY 1 ,β R,S (=2) V (M ), V (M ),V and V are employed to faciliate analysis. In Table 2, we present ZJ 1 1 GK G the results obtained based on different measures to facilitate comparative analysis. We can note that IFS M will be assigned more entropy than the IFS M when entropy measures V (M ),V (M ),and V (M ) are applied.The ranking orders obtained based on ZL 1 BB 1 SK 1 these measures are listed below. 2 3 4 V (M )> V (M )> V (M )> V (M )< V (M ), ZL 1 ZL ZL ZL ZL 1 1 1 2 4 3 V (M )> V (M )> V (M )> V (M )< V (M ), BB 1 BB BB BB BB 1 1 1 2 2 3 4 V (M )> V (M )> V (M ) > H (M )> V (M ). SK 1 SK SK 1 SK SK 1 1 1 It is shown that these ranked orders do not satisfy intuitive analysis in Equation (25), while ,β R,S other entropy measures can induce desirable results. In this example, V (M ), V , V HY 1 GK G (=2) and V (M ) perform well. This illustrates that these entropy measures are not robust enough to distinguish the uncertainty of IFSs with linguistic information. Example 4.2: Let us consider another IFS M defines in K. The IFS is defined as : M ={(3, 0.2, 0.5), (4, 0.3, 0.7), (5, 0.4, 0.3), (6, 0.7, 0.2), (7, 0.8.0.1)}. 2 FUZZY INFORMATION AND ENGINEERING 371 ,β R,S 2 2 3 4 We calculate M ,M ,M and M . Now we compare only V (M ),V (M ), V , V and HY 2 ZJ 2 2 2 2 2 GK G V (M ). (=2) 2 ,β Moreover, the results produced by entropy measure V (M ) and V are not reasonable ZJ 2 GK in this example, which are shown as the equations below. 2 3 4 V (M )< V (M )> V (M )> V (M )> V (M ), ZJ 2 ZJ ZJ ZJ ZJ 2 2 2 ,β ,β ,β ,β 2 2 4 3 V (M )> V (M )> V (M )> V (M )> V (M ). 2 HY GK GK 2 GK 2 GK 2 2 ,β Therefore, the entropy measures V (M ),V (M ) are not suitable for differentiating the ZJ 2 2 GK R,S information conveyed by IFSs. But V (M ), V and V are also satisfy the ranking order HY 2 R,S (=2) in Equation (25). The effectiveness of the proposed fuzzy measure V (M ), V and V (M ) is indicated by this example once again. Hence, the proposed measure consider HY 2 one parameter which increases the flexibility due to the setting whereas V does not HY due to the absence of parameters . Therefore, the proposed measure is more encouraging and has broad scope of practical applications in numerous areas. Moreover, we can see that proposed entropy and Garg and Kaur [19] have the same order preference. It proves that the new entropy formula is feasible, but the presence of parameter in proposed entropy and Garg and Kaur [19] is more convincing. In the next section, the traditional VIKOR and TODIM methods are integrated to solve MCDM problems in intuitionistic fuzzy environments. 5. Intuitionistic Fuzzy VIKOR-TODIM Approach and Its Application 5.1. Vikor Method In 1998, Opricovic [36], the notion of the VIKOR technique was developed to compute a compromise solution(s) based on the L – metric and should be as close to be a positive solution and as far from a negative solution . This method offers a compromise program- ming providing a group utility for the ‘majority’ rule and a less individual regret of the ‘opponent’ with clashing criteria. Assume that each alternative ψ according to each criteria E are given as γ i = 1(1)m; j = i j ij 1(1r). The multi-criteria measure for compromise ranking is created from the L -metric with an aggregating function by Yu [60]. Advancement of the VIKOR method by Yu [60]isgiven by ⎧ ⎫ 1 p p ⎨ ⎬ (H − t ) ij L = u ,1 ≤ p ≤∞, i = 1(1)m; (26) p,i + − ⎩ ⎭ (H − H ) j j j=1 + − where H = max t and H = min t are the best and worst solutions for each criteria. u i ij i ij j j j th ∗ represents the weight of j criteria. The measure L denotes the distance of alternative ψ p,i ∗ ∗ to the ideal solution. Here, L (as S ) and L (as R ), has been used to formulate the grading i i 1,j ∞,i of alternatives ψ . Now, an algorithm is exhibited to solve the MCDM problems by using the IFSs and proposed VIKOR-TODIM model based on distance measures with the following steps: 372 V. ARYA AND S. KUMAR Step 1: Establish the IF decision matrix P = [γ ] with the help of γ = (ξ , ν ) ,where ij m×r ij ij ij γ is an IFN. ij E E ... E 1 2 r ⎡ ⎤ ψ (ξ , ν )(ξ , ν ) ··· (ξ , ν ) 1 σ σ σ σ σ σ 11 11 12 12 1n 1r ⎢ ⎥ ψ (ξ , ν )(ξ , ν ) ··· (ξ , ν ) 2 σ σ σ σ σ σ 21 21 22 22 2n 2r ⎢ ⎥ P = [γ ] = (27) ⎢ ⎥ ij m×r . . . . . . . . . ⎣ ⎦ . . . . ψ (ξ , ν )(ξ , ν ) ··· (ξ , ν ) m σ σ σ σ σ σ m1 m1 m2 m2 mn mr Step 2: Change the decision matrix P = (γ ) into a normalised IF-decision matrix by: ij m×r γ , for benefit type criteria ij q = (28) ij γ , for cost type criteria ij where γ = (ν , ξ ) is the complement of γ [61]. Then, we obtain a new normalised IF- ij ij ij ij decision matrix P = (q ) ij m×r Step 3: