You are on page 1of 6

SoftwareX 20 (2022) 101271

Contents lists available at ScienceDirect

SoftwareX
journal homepage: www.elsevier.com/locate/softx

Original software publication

pyFDM: A Python library for uncertainty decision analysis methods



Jakub Więckowski a , Bartłomiej Kizielewicz a , Wojciech Sałabun b ,
a
National Institute of Telecommunications, Szachowa 1, 04-894 Warsaw, Poland
b
Research Team on Intelligent Decision Support Systems, Department of Artificial Intelligence and Applied Mathematics, Faculty of Computer Science
and Information Technology, West Pomeranian University of Technology, 71-210 Szczecin, Poland

article info a b s t r a c t

Article history: Decision-making is an integral part of life. Many situations require us to make rational choices among
Received 26 September 2022 numerous decision options. In addition, real-world problems are characterized by uncertainties that
Received in revised form 9 November 2022 hinder the entire process. Multi-Criteria Decision Analysis (MCDA) comes with help. Many evaluation
Accepted 15 November 2022
techniques have been developed along with fuzzy logic assumptions, classified as fuzzy MCDA tools.
Keywords: These methods allow efficient and effective evaluation of decision variants in an uncertain environ-
Fuzzy logic ment. Existing tools facilitate the work of these methods, but they are limited or not updated with
Uncertain data newly established technologies. Therefore, this article proposes the pyFDM (Python Fuzzy Decision
Python Making): a new Python 3 software library to facilitate calculations using methods that operate on
Triangular fuzzy numbers uncertain data. In addition to evaluation methods, various techniques for normalization, defuzzification,
Fuzzy Decision Making distance measure, or objective criteria weighting are also included. It provides a comprehensive set of
tools for using available techniques to perform calculations in this field. The operation of the proposed
library is demonstrated through two practical examples. The first is directed to selecting the most
appropriate stock, while the second is devoted to the ERP system selection firms. The presented
examples show that the implemented library can be effectively used for fuzzy decision-making
problems based on Triangular Fuzzy Numbers. The modular structure allows the interchangeable use
of available methods, making the tool a comprehensive environment for calculations in this area.
© 2022 The Author(s). Published by Elsevier B.V. This is an open access article under the CC BY license
(http://creativecommons.org/licenses/by/4.0/).

Code metadata

Current code version v1.0.0


Permanent link to code/repository used for this code version https://github.com/ElsevierSoftwareX/SOFTX-D-22-00305
Legal Code License MIT Licence
Code versioning system used Git
Software code languages, tools, and services used Python
Compilation requirements, operating environments & dependencies Python v3.6, numpy
Support email for questions j.wieckowski@il-pib.pl

Software metadata

Current software version v1.0.0


Permanent link to executables of this version https://github.com/jwieckowski/pyfdm
Legal Software License MIT Licence
Computing platforms/Operating systems Microsoft Windows, Linux
Software code languages, tools, and services used Python
Compilation requirements, operating environments & dependencies Python v3.6, numpy
Link to user manual https://pyfdm.readthedocs.io/en/latest/
Support email for questions j.wieckowski@il-pib..pl

∗ Corresponding author.
E-mail address: wojciech.salabun@zut.edu.pl (Wojciech Sałabun).

https://doi.org/10.1016/j.softx.2022.101271
2352-7110/© 2022 The Author(s). Published by Elsevier B.V. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).
Jakub Więckowski, Bartłomiej Kizielewicz and Wojciech Sałabun SoftwareX 20 (2022) 101271

