INTERNATIONAL UNIVERSITY – VNU HCMC

SCHOOL OF INDUSTRIAL ENGINEERING AND MANAGEMENT

MULTI-CRITERIA DECISION MAKING MODEL

(MCDM) TO EVALUATE AND SELECT
ELECTRONIC CONSTITUENT SUPPLIERS
A CASE STUDY: CMC CORPORATION

Lecturer: Dr. Dao Vu Truong Son

Course: Capstone Design
Group Number: 35

Student ID Member name % Contribution

IELSIU21284 Huynh Tran Tra Giang 100%

IELSIU21312 Hong Minh Ky 100%

IELSIU20408 Vo Minh Van Quynh 100%

Ho Chi Minh City

June/2024
1
 ABSTRACT
Selecting suppliers for electronic constituents is a critical and complex task influenced
by various uncertain and imprecise criteria. This study introduces a novel integrated approach
that combines Pythagorean fuzzy numbers with multiple Multi-Criteria Decision Making
(MCDM) methods to assist a manufacturer in Hanoi in selecting the most suitable electronic
constituent suppliers. The proposed methodology involves experts filling out forms regarding
the suppliers' locations, with the responses transformed into Pythagorean Fuzzy Numbers
(PFNs) for subsequent analysis. By leveraging Pythagorean fuzzy numbers, this integrated
approach effectively handles the fuzziness inherent in supplier evaluation, ensuring robust and
reliable supplier selection. The evaluation and selection process employs three independent
hybrid methods: AHP-TOPSIS, DEMATEL-VIKOR, and SWARA-PROMETHEE. The
study's findings reveal that all three methods consistently recommend the same suppliers,
demonstrating robustness and reliability for the initial selection and future changes. This
approach ensures that the selected suppliers meet the manufacturer's criteria effectively,
providing a comprehensive and adaptable solution for supplier evaluation and selection in
dynamic environments.

Keywords: Pythagorean fuzzy number, AHP-TOPSIS, DEMATEL-VIKOR SWARA-

PROMETHEE, etronic constituents