Evaluate the Criteria’s weights: If the weights are completely known, at that point MCDM problem can be solved by aggregating all intuitionistic fuzzy information under dif- ferent criterions and comparing the final intuitionistic fuzzy values. However in a partial application, the criteria weights are usually partially known or completely unknown. Conse- quently, the criteria weights must be resolved before solving the MCDM models. The criteria weights can be observationally allocated by decision makers. Nonetheless, this method is subjective and the partial information on the weights vector may not be utilised adequately. Therefore, we can propose another model to compute the weights based on the proposed measure. We can set the total information amount as the objective function of optimisation. By minimising the sum of all information amount, we may construct the following model. m r Min T = V (ξ ) u ij j i=1 j=1 u ∈ S (29) u = 1 such that j j=1 ≥ 0, j = 1, 2, ... , r where S denotes the set of all incomplete information about criterion weights and V (ξ ) ij is the information measure calculated by proposed measure. Determining the criteria weights of each criterion E based on the weights of criteria u = (u , u , ... , u ) as: 1 2 r u = , j, r = 1, 2, ... , r; (30) js where u is the weight of the criterion E , u = max{u /j = 1, 2, ... , r} and 0 ≤ u ≤ 1. Note j j s j js that all criteria have equal importance, any criterion may be chosen. FUZZY INFORMATION AND ENGINEERING 373 5.2. TODIM Method Step 4: The TODIM technique is a discrete multi-criteria technique used for qualitative and quantitative criteria based on Prospect Theory. By Equation (30), we can derive the dominance degree of ψ under each alternative ψ with respect to each criterion E : i i j u d (q , q ) js h ij t j ,if q > q ij t j ⎪ r 1 ⎪ js j=1 null,if q = q Z (ψ , ψ ) = ij t j (31) j i t ⎪ r u d (q , q ) ⎪ js h ij t j j=1 1 1 % ⎩ − if q < q ij t j γ u js where the parameter γ indicates the attenuation factor of the losses, and d (q , q )is to h ij t j measure the distance between the IFNs q and q the parameter γ represents the attenua- ij tj tion factor of losses. By definition if q > q ,then Z (ψ , ψ ) signifies a gain; if q < q , then ij tj j i t ij t j Z (ψ , ψ ) represents a loss. j i t1 Step 5: Constructing the dominance matrix of each alternative ψ with respect to each criterion E is shown below: E E ... E 1 2 r ⎡ ⎤ ψ 0 Z (ψ , ψ ) ··· Z (ψ , ψ ) 1 j 1 2 j 1 m ⎢ ⎥ ψ Z (ψ , ψ ) 0 ··· Z (ψ , ψ ) j 2 1 j 2 m ⎢ ⎥ Z = [Z (ψ , ψ )] = ⎢ ⎥ (32) j j i t m×r . . . . 1 . . . . . . ⎣ ⎦ . . . . ψ Z (ψ , ψ ) Z (ψ , ψ ) ··· 0 j m 1 j m 2 Step 6: Computing the totally dominance degree of each alternative ψ under the criterion E with respect to another alternatives ψ (t = 1, 2, ... , m) as follows: j t 1 (ψ ) = Z (ψ , ψ ) (33) j i j i t t =1 Total dominance results by considering all alternatives are shown as: ⎡ ⎤ ψ Z (ψ , ψ ) j 1 t t =1 1 ⎢ ⎥ ψ Z (ψ , ψ ) 2 j 2 t t =1 1 ⎢ 1 ⎥ [Z (ψ , ψ )] = ⎢ ⎥ j i t m×m . . . . ⎣ ⎦ . . ψ Z (ψ , ψ ) m j m t t =1 1 The dominance matrix t = [t ] from the set of n criteria obtained as: ij m×r E E ... E 1 2 r ⎡ ⎤ m m m ψ Z (ψ , ψ ) Z (ψ , ψ ) ··· Z (ψ , ψ ) 1 1 t 2 1 t r 1 t t =1 1 t =1 1 t =1 1 1 1 1 m m m ⎢ ⎥ ψ Z (ψ , ψ ) Z (ψ , ψ ) ··· Z (ψ , ψ ) 1 2 t 2 2 t r 2 t1 t =1 1 t =1 1 t =1 ⎢ 1 1 1 ⎥ [t ] = ⎢ ⎥ ij m×r . . . . . . . . . ⎣ ⎦ . . . . m m m ψ Z (ψ , ψ ) Z (ψ , ψ ) ··· Z (ψ , ψ ) 1 m t 2 m t r m t t =1 1 t =1 1 t =1 1 1 1 1 (34) 374 V. ARYA AND S. KUMAR The positive ideal solution (PIS) and negative ideal solution(NIS) are calculated by using the dominance matrix. Therefore, to rank the alternatives, it is necessary to determine the ideal solutions. Since, the core idea of VIKOR method is based on the distance measures of alternatives from the ideal solutions and so traditional VIKOR method remains neu- tral regarding the attitudinal character of experts. In the proposed VIKOR-TODIM model, TODIM method was developed with IFNs to calculate the dominance matrix and then it was integrated with VIKOR method (which aims to obtain compromise solution) to obtain extraordinary ranking order. Step 7: Acquire PIS and NIS for each criterion. The positive solution is the best value H and the negative solution is the worst value H as follows: + + + + H = (H , H , ... , H ) 1 2 r ⎛ ⎞ m m m ⎝ ⎠ = max Z (ψ , ψ ),max Z (ψ , ψ ), ... ,max Z (ψ , ψ ) (35) 1 i t 2 i t r i t 1 1 1 i=1 i=1 i=1 t =1 t =1 t =1 1 1 1 − − − − and H = (H , H , ... , H ) 1 2 r ⎛ ⎞ m m m ⎝ ⎠ = min Z (ψ , ψ ),min Z (ψ , ψ ), ... ,min Z (ψ , ψ ) (36) 1 i t 2 i t r i t 1 1 1 i=1 i=1 i=1 t =1 t =1 t =1 1 1 1 ∗ ∗ Step 8: Obtain