1. Motivation and significance

Many daily problems are complex and require the analysis of


multiple parameters that make up the final outcome [1]. Multi-
Criteria Decision Analysis (MCDA) methods are used to help find
rational solutions [2]. For MCDA methods to function properly,
it is necessary to define a set of criteria to create a decision
matrix on its basis, which is used to assess the attractiveness of
the alternatives [3]. In addition, it is necessary to indicate the
relevance of each criterion through a method for determining the
criterion weights or to provide the criterion weights explicitly [4].
Furthermore, apart from the vector of criterion weights, which
allows the influence of each parameter on the final result to be
differentiated, the type of criteria (Profit or Cost) should also
be indicated, which will determine the need to maximize or
minimize the value of a criterion. Fig. 1 shows the flowchart of
Fig. 1. Subsequent steps of performing multi-criteria decision analysis
general steps in applying multi-criteria methods. Using MCDA assessment.
methods, Decision Support Systems (DSS) [5] are designed and
are readily and effectively used in many areas, such as supplier
selection [6], location selection [7], industry [8], healthcare [9],
is defuzzified to convert a fuzzy set into a crisp value [31]. With
and sports management [10], among others. Due to the practical the existing possibilities of using different metrics, it is necessary
applicability of MCDA methods, much work is directed toward to enable the expert to use these techniques modularly, providing
developing these methods to improve their effectiveness and him with a comprehensive tool for researching the field of fuzzy
efficiency. In addition, new extensions to existing methods are logic in the multi-criteria area.
being developed to improve the decision support process. Despite the popularity of MCDA methods, there is a significant
The initial concept of MCDA methods was to operate on crisp lack of tools that enable intuitive and compact use of these tech-
values [11]. However, with the development of work on these niques in practical terms. Most of the publicly available software
techniques, it was noticed that the characteristics of real prob- lacks support for the latest technologies, contains only several
lems deviate from such data representation. Fuzzy logic principles MCDA techniques [32], is not updated, or operates in a multi-
came with help, which made it possible to model uncertain data criteria environment based on crisp data [33,34]. There are also
in MCDA methods evaluation [12]. Introducing fuzzy sets into some commercial solutions [35], but their main drawback is the
multi-criteria techniques made it possible to define criteria values need to purchase a license to use the tool, making them less
as a range of values within which the actual value describing attractive to most of the potential users. Due to the increasing im-
a given decision option falls [13]. One of the most widely used portance of fuzzy logic in multi-criteria problems, it is necessary
solutions representing traditional fuzzy sets is Triangular Fuzzy to fill the gap created by providing a tool that takes into account
Numbers (TFN), whose value is represented as x = [a, m, b] [14, fuzzy sets in the operation of MCDA methods.
15]. In addition, the TFN assumptions allow the range of values In this paper, we propose the pyFDM (Python Fuzzy Deci-
to be represented asymmetrically [16]. The practical potential sion Making) package, which contains the functionality of the
of fuzzy logic has influenced the introduction of extensions to selected MCDA methods operating in a fuzzy environment. It also
existing MCDA methods, enriched by using fuzzy set arithmetic provides different data normalization methods, distance metrics,
in calculating the attractiveness of decision options [17,18]. and defuzzifications. The proposed library is based on Python 3
Among the conventional MCDA methods are Multi-Objective and can be downloaded with the Python Package Index (PyPI). It
Optimization on the basis of Ratio Analysis (MOORA) [19], VlseKri- makes it freely accessible for a broad group of users operating
terijumska Optimizacija I Kompromisno Resenje (VIKOR) [20], in a scientific field or for personal purposes. The source code
Technique for Order of Preference by Similarity to Ideal Solution is also available on Github [36]. The documentation is provided
(TOPSIS) [21], for example. To make it possible to maintain a on Read the Docs website [37]. The pyFDM contains ten fuzzy
customized approach to calculations, taking into account the MCDA methods, namely Additive Ratio ASsessment (ARAS) [38],
assumptions of fuzzy logic, many crisp methods have been ex- COmbinative Distance-based ASsessment (CODAS) [39], Complex
tended to use fuzzy sets. Methods such as Fuzzy TOPSIS, Fuzzy Proportional Assessment (COPRAS) [40], Evaluation based on Dis-
VIKOR, or Fuzzy MOORA have been repeatedly used in practical tance from Average Solution (EDAS) [41], Multi-Attribute Bound-
decision-making problems, where they have shown higher per- ary Approximation Area Comparison (MABAC) [42],
formance compared to the traditional variant operating on crisp Multi-Attributive Ideal-Real Comparative Analysis (MAIRCA) [43],
numbers [22–25]. However, the number of methods developed MOORA [44], Operational Competitiveness RAting (OCRA) [45],
causes difficulties in choosing the appropriate technique for the TOPSIS [46], and VIKOR [47]. All the implemented methods op-
indicated problem. The use of different approaches in the data erate on the Triangular Fuzzy Numbers, as this extension finds
conversions translates into differences in the proposed rankings the greatest interest while handling the fuzzy multi-criteria prob-
between the various methods [26,27]. Consequently, the results lems.
obtained by these methods are only suggestions for the expert
making the final decision. 2. Software description
In addition, it is also essential to mention the influence of
other parameters on the obtained ambiguity of the results. One MCDA methods are based on different approaches to calcu-
factor is the data normalization method used [28]. Others are the lating the alternative preference values. As in the crisp environ-
metrics used to calculate distances between values [29]. Studies ment, this concept is maintained when handling the uncertainty
show that using different techniques within the same test data with the fuzzy extensions. However, these methods use similar
yields different results [30]. Another element that affects the mechanisms at different stages of the calculation. The general
outcomes calculated using fuzzy MCDA methods is how the data schema of the fuzzy MCDA methods performance includes: (1)
2
Jakub Więckowski, Bartłomiej Kizielewicz and Wojciech Sałabun SoftwareX 20 (2022) 101271