the values S and R bu using the below formulas: i i d (H , t ) ij S = u (37) + − d (H , H ) j j 1≤j≤r d (H , t ) h ij R = max u (38) + − 1≤j≤r d (H , H ) j j where m m d (H , t ) = max Z (ψ , ψ ) − Z (ψ , ψ ), h ij j i t j i t j 1 1 1≤i≤n t =1 t =1 1 1 m m + − d (H , H ) = max Z (ψ , ψ ) − min Z (ψ , ψ ). h j i t j i t j j 1 1 1≤i≤n 1≤i≤n t =1 t =1 1 1 th r and u ; j = 1(1)r denotes the weight of j criteria satisfying u = 1. j j j=1 Step 9: Obtain the influence values Q , i = 1(1)m with Equation (39) as : ∗ − ∗ − S − S R − R ∗ i i Q = τ + (1 − τ) . (39) − − ¯ ¯ S − S R − R − − ¯ ¯ Here S = max(S ), S = min(S ), R = max(R ) and R = min(R ). The coefficient τ and 1 − τ i i i i ∗ ∗ are introduced as a weight for ‘maximum group utility’ S and individual regrets (R ).In i i 1 1 general, we set the value to τ = and τ = denotes a consensus. 2 2 FUZZY INFORMATION AND ENGINEERING 375 ∗ ∗ Step 10: Determining the rank of the alternatives, sorting by the values of (S ), (R ) and i i (Q ) in descending order. Step 11: Calculating the compromise solution, the solution satisfying the following two conditions will be the most desirable solution. ∗ (2) ∗ (1) (1) (2) C1:If Q (ψ ) − Q (ψ ) ≥ , where ψ and ψ , respectively, stand at initial and r−1 second positions in the ranking list of Q and r denotes the number of alternatives. (1) ∗ ∗ C2: The alternative ψ should also be ranked first in the list of S or /and R . The com- i i promise solution Q is more stable in a decision making process and calculated with ’voting by majority rule’ (τ> 0.5), or ’by conseness’(τ = 0.5), or ’with veto’(τ< 0.5). If the condi- tions C1 and C2 are not simultaneously satisfied, then we seek the compromise solution as given below: (1) (2) (a) (Acceptable advantage): If only C2 is not satisfied then ψ and ψ are the compro- mise solutions. (b) (Acceptable stability): If C1 is not satisfied, then we explore the utmost value M by the following equation: ∗ (M) ∗ (1) Q (ψ ) − Q (ψ )< (40) r − 1 (i) Hence, it shows the alternatives ψ (i=1(1)M) are the real compromised solution(s). 6. Numerical Example This section presents a problem of evaluating university teaching hospital (Chandigarh, India) has been studied intends to construct a ‘Management Information System’ (MIS) in order to enhance better services and work productivity. Practical use of the pro- posed approach involves for illustrating the applicability of the proposed VIKOR-TODIM method. Assume that there are five possible alternatives as: ψ , ψ , ψ ,ψ and ψ have 1 2 3 4 5 been considered as alternatives. To perform the evaluation, a commission composed of five decision experts (say)DE , DE , DE , DE and DE have been established and five eval- 1 2 3 4 5 uation criteria which include Functionality (E ), Reliability (E ), Maintenance (E ),Cus- 1 2 3 tomer Satisfaction (E ), and Price (E ) are considered. Then, the evaluation by the rel- 4 5 evant experts of the alternatives ψ regarding each given criterion is presented by an IFN. Intuitionistic fuzzy decision matrices with their evaluation results are presented in Table 3. Now we put forward the proposed measure to prioritise these information management: Table 3. IFVs with different entropy measures. 2 2 3 4 Entropies M M M M M 2 2 2 4 V 0.3627 0.3620 0.2830 0.2423 0.2185 HY V 0.2119 0.2486 0.2253 0.1769 0.1478 ZJ ,β V 0.4650 0.4876 0.3241 0.2847 0.2417 GK R,S V 0.4352 0.4157 0.3476 0.3316 0.3017 =2 V 0.4168 0.4024 0.3945 0.3274 0.2912 376 V. ARYA AND S. KUMAR 6.1. Decision Analysis with Proposed Method Step1. We construct an Intuitionistic fuzzy decision matrix D which contains five criteria and five alternatives and is showed as below: Decision value E E E E E 1 2 3 4 5 ψ (0.60,0.10) (0.20,0.70) (0.50,0.40) (0.50,0.20) (0.20,0.80) ψ (0.30,0.50) (0.70,0.10) (0.20,0.80) (0.60,0.40) (0.50,0.30) ψ (0.30,0.20) (0.40,0.30) (0.40,0.20) (0.50,0.30) (0.40,0.20) ψ (0.10,0.40) (0.10,0.50) (0.40,0.60) (0.60,0.30) (0.70,0.20) ψ (0.20,0.30) (0.30,0.60) (0.40,0.50) (0.40,0.20) (0.10,0.50) Step2. We assume that each criterion is the cost criterion, the normalised decision matrix will become the following one: Decision value E E E E E 1 2 3 4 5 ψ (0.10,0.60) (0.70,0.20) (0.40,0.50) (0.20,0.50) (0.80,0.20) ψ (0.50,0.30) (0.10,0.70) (0.80,0.20) (0.40,0.60) (0.30,0.50) ψ (0.20,0.30) (0.30,0.40) (0.20,0.40) (0.30,0.50) (0.20,0.40) ψ (0.40,0.10) (0.50,0.10) (0.60,0.40) (0.30,0.60) (0.20,0.70) ψ (0.30,0.20) (0.60,0.30) (0.50,0.40) (0.20,0.40) (0.50,0.10) Step 3. The partial information on criterion weights is listed in set T: The entropy values for each criterion are given as below: 5 5 5 5 ˜ ˜ K = ξ = V (r ) = 0.102; K = ξ = V (r ) = 0.055 1 1j 1j 2 2j 2j i=1 i=1 i=1 i=1 5 5 5 5 ˜ ˜ K = ξ = V (r ) = 0.013; K = ξ = V (r ) = 0.013 3 3j 3j 4 4j 4j i=1 i=1 i=1 i=1 5 5 K = ξ = V (r ) = 0.093. 5 5j 5j i=1 i=1 The optimal model to determine the attribute weights can be constructed as; min T = 0.102u + 0.055u + 0.013u + 0.013u + 0.093u 1 2 3 4 5 ⎪ u ∈ S u = 1 such that j j=1 u ≥ 0, j = 1, 2, ... ,5. Corresponding to Equation (29), we have u = (0.2271, 0.2615, 0.1826, 0.1808, 0.148) . Step 4. We compute the dominance of each alternative ψ over the alternative ψ with i t respect to the given criterion (i, t = 1, 2, 3, 4, 5). Assume the value of τ = 2.5. From the set 1 FUZZY INFORMATION AND ENGINEERING 377 of criteria and the decision matrix , we can construct five dominance matrices Z − Z as 1 5 follows: ⎡ ⎤ E E E E E 1 2 3 4 5 ⎢ ⎥ 0.0000 −0.5310 −0.4598 −0.5936 −0.5310 ⎢ ⎥ ⎢ ⎥ ⎢ 0.3013 0.0000 0.2610 −0.4598 0.2610 ⎥ Z = ⎢ ⎥ ⎢ 0.2610 −0.4598 0.0000 −0.4598 −0.2655⎥ ⎢ ⎥ ⎣ 0.3369 0.2610 0.2131 0.0000 0.1507 ⎦ 0.3013 −0.4598 0.1507 −0.2655 0.0000 ⎡ ⎤ E E E E e 1 2 3 4 5 ⎢ ⎥ 0.0000 0.3960 0.3234 0.2800 0.1617 ⎢ ⎥ ⎢ ⎥ ⎢ −0.6060 0.0000 −0.4285 −0.6060 −0.5532⎥ Z = ⎢ ⎥ ⎢ −0.4948 −0.2800 0.0000 −0.4285 −0.4285⎥ ⎢ ⎥ ⎣ −0.4285 0.3960 0.2800 0.0000 0.2800 ⎦ −0.2474 0.3615 0.2800 −0.4285 0.0000 ⎡ ⎤ E E E E E 1 2 3 4 5 ⎢ ⎥ 0.0000 −0.5920 0.2340 −0.4186 −0.2960 ⎢ ⎥ ⎢ ⎥ ⎢ 0.2703 0.0000 0.3310 0.1911 0.2340 ⎥ Z = ⎢ ⎥ ⎢ −0.5127 −0.7251 0.0000 −0.5920 −0.5127⎥ ⎢ ⎥ ⎣ 0.1911 −0.4186 0.2703 0.0000 0.1351 ⎦ 0.1351 −0.5127 0.2340 −0.2960 0.0000 ⎡ ⎤ E E E E E 1 2 3 4 5 ⎢ ⎥ 0.0000 −0.5151 −0.2974 0.0000 −0.2974 ⎢ ⎥ ⎢ ⎥ 0.2330 0.0000 0.0000 0.1345 0.0000 ⎢ ⎥ Z = ⎢ ⎥ ⎢ 0.1345 0.0000 0.0000 0.1345 0.0000 ⎥ ⎢ ⎥ ⎣ 0.0000 −0.2974 −0.2974 0.0000 −0.5151⎦ 0.1345 0.0000 0.0000 0.2330 0.0000 ⎡ ⎤ E E E E E 1 2 3 4 5 ⎢ ⎥ 0.0000 0.2721 0.2981 0.2981 0.2434 ⎢ ⎥ ⎢ ⎥ −0.7350 0.0000 0.0000 0.1721 −0.6574 ⎢ ⎥ Z = ⎢ ⎥ ⎢ −0.8051 0.0000 0.0000 0.2108 −0.5693⎥ ⎢ ⎥ ⎣ −0.8051 −0.4649 −0.5693 0.0000 −0.8051⎦ −0.6574 0.2434 0.2108 0.2981 0.0000 Step5and6. With the help of Equation (34), computing the total dominance of each alternative ψ with respect to the alternatives ψ . Following the consequences of Step 4, i t the dominance matrix is therefore created as below: ⎡ ⎤ ψ ψ ψ ψ ψ 1 2 3 4 5 ⎢ ⎥ ψ −2.1154 0.3635 −0.9241 0.9617 −0.2733 ⎢ ⎥ ⎢ ⎥ ψ 1.1611 −2.1937 −1.6318 0.5275 −0.0344 ⎢ 2 ⎥ ⎢ ⎥ ⎢ ψ −1.0726 1.0264 −0.3425 0.1779 −0.4396⎥ ⎢ ⎥ ⎣ ⎦ ψ −1.1099 0.3675 0.2690 −1.1099 0.3675 ψ 1.1117 −1.2203 −1.1636 −2.6444 0.0949 5 378 V. ARYA AND S. KUMAR + − Step 7. With the help of Equations (35) and (36), the PIS H and the NIS H are acquiring as follows: H = (1.1611, 1.0264, 0.2690, 0.9617, 0.3675); H = (−2.1154, −2.1937, −2.3425, −2.6444, −0.4396). ∗ ∗ Steps 8. In this step, we calculate S and R as below: i i ∗ ∗ ∗ ∗ ∗ S = 0.6787, S = 0.5326, S = 0.6581, S = 0.2712, S = 0.5900; 1 2 3 4 5 ∗ ∗ ∗ ∗ ∗ R = 0.2270, R = 0.2270, R = 0.2270, R = 0.1573, R = 0.2270. 1 2 3 4 5 Step 9. Let the value of τ be 0.5. Compute the values of Q i = 1(1)5, ∗ ∗ ∗ ∗ ∗ Q = 1.0000, Q = 0.8207, Q = 0.9747, Q = 0.0000, Q = 0.8912. 1 2 3 4 5 ∗ ∗ ∗ Step 10. Sorting the values S , R and Q on the basic of ascending order, we can compute ranking of alternatives. Table 4 shows the results. From Table 5, we conclude that the most desirable alternative is ψ corresponding to different values of τ . From this analysis, it has been observed that the optimal alternative is remain same which shows that the proposed approach is more consistent and can be effectively used to solve the decision-making problems in a more efficient ways (Figure 2). Step 11. From Table 5, it is obvious that alternatives ψ and ψ stands at the initial and 4 2 ∗ ∗ (4) second positions in the ranking list of Q . However,applying the condition C1, Q (ψ ) − ∗ (2) 1 Q (ψ ) = 0.0000 − 0.8207 =−0.8207 < = 0.25. It implies that C1 is not satisfied. 5−1 Then , we search for compromise solution as follows: ∗ (5) ∗ (2) 1 Q (ψ ) − Q (ψ ) = 0.8912 − 0 = 0.8912 > = 0.25. Thus, ψ and ψ are taken 2 4 5−1 as compromised solutions. ∗ ∗ ∗ Table 4. The ranking C , D and Q and the compromise measure. ψ ψ ψ ψ ψ Ranking Compromise solutions 1 2 3 4 5 S 0.6787 0.5326 0.6581 0.2712 0.5900 ψ ψ ψ ψ ψ ψ 4 2 5 3 1 4 R 0.2270 0.2270 0.2270 0.1573 0.2270 ψ ψ = ψ = ψ = ψ ψ , ψ 4 2 5 3 1 4 2 Q 1.0000 0.8207 0.9747 0.0000 0.8912 ψ ψ ψ ψ ψ ψ , ψ 4 2 5 3 1 4 2 ∗ ∗ ∗ Table 5. The S , R , Q values obtained by the weight τ changes and the compromise solution. i i i ψ ψ ψ ψ ψ Ranking Compromise solutions 1 2 3 4 5 S 0.6787 0.5326 0.6581 0.2712 0.5900 ψ ψ ψ ψ ψ ψ 4 2 5 3 1 4 R 0.2270 0.2270 0.2270 0.1573 0.2270 ψ ψ = ψ = ψ = ψ ψ 4 2 5 3 1 4 0 1.0000 1.0000 1.0000 