(4) Min–Max (minmax_normalization),


(5) Vector (vector_normalization), and
(6) SAW normalization (saw_normalization)
• defuzzifications: this module includes four methods that
can be used to defuzzify the Triangular Fuzzy Numbers into
the crisp value. The techniques that can be used for this
purpose are:
(1) Graded mean average
(graded_mean_average_defuzzification),
(2) Mean (mean_defuzzification),
(3) Mean area (mean_area_defuzzification),
(4) Weighted mean (weighted_mean_defuzzification)
• distances: to calculate the distance between two Triangular
Fuzzy Numbers, the specific methods should be used. In the
case of TOPSIS and CODAS methods, the user can select the
following distance measures:
(1) Euclidean (euclidean_distance),
(2) Weighted Euclidean (weighted_euclidean_distance),
(3) Hamming (hamming_distance),
(4) Weighted Hamming (weighted_hamming_distance),
(5) Vertex (vertex_distance),
(6) Tran Duckstein (tran_duckstein_distance),
(7) L-R (lr_distance),
(8) Mahdavi (mahdavi_distance)
• correlations: comparing results similarity is an integrated
part of the MCDA usage. To this end, in this module four co-
Fig. 2. pyFDM package structure. efficients are available to use, namely: (1) Spearman corre-
lation coefficient (spearman_coef), (2) Pearson correlation
coefficient (pearson_coef), (3) Spearman weighted corre-
Determination of the initial data, namely the sets of criteria, al- lation coefficient (weighted_spearman_coef), and (4) WS
ternatives, and weights, (2) Selection of the fuzzy MCDA method, ranking similarity coefficient (ws_rank_similarity_coef)
(3) Selection of the additional measures, such as normalization, • helpers: it includes additional methods which can be useful
defuzzification, distance methods, (4) Calculation of the alter- when performing a multi-criteria assessment. The meth-
native preference values, (5) Calculation of the results rankings ods for calculating the ranking of the obtained preference
and their similarity. The pyFDM library covers all the mentioned- values (rank) and for the random generation of the Tri-
above steps and allows for changing the measures modularly. It is angular Fuzzy Number decision matrix with a given size
one of the main advantages of the proposed tool, as it can be used (generate_fuzzy_matrix) are provided
with the default parameters or with modified measures. Fig. 2 • validator: this module provides methods for testing the
presents the general structure of the pyFDM library modules. correctness of the input data given by the user. It enables
to verify of the input data dimensionality, type of decision
2.1. Software architecture matrix, the sum of weights, the types of criteria