0.0000 1.0000 ψ = ψ = ψ = ψ = ψ ψ 4 2 5 3 1 4 0.1 1.0000 0.9641 0.9949 0.0000 0.9782 ψ ψ ψ ψ ψ ψ 4 2 5 3 1 4 0.2 1.0000 0.9283 0.9899 0.0000 0.9565 ψ ψ ψ ψ ψ ψ 4 2 5 3 1 4 0.3 1.0000 0.8924 0.9848 0.0000 0.9347 ψ ψ ψ ψ ψ ψ 4 2 5 3 1 4 Q (τ ) 0.5 1.0000 0.8207 0.9747 0.0000 0.8912 ψ ψ ψ ψ ψ ψ , ψ 4 2 5 3 1 2 4 0.6 1.0000 0.7849 0.9697 0.0000 0.8694 ψ ψ ψ ψ ψ ψ , ψ 4 2 5 3 1 2 4 0.7 1.0000 0.7490 0.9646 0.0000 0.8476 ψ ψ ψ ψ ψ ψ 4 2 5 3 1 4 0.8 1.0000 0.7132 0.9596 0.0000 0.8259 ψ ψ ψ ψ ψ ψ 4 2 5 3 1 4 0.9 1.0000 0.6773 0.9545 0.0000 0.8041 ψ ψ ψ ψ ψ ψ 4 2 5 3 1 4 1 1.0000 0.6415 0.9494 0.0000 0.7823 ψ ψ ψ ψ ψ ψ 4 2 5 3 1 4 FUZZY INFORMATION AND ENGINEERING 379 Figure 2. Sesnitivity analysis of the alternatives under IFSs. 6.2. Sensitivity Analysis During the decision analysis of proposed approach, the parameter τ plays a significant role for the assurance of Q , see (39), which demonstrates the weight age of the strategy of the ‘utmost group utility’. Specifically, when τ is 0, then the ranking is as ψ = ψ = ψ = 4 2 5 ψ = ψ =ψ . If 0.1 ≤ τ ≤ 1, then the ranking order is as follows: ψ ψ ψ ψ ψ 3 1 4 4 2 5 3 1 where means ‘preferred to’. In what pursues, we examine the impact of the parameter τ to the estimation of Q in detail. The results are appeared in Table 5. From this analysis, we conclude that the decision-maker can plan to choose the values of parameter and therefore, their respective alternatives according to his goal. Therefore, the proposed mea- sures give various choices to the decision-maker to reach the target. It can be seen that the change of τ has no impact on the ranking order. 6.3. Comparative Analysis and Further Discussion To verify the usefulness of the proposed MCDM method, the above example was solved by using the methods in the existing literature as given by the researchers with same attribute weights information and results are depicted in Table 6. Table 6. Comparison of the alternatives by different methods under IF environment. Methods Environment Model Ranking Boran et al. [38] IFSs TOPSIS ψ ψ ψ = ψ ψ 4 3 5 2 1 Gomes and Rangel [62] IFSs TODIM ψ ψ ψ ψ ψ 4 5 3 2 1 You and Liu [37] IFSs VIKOR ψ ψ ψ ψ ψ 4 1 2 3 5 Proposed method IFSs VIKOR-TODIM ψ ψ ψ ψ ψ 4 2 5 3 1 380 V. ARYA AND S. KUMAR The ranking of alternatives show that the first choice ψ as the most suitable alter- native in above methods. In our proposed method ψ is best choice but ranking order does not matter for other alternatives. Now, a very common query that arises in mind is that how the proposed method is superior to other methods for comparison. The weight assigned to each criteria can vary much ultimately obtained reflected in result of a method. This example proves that the proposed decision-making approach is competent to getting reasonable results. Thus , the priority of the new information measure is also verified. (1) Gomes and Rangel [62] approach does not consider the entropy information, which produces various results for different operators. In the proposed method, there is no risk of loss information because it is useful to consider the entropy information of the criteria as well as the discrimination between the alternatives. (2) Boran et al. [38] method only considers the distances from the ideal solution and from the negative ideal solution, without considering their relative importance. The final results yielded by the intuitionistic fuzzy TOPSIS are not always the closet to the ideal solution. (3) In comparison with You and Liu [37], the proposed method can effectively handle the MCDM problems by taking decision makers behaviours into consideration. The approach given by You and Liu [37] fails to get sensible decision reliability. (4) The results obtained by the proposed method found to be more accurate as it takes DMs’ bounded rationality into account as well as an aggregating function of all criteria with an advantage rate. As a result, the combined VIKOR-TODIM method can make full use of their merits in the complicated decision-making problems. The advantages of VIKOR and TODIM are all fully utilised. Hence, it can efficiently address the circumstance where DMs have limited rationality and some conflicting criteria. The modified VIKOR- TODIM method is more generalised and suitable to solve the real-life problem more accurately than the existing ones. Figure 3. Statistics of Comparison rankings. FUZZY INFORMATION AND ENGINEERING 381 A Graphical Analysis: For a quick analysis , the results of the above discussion are displayed in Figure 3. The Bar chart in Figure 3 has illustrated that all results from three methods are not exactly the same. In the present case, some parts of the outcomes are totally different. The functional utility of proposed technique is altogether clarified dealing with the issue of the board information framework in an educating clinic. 7. Conclusions In this study, when there is an absence of data and partial complete information and deci- sion experts are in an intuitionistic fuzzy environment, then we proposed a new intuition- istic information measure for IFSs. The largest contribution of the proposed work is that it integrated the advantages of IFNs, entropy information of the criteria, TODIM and VIKOR to well address the decision-making problem. Firstly, we have developed an entropy-weight method to determine the weights of criterion’s under intuitionistic fuzzy environment with the weights of the criteria are completely unknown. An advantage of the proposed method is it deduces the subjectivity influence and remains the original decision information suf- ficiently. Secondly, we have stretched out the VIKOR-TODIM method to deal with become an appropriate MCDM solution in which the importance of criteria are given as linguistics described by IFNs and the preference rating of alternatives. The functional utility of the proposed IF-approach is thoroughly explained dealing with the problem of management information system in a teaching hospital. The entropy-based VIKOR-TODIM method coor- dinated with IFS has excellent opportunities for success in multi-criteria decision-making because it implies relatively objective, intuitive and vague information of decision experts. The proposed VIKOR-TODIM approach has less loss of information and delivers more cor- rect, more general and precise results and can be utilised to tackle the uncertainties in decision-making problems under intuitionistic fuzzy environment. In the future, authors can expand the proposed method by applying various MCDM tech- niques (e.g. PROMETHEE, LINMAP (Linear Programming Method for Multidimensional Anal- ysis of Preference) , ELECTRE or TOPSIS ) to select best criteria. Furthermore, we should also be considered that the method could be found applications on some other problems like as strategic project evaluation, supplier selection, factory location, marketing management, renewable energy technologies, etc. The proposed IF-entropy measure can be extended to the concept of the symmetric or parametric divergence measure for Intuitionistic fuzzy sets, Picture fuzzy sets, Pythagorean fuzzy sets. Disclosure statement No potential conflict of interest was reported by the author(s). Notes on contributors Mr. Vikas Arya graduated from Kurukshetra University, Kurukshetra (KUK), in 2008. He completed his post-graduation in Mathematics in 2012 from KUK. In 2017, he joined Ph.D. at Maharishi Markandesh- war (Deemed to be University), Mullana, Ambala, and to date he is Research Scholar in MMDU, Mul- lana. His research interests include Information Theory, Fuzzy Information Measures and Divergence Measures. 382 V. ARYA AND S. KUMAR Dr. Satish Kumar completed his graduation in 1996 and postgraduation in 1998 from Kurukshetra University, Kurukshetra (KUK). He obtained his M.Phil. degree in 2000 from KUK. He was awarded Ph.D. degree in 2005 from Chaudhary Charan Singh (CCS) University, Meerut. After his Ph.D., he joined Jai Prakash Mukund Lal Institute of Engineering and Technology (JMIT) Radaur (Haryana) as a lecturer. In 2006, he shifted to Mekelle University Ethiopia as an Assistant Professor. After serving in Ethiopia for 2 years, he returned back to India and joined Geeta Institute of Management and Technology (GIMT), Kanipla (Haryana), as an Associate Professor and worked there till 2012. In 2012, he again shifted to Ethiopia and worked as a Professor in Arba Minch University till 2014. From 2014 to till date, he is working as a Professor in Maharishi Markandeshwar University, Mullana. His area of interest includes Information Theory and Coding, Fuzzy Information Measures. References [1] Shannon CE. The mathematical theory of communication. Bell Syst Tech J. 1948;27(3):379–423. [2] Zadeh LA. Fuzzy sets. Inf Cont. 1965;8:338–353. [3] Zadeh LA. Probability measures of fuzzy events. J Math Anal Appl. 1968;23(2):421–427. [4] DeLuca A, Termini S. A definition of non-probabilistic entropy in the setting of fuzzy set theory. Inf Cont. 1972;20:301–312. [5] Yager RR. On the measure of fuzziness and negation part I: membership in the unit interval. Int J Gen Syst. 1979;5(4):221–229. [6] Higashi M, Klir G. On measures of fuzziness and fuzzy complements. I J Gen Syst. 1982;8:169–180. [7] Arya V, Kumar S. Fuzzy entropy measure with an application in decision making under Bipolar fuzzy environment based on TOPSIS method. Int J Inf Manag Sci. 2020;31(2):99–121. [8] Arya V, Kumar S. Knowledge measure and entropy: a complementary concept in fuzzy theory. Granular Comput. 