The pyFDM library is a tool that provides fuzzy MCDA meth- 2.2. Software functionalities
ods, objective weighting methods, normalization methods,
defuzzification methods, distance metrics, and correlation coeffi- To pyFDM was created as a comprehensive tool covering the
cients. Moreover, there are placed helper methods and a validator main functionalities of the multi-criteria assessment in the fuzzy
to verify the correctness of the input parameters given by the environment based on the Triangular Fuzzy Numbers. It enables
to use of features like:
user. The tool modules are described as follows:
• Evaluating decision variants when their characteristics can-
• methods: this module contains ten fuzzy MCDA methods
not be represented as the crisp value
based on the Triangular Fuzzy Numbers, namely (1) fuzzy
• Calculating the alternative preference values based on se-
ARAS (fARAS), (2) fuzzy CODAS (fCODAS), (3) fuzzy COR-
lected fuzzy MCDA method with Triangular Fuzzy Number
PAS (fCOPRAS), (4) fuzzy EDAS (fEDAS), (5) fuzzy MABAC
extension
(fMABAC), (6) fuzzy MAIRCA (fMAIRCA), (7) fuzzy MOORA
• Selecting the different metrics responsible for normaliza-
(fMOORA), (8) fuzzy OCRA (fOCRA), (9) fuzzy TOPSIS tion, distance calculation, or defuzzification
(fTOPSIS), (10) fuzzy VIKOR (fVIKOR) • Calculating the weights with the given objective weighting
• weights: four objective weighting methods was provided: method operating in the fuzzy environment
(1) Equal (equal_weights), • Establishing the similarity of the results with selected cor-
(2) Shannon entropy (shannon_entropy_weights), (3) relation coefficients
Standard deviation (standard_deviation_weights), (4) • Calculating the positional ranking based on the preference
Variance (variance_weights) values
• normalizations: to provide comprehensive tool for initial • Determinating the random Triangular Fuzzy Number deci-
data manipulation, this module provides six normalization sion matrix that can be used especially for research pur-
methods, namely: poses to examine the generalized performance of this fuzzy
(1) Sum (sum_normalization), extension
(2) Max (max_normalization), • Testing performance of different combinations of the avail-
(3) Linear (linear_normalization), able methods changing them modularly
3
Jakub Więckowski, Bartłomiej Kizielewicz and Wojciech Sałabun SoftwareX 20 (2022) 101271

Table 1
Decision matrix for Example 1.
C1 C2 C3
A1 (3, 4, 5) (4, 5, 6) (8, 9, 9)
A2 (6, 7, 8) (4, 5, 6) (1, 2, 3)
A3 (5, 6, 7) (2, 3, 4) (3, 4, 5)
A4 (8, 9, 9) (2, 3, 4) (2, 3, 4)
A5 (7, 8, 9) (7, 8, 9) (5, 6, 7)
Weight 0.394 0.084 0.522
Type Cost Profit Profit

2.3. Sample code snippets

Below we present a sample code snippet 1, where one of


the fuzzy MCDA methods, namely fuzzy ARAS, is used to assess
alternatives. The shown calculation is based on the given decision
matrix, criteria weights, and their types. The user can provide the
data directly, or he can choose to randomly generate a triangular
fuzzy decision matrix with the ’generate_fuzzy_matrix’ method
and calculate the criteria weights with one of the objective tech-
niques from the ’weights’ module implemented in the library.
Moreover, the criteria types can be represented as ’1’ for the profit
and ’-1’ for the cost. The presented decision matrix contains 4
alternatives and 3 criteria, which reflect the number of rows and Fig. 3. Rankings similarity obtained by different fuzzy MCDA methods for
columns in the array respectively. Example 1.

Listing 1: Example of fuzzy assessment based on the pyFDM


library. Example 1. The first problem is directed to selecting the most ap-
 ⊵
1 from pyfdm . methods import fARAS
propriate stock and was already addressed by Narang in 2021 [48],
2 from pyfdm . helpers import rank where we can find names of the used criteria and alternatives.
3 import numpy as np The decision matrix used by Narang in his research was presented
4 in Table 1. The weights and types of criteria were also provided
5 if __name__ == ' __main__ ' :
6 matrix = np. array ([
to indicate all data needed for the assessment.
7 [[5 , 7, 9], [5, 7, 9], [7, 9, 9]], The authors used the fuzzy COPRAS method to provide the
8 [[1 , 3, 5], [3, 5, 7], [3, 5, 7]],
9 [[1 , 1, 3], [1, 3, 5], [1, 3, 5]],
alternatives’ ranking. Nevertheless, we decided to verify all im-
10 [[7 , 9, 9], [7, 9, 9], [7, 9, 9]] plemented multi-criteria techniques that were included in the
11 ]) library. Thanks to this, 10 rankings were obtained and shown
12 in Table 2. Due to different approaches used by fuzzy evalua-
13 weights = np. array ([[5 , 7, 9], [7, 9, 9], [3, 5,
7]]) tion methods, some discrepancies in the rankings can be no-
14 types = np. array ([1, -1, 1]) ticed. However, it is worth noting that 1st and 2nd places in
15 the rankings indicated by all used methods remain the same.
16 f_aras = fARAS () Different alternatives’ hierarchies are obtained below 3rd position
17 pref = f_aras (matrix , weights , types)
18
for selected fuzzy techniques.
19 print (f ' Fuzzy ARAS preferences : {pref} ' ) In addition, the similarity of the obtained results should be
20 print (f ' Fuzzy ARAS ranking : {rank(pref)} ' ) examined to indicate how coherent is methods’ performance. For
  this purpose, the ’ws_rank_similarity_coef’ implemented in the
Using the assumptions of Object-Oriented Programming (OOP), library was used. The visualization of the obtained rankings’ sim-
the fuzzy ARAS object is then created, with the default parame- ilarities is presented in Fig. 3. High correlations are represented
ter for the normalization technique set to ‘sum_normalization’. as values close to 1, and lower values indicate relatively lower
It can be replaced with one of the methods implemented in rankings similarity. Then, the fuzzy MAIRCA method provides the
the ’methods/fuzzy_sets/tfn/normalizations’ module. Calling the lowest correlated results compared to other techniques.
fuzzy ARAS object allows for obtaining the preference values of
the considered alternatives. Based on the obtained assessment, Example 2. The second practical example cover the problem of
the positional ranking of the alternatives can be calculated with ERP system selection firms presented by Karande in 2012 [44],
the ‘rank’ method from the ‘helpers’ module. Knowing that the where we can find names of the used criteria and alternatives. In
fuzzy ARAS method assigns higher preference values to better- Table 3, the decision matrix, weights, and types of criteria used
ranked alternatives, the preference array should be sorted in in the assessment were presented. The fuzzy MOORA method
descending order, and this is the default parameter value for combined with vector normalization was used to perform the
the ‘rank’ method, as most of the methods are based on this evaluation. To verify the normalization techniques provided in
assumption. the library and indicate their impact on the outcome, the available
methods were tested within the given problem. In Fig. 4 the
3. Illustrative examples visualization of the normalized values for each criterion calcu-
lated with different techniques are presented. Diverse formulas
To present the selected functionalities of the implemented applied in the calculations caused significant differences in the
library, two practical examples are shown below. They illustrate obtained results. Table 4 includes the rankings determined based
how the proposed tool can be used in multi-criteria problems on the performed evaluations using the fuzzy MOORA technique
based on the Triangular Fuzzy Numbers. combined with 6 normalization methods.
4
Jakub Więckowski, Bartłomiej Kizielewicz and Wojciech Sałabun SoftwareX 20 (2022) 101271

Table 2
Rankings calculated with different fuzzy MCDA methods for Example 1.
fARAS fCODAS fCOPRAS fEDAS fMABAC fMAIRCA fMOORA fOCRA fTOPSIS fVIKOR
A1 1 1 1 1 1 1 1 1 1 1
A2 4 4 4 4 4 4 4 5 5 4
A3 3 3 3 3 3 3 5 3 3 3
A4 5 5 5 5 5 5 4 4 4 5
A5 2 2 2 2 2 2 2 2 2 2

Table 3
Decision matrix for Example 2.
Ai C1 C2 C3 C4 C5
A1 (0.6, 0.8, 1.0) (0.6, 0.8, 1.0) (0.4, 0.6, 0.8) (0.2, 0.4, 0.6) (0.8, 1.0, 1.0)
A2 (0.4, 0.6, 0.8) (0.6, 0.8, 1.0) (0.4, 0.6, 0.8) (0.8, 1.0, 1.0) (0.2, 0.4, 0.6)
A3 (0.8, 1.0, 1.0) (0.4, 0.6, 0.8) (0.6, 0.8, 1.0) (0.0, 0.2, 0.4) (0.0, 0.2, 0.4)
A4 (0.2, 0.4, 0.6) (0.4, 0.6, 0.8) (0.4, 0.6, 0.8) (0.6, 0.8, 1.0) (0.4, 0.6, 0.8)
Weight 0.364 0.271 0.203 0.093 0.068
Type Cost Profit Profit Profit Profit

development, industry, and sports management, among other


fields, where many factors influence the results and partial data
is not precisely known. Furthermore, the proposed library could
be used in the academic educational field. It could introduce to
students the assumptions of the fuzzy logic and MCDA methods
and be tested for different use cases.
The pyFDM package architecture was designed to provide
modularity and transparency in its usage. The aim was to present
a tool allowing for complex calculations in the fuzzy environ-
ment that can be performed by less and more experienced users
and with different level of MCDA methods familiarity. A set of
examples for each functionality shows how to properly perform
calculations. It makes it possible to examine available methods’
performance and establish a comparative analysis of the obtained
results to emphasize the differences and similarities between
implemented techniques.