2020;1–13. https://doi.org/10.1007/s41066-020-00221-7 [9] Atanassov KT. Intuitionistic fuzzy sets. Fuzzy Sets Syst. 1986;20(1):87–96. [10] Kaufmann A. Introduction to the theory of fuzzy subsets. New York (NY): Academic Press; 1975. [11] Burillo P, Bustince H. Entropy on intuitionistic fuzzy sets and on interval-valued fuzzy sets. Fuzzy Sets Syst. 2001;118:305–316. [12] Szmidt E, Kacprzyk J. Entropy for intuitionistic fuzzy sets. Fuzzy Sets Syst. 2001;118(3):467–477. [13] Hung WL, Yang MS. Fuzzy entropy on intuitionistic fuzzy sets. Int J Intell Syst. 2006;21(4):443–451. [14] Verma R, Sharma BD. Exponential entropy on intuitionistic fuzzy sets. Kybernetika. 2013;49: 114–127. [15] Pal NR, Pal SK. Object background segmentation using new definitions of entropy. IEE Proc E. 1989;366:284–295. [16] Ye J. Two effective measures of intuitionistic fuzzy ertropy. Computing. 2010;87:55–62. [17] Ali MI, Feng F, Mahmood T, et al. A graphical method for ranking Atanassov’s intuitionistic fuzzy values using the uncertainty index and entropy. Int J Intell Syst. 2019;34(10):2692–2712. [18] Garg H, Agarwal N, Tripathi A. Generalized intuitionistic fuzzy entropy measure of order α and degree β and its applications to multi-criteria decision making problem. Int J Fuzzy Syst Appl. 2017;6(1):86–107. [19] Garg H, Kaur J. A novel (R, S)-norm entropy measure of intuitionistic fuzzy sets and its applications in multi-attribute decision-making. Mathematics. 2018;6(6):92. [20] Joshi R, Kumar S. Parametric (R, S)-norm entropy on intuitionistic fuzzy sets with a new approach in multiple attribute decision making. Fuzzy Inf Eng. 2017;9(2):181–203. [21] Garg H. Generalized intuitionistic fuzzy entropy-based approach for solving multi-attribute decision-making problems with unknown attribute weights. Proc Natl Acad Sci, India Sec A Phys Sci. 2017;89(1):129–139. [22] Ullah K, Hassan N, Mahmood T, et al. Evaluation of investment policy based on multi-attribute decision-making using interval valued T-spherical fuzzy aggregation operators. Symmetry. 2019;11(3):357. [23] Ullah K, Garg H, Mahmood T, et al. Correlation coefficients for T-spherical fuzzy sets and their applications in clustering and multi-attribute decision making. Soft Comput. 2020;24(3):1647–1659. FUZZY INFORMATION AND ENGINEERING 383 [24] Garg H, Gwak J, Mahmood T, et al. Power aggregation operators and VIKOR methods for complex q-rung orthopair fuzzy sets and their applications. Mathematics. 2020;8(4):538. [25] Arya V, Kumar S. Multi-criteria decision making problem for evaluating ERP system using entropy weighting approach and q-rung orthopair fuzzy TODIM approach. Granular Comput. 2020; https://doi.org/10.1007/s41066-020-00242-2. [26] Mahmood T Ashraf S, S Abdullah, et al. Different approaches to multi-criteria group deci- sion making problems for picture fuzzy environment. Bull Brazilian Math Soc, New Ser. 2019;50(2):373–397. [27] Garg H, Kaur G. Extended TOPSIS method for multi-criteria group decision-making problems under cubic intuitionistic fuzzy environments. Scientia Iranica. 2020;27(1):396–410. [28] Garg H, Arora R. TOPSIS method based on correlation coefficient for solving decision-making problems with intuitionistic fuzzy soft set information. AIMS Math. 2020;5(4):2944–2966. [29] Hwang CL, Yoon K. Multiple attribute decision making: methods and applications. Berlin: Springer-Verlag; 1981. [30] Krohling RA, Souza DE. Combining prospect theory and fuzzy numbers to multi-criteria decision making. Expert Syst Appl. 2012;39:11487–11493. [31] Fan ZP, Zhang X, Chen FD, et al. Extended TODIM method for hybrid multiple attribute decision making problems. Knowledge Based Syst. 2013;42:40–48. [32] Krohling RA, Pacheco AGC, Siviero ALT. IF-TODIM: an intuitionistic fuzzy TODIM to multicriteria decision making. Knowledge Based Syst. 2013;53:142–146. [33] Lourenzutti R, Krohling RA. A study of TODIM in a intuitionistic fuzzy and random environment. Expert Syst Appl. 2013;40:6459–6468. [34] Wang CZ, Shao MW, He Q. Feature subset selection based on fuzzy neighborhood rough sets. Knowledge Based Syst. 2016;111:173–179. [35] Wei CP, Ren ZL, Rodríguez RM. A hesitant fuzzy linguistic TODIM method based on a score function. I J Comput Intel Syst. 2015;8(4):701–712. [36] Opricovic S. Multicriteria optimization of civil engineering systems. Belgrade: Facultyu of Civil Engineering; 1998. [37] You XY, Liu HC. An extended vikor method using intuitionistic fuzzy sets and combination weights for supplier selection. Symmetry. 