5. Conclusions

Fig. 4. Normalized values from decision matrix obtained by different


In this paper, we present the Python 3 library dedicated to
normalizations methods for Example 2. the multi-criteria assessment combined with the Triangular Fuzzy
Numbers to allow performing calculations in the fuzzy environ-
Table 4 ment. Additional measures like normalizations, distance metrics,
Ranking comparison obtained by using different normalizations and the fuzzy defuzzifications, objective weighting methods, and correlation co-
MOORA method for Example 2. efficients are provided to create a comprehensive tool for decision
Ai Sum Max Linear Min–Max Vector SAW variants assessments. Illustrative examples show the practical
A1 3 1 1 1 3 3 potential of the proposed tool. In addition, the example code snip-
A2 2 3 2 3 2 2 pets for each functionality are uploaded on the GitHub repository,
A3 4 2 3 2 4 4
A4 1 4 4 4 1 1 which, combined with the provided documentation, allows for
the use of available techniques for research and personal pur-
poses. The library structure was designed to extend current func-
tionalities with new methods and fuzzy extensions. The pyFDM
4. Impact package includes selected fuzzy MCDA methods and other addi-
tional techniques. However, the range of covered methods could
Currently, there exist only several solutions supporting the be widened to provide a complete research tool. Moreover, the
multi-criteria evaluation functionalities. Moreover, most of them main aim for future directions is to extend the package with other
operate in a crisp environment or offer a single method to assess fuzzy extensions.
without additional techniques that helps to examine the consid-
ered set. To fill the existing gap of freely available tools dedicated
to performing multi-criteria calculations in a fuzzy environment, Declaration of competing interest
the pyFDM library is presented. It could find interest in a group
of researchers handling the multi-criteria assessment combined The authors declare that they have no known competing finan-
with the Triangular Fuzzy Numbers. The practical potential of the cial interests or personal relationships that could have appeared
implemented techniques could also be exploited in sustainable to influence the work reported in this paper.
5
Jakub Więckowski, Bartłomiej Kizielewicz and Wojciech Sałabun SoftwareX 20 (2022) 101271