2017;9(9):169. [38] Boran FE, Genc S, Akay D. Personnel selection based on intuitionistic fuzzy sets. Hum Factors Ergon Manuf Sere Ind. 2011;21:493–503. [39] Garg H, Kumar K. A novel exponential distance and its based TOPSIS method for interval- valued intuitionistic fuzzy sets using connection number of SPA theory. Artif Intell Rev. 2020;53(1):595–624. [40] Kumar K, Garg H. Connection number of set pair analysis based TOPSIS method on intuitionistic fuzzy sets and their application to decision making. Appl Intell. 2018;48(8):2112–2119. [41] Benayoun R, Roy B, Sussman B. Electre: une methode pour guider le choix en presence de points de vue multiples, Note de travail 49. Direction Scientifique: SEMAMETRA International; 1966. [42] Brans JP, Mareschel V. PROMETHEE: a new family of outranking methods in multicriteria analysis. Multiple Criteria Decision Analysis: State of the Art Surveys, Editor: Brans JP. New York: North- Holland; 1984. vol. 84. p. 477–490. [43] Keshavarz Ghorabaee M, Zavadskas EK, Amiri M, et al. Multi-criteria evaluation of green sup- pliers using an extended WASPAS method with interval type-2 fuzzy sets. J Clean Prod. 2016;137:213–229. [44] Yuan J, Li C, Li W, et al. Linguistic hesitant fuzzy multi-criterion decision-making for renewable energy: a case study in Jilin. J Clean Prod. 2018;172:3201–3214. [45] Xu ZS. Intuitionistic fuzzy aggregations operators. IEEE Trans Inf Forensics Secur. 2007;15:1179– [46] Chen SM, Tan JM. Handling multi-criteria fuzzy decision-making problems based on vague set theory. Fuzzy Sets Syst. 1994;67:163–172. [47] Hong DH, Choi CH. Multi-criteria fuzzy decision-making problems based on vague set theory. Fuzzy Sets Syst. 2000;114:103–113. 384 V. ARYA AND S. KUMAR [48] Szmidt E, Kacprzyk J, Bujnowski P. How to measure amount of knowledge conveyed by Atanassov’s intuitionistic fuzzy sets. Inf Sci. 2014;7:276–285. [49] Renyi A. On measures of entropy and information. Proceedings of the 4th Barkley symposium on Mathematical statistics and probability; vol. 1. University of California Press; 1961. p. 547. [50] Tsallis C. Possible generalization of Boltzman–Gibbs statistics. J Stat Phys. 1988;52(1–2): 479–487. [51] Boekee DE, Vander Lubbe JCA. The R-norm information measure. Inf Cont. 1980;45:136–155. [52] Taneja IJ. A study of generalized measures of information theory [Ph.D. thesis]. University of Delhi; 1975. [53] Havrda JH, Charvat F. Quantification method classification process: concept of structural α- entropy. Kybernetika. 1967;3:30–35. [54] Aczel J, Darocozy Z. On measures of information and their characterization. New York: Academic Press; 1975. [55] Arimoto K. Information-theoretical considerations on estimations problems. Inf Cont. 1971;19: 181–194. [56] Bhandari D, Pal NR. Some new information measures for fuzzy sets. Inf Sci. 1993;67(3):209–228. [57] Zeng WY, Li HX. Relationship between similarity measure and entropy of interval-valued fuzzy sets. Fuzzy Sets Syst. 2006;157(11):1477–1484. [58] Zhang QS, Jiang SY. A note on information entropy measure for vague sets. Inf Sci. 2008;178(21):4184–4191. [59] De SK, Biswas R, Roy AR. Some operations on intuitionistic fuzzy sets. Fuzzy Sets Syst. 2000;114:477–484. [60] Yu PL. A class of solutions for group decision problems. Manage Sci. 1973;19(8):936–946. [61] Xu ZS, Hu H. Projection models for intuitionistic fuzzy multiple attribute decision making. Int J Inf Tech Deci Making. 2010;9(2):267–280. [62] Gomes L, Rangel L. An application of the todim method to the multicriteria rental evaluation of residential properties. Eur J Oper Res. 2009;193:204–211.
Fuzzy Information and Engineering – Taylor & Francis
Published: Jul 3, 2019
Keywords: Intuitionistic fuzzy set; Havrda–Charvat–Tsallis entropy; hamming distance; modified VIKOR-TODIM model; 94A15; 94A17; 26D15; 68R10; 68R05
You can share this free article with as many people as you like with the url below! We hope you enjoy this feature!
Read and print from thousands of top scholarly journals.
Already have an account? Log in
Bookmark this article. You can see your Bookmarks on your DeepDyve Library.
To save an article, log in first, or sign up for a DeepDyve account if you don’t already have one.
Copy and paste the desired citation format or use the link below to download a file formatted for EndNote
Access the full text.
Sign up today, get DeepDyve free for 14 days.
All DeepDyve websites use cookies to improve your online experience. They were placed on your computer when you launched this website. You can change your cookie settings through your browser.