Data availability [22] Emovon I, Okpako OS, Edjokpa E. Application of fuzzy MOORA method in
the design and fabrication of an automated hammering machine. World J
Eng 2020.
No data was used for the research described in the article.
[23] Kizielewicz B, Bączkiewicz A. Comparison of Fuzzy TOPSIS, Fuzzy VIKOR,
Fuzzy WASPAS and Fuzzy MMOORA methods in the housing selection
Acknowledgment problem. Procedia Comput Sci 2021;192:4578–91.
[24] Matawale CR, Datta S, Mahapatra S. Supplier selection in agile supply
The work was supported by the National Science Centre, chain: Application potential of FMLMCDM approach in comparison with
Fuzzy-TOPSIS and Fuzzy-MOORA. Benchmarking Int J 2016.
Poland, Decision number 2018/29/B/HS4/02725 (J.W., B.K. and [25] Petrović G, Mihajlović J, Ćojbašić Ž, Madić M, Marinković D. Comparison
W.S.). of three fuzzy MCDM methods for solving the supplier selection problem.
Facta Univ Ser Mech Eng 2019;17(3):455–69.
References [26] Bikmukhamedov R, Yeryomin Y, Seitz J. Evaluation of MCDA-based han-
dover algorithms for mobile networks. In: 2016 eighth international
conference on ubiquitous and future networks. ICUFN, IEEE; 2016, p.
[1] Liao H, Jiang L, Xu Z, Xu J, Herrera F. A linear programming method for
810–5.
multiple criteria decision making with probabilistic linguistic information.
[27] Sałabun W, Wątróbski J, Shekhovtsov A. Are MCDA Methods Benchmark-
Inform Sci 2017;415:341–55.
able? A Comparative Study of TOPSIS, VIKOR, COPRAS, and PROMETHEE II
[2] Mühlbacher AC, Kaczynski A. Making good decisions in healthcare with
Methods. Symmetry 2020;12(9):1549.
multi-criteria decision analysis: The use, current research and future
[28] Banaeian N, Mobli H, Fahimnia B, Nielsen IE, Omid M. Green supplier
development of MCDA. Appl Health Econ Health Policy 2016;14(1):
selection using fuzzy group decision making methods: A case study from
29–40.
the agri-food industry. Comput Oper Res 2018;89:337–47.
[3] Frazão TD, Camilo DG, Cabral EL, Souza RP. Multicriteria decision analysis
[29] Rusdiana U, Ernawati I, Falih N, Arista A. Comparison of distance metrics
(MCDA) in health care: a systematic review of the main characteristics and
on fuzzy C-means algorithm through customer segmentation. In: 2021 in-
methodological steps. BMC Med Inform Decis Mak 2018;18(1):1–16.
ternational conference on informatics, multimedia, cyber and information
[4] Kizielewicz B, Wątróbski J, Sałabun W. Identification of relevant crite-
system (ICIMCIS). IEEE; 2021, p. 307–11.
ria set in the MCDA process—Wind farm location case study. Energies
[30] Lin S-S, Shen S-L, Zhang N, Zhou A. Comprehensive environmental impact
2020;13(24):6548.
evaluation for concrete mixing station (CMS) based on improved TOPSIS
[5] de Souza Melaré AV, González SM, Faceli K, Casadei V. Technologies and
method. Sustainable Cities Soc 2021;69:102838.
decision support systems to aid solid-waste management: a systematic
[31] Tang C, Xu D, Chen N. Sustainability prioritization of sewage sludge to
review. Waste Manage 2017;59:567–84.
energy scenarios with hybrid-data consideration: a fuzzy decision-making
[6] Maghsoodi AI, Kavian A, Khalilzadeh M, Brauers WK. CLUS-MCDA: A novel
framework based on full consistency method and fusion ranking model.
framework based on cluster analysis and multiple criteria decision theory
Environ Sci Pollut Res 2021;28(5):5548–65.
in a supplier selection problem. Comput Ind Eng 2018;118:409–22.
[32] Duminy N. Package for multi-criteria decision analysis. 2021, https://pypi.
[7] Ulutaş A, Balo F, Sua L, Demir E, Topal A, Jakovljević V. A new integrated
org/project/mcda//. [Online; accessed 08-July-2022].
grey MCDM model: Case of warehouse location selection. Facta Univ Ser
[33] Meyer P, Bigaret S. Diviz: A software for modeling, processing and sharing
Mech Eng 2021.
algorithmic workflows in MCDA. Intell Decis Technol 2012;6(4):283–96.
[8] Youssef MI, Webster B. A multi-criteria decision making approach to
[34] Wątróbski J, Bączkiewicz A, Sałabun W. pyrepo-mcda—Reference objects
the new product development process in industry. Rep Mech Eng
based MCDA software package. SoftwareX 2022;19:101107.
2022;3(1):83–93.
[35] Dabbagh N, Fake H. Tech select decision aide: A mobile application to facil-
[9] Dell’Ovo M, Capolongo S, Oppio A. Combining spatial analysis with MCDA
itate just-in-time decision support for instructional designers. TechTrends
for the siting of healthcare facilities. Land Use Policy 2018;76:634–44.
2017;61(4):393–403.
[10] Barajas A, Castro-Limeres O, Gasparetto T. Application of MCDA to evaluate
[36] Wiȩckowski J. PyFDM library package repository. 2022, https://github.com/
financial fair play and financial stability in European football clubs. J. Sports
jwieckowski/pyfdm. [Online; accessed 26-September-2022].
Econ Manage 2017;7(3):143–64.
[37] Wiȩckowski J. PyFDM library documentation. 2022, https://pyfdm.
[11] Regier DA, Peacock S. Theoretical foundations of MCDA. In: Multi-
readthedocs.io/en/latest/. [Online; accessed 26-September-2022].
criteria decision analysis to support healthcare decisions. Springer; 2017,
[38] Fu Y-K, Wu C-J, Liao C-N. Selection of in-flight duty-free product suppliers
p. 9–28.
using a combination fuzzy AHP, fuzzy ARAS, and MSGP methods. Math
[12] Stewart TJ. Dealing with uncertainties in MCDA. In: Multiple criteria
Probl Eng 2021;2021.
decision analysis: state of the art surveys. Springer; 2005, p. 445–66.
[39] Panchal D, Chatterjee P, Shukla RK, Choudhury T, Tamosaitiene J. Integrated
[13] Lootsma FA. Fuzzy logic for planning and decision making. vol. 8, Springer
Fuzzy AHP-Codas Framework for Maintenance Decision in Urea Fertilizer
Science & Business Media; 2013.
Industry. Econ Comput Econ Cybern Stud Res 2017;51(3).
[14] Chen S-M. Fuzzy system reliability analysis using fuzzy number arithmetic
[40] Dhiman HS, Deb D. Fuzzy TOPSIS and fuzzy COPRAS based multi-criteria
operations. Fuzzy Sets and Systems 1994;64(1):31–8.
decision making for hybrid wind farms. Energy 2020;202:117755.
[15] Kumar R, Khepar J, Yadav K, Kareri E, Alotaibi SD, Viriyasitavat W, et al.
[41] Zindani D, Maity SR, Bhowmik S. Fuzzy-EDAS (evaluation based on distance
A Systematic Review on Generalized Fuzzy Numbers and Its Applications:
from average solution) for material selection problems. In: Advances in
Past, Present and Future. Arch Comput Methods Eng 2022;1–24.
computational methods in manufacturing. Springer; 2019, p. 755–71.
[16] Chen F, Chen Y, Zhou J, Liu Y. Optimizing h value for fuzzy linear
[42] Bozanic D, Tešić D, Milićević J. A hybrid fuzzy AHP-MABAC model:
regression with asymmetric triangular fuzzy coefficients. Eng Appl Artif
Application in the Serbian Army–The selection of the location for deep
Intell 2016;47:16–24.
wading as a technique of crossing the river by tanks. Decis Mak Appl
[17] Pelissari R, Oliveira MC, Abackerli AJ, Ben-Amor S, Assumpção MRP. Tech-
Manag Eng 2018;1(1):143–64.
niques to model uncertain input data of multi-criteria decision-making
[43] Boral S, Howard I, Chaturvedi SK, McKee K, Naikan VNA. An integrated
problems: a literature review. Int Trans Oper Res 2021;28(2):523–59.
approach for fuzzy failure modes and effects analysis using fuzzy AHP and
[18] Yatsalo B, Radaev A, Martínez L. From MCDA to fuzzy MCDA: Presump-
fuzzy MAIRCA. Eng Fail Anal 2020;108:104195.
tion of model adequacy or is every fuzzification of an mCDA method
[44] Karande P, Chakraborty S. A Fuzzy-MOORA approach for ERP system
justified? Inform Sci 2022;587:371–92.
selection. Decis Sci Lett 2012;1(1):11–21.
[19] Dhanalakshmi CS, Mathew M, Madhu P. Biomass material selection for
[45] Ulutas A. Supplier selection by using a fuzzy integrated model for a textile
sustainable environment by the application of multi-objective optimiza-
company. Eng Econ 2019;30(5):579–90.
tion on the basis of ratio analysis (MOORA). In: Materials, design, and
[46] Chen C-T. Extensions of the TOPSIS for group decision-making under fuzzy
manufacturing for sustainable environment. Springer; 2021, p. 345–54.
environment. Fuzzy Sets and Systems 2000;114(1):1–9.
[20] Babashamsi P, Golzadfar A, Yusoff NIM, Ceylan H, Nor NGM. Integrated
[47] Opricovic S. A fuzzy compromise solution for multicriteria problems. Int J
fuzzy analytic hierarchy process and VIKOR method in the prioritiza-
Uncertain Fuzziness Knowl-Based Syst 2007;15(03):363–80.
tion of pavement maintenance activities. Int J Pavement Res Technol
[48] Narang M, Joshi M, Pal A. A hybrid fuzzy COPRAS-base-criterion method
2016;9(2):112–20.
for multi-criteria decision making. Soft Comput 2021;25(13):8391–9.
[21] Zyoud SH, Fuchs-Hanusch D. A bibliometric-based survey on AHP and
TOPSIS techniques. Expert Syst Appl 2017;78:158–81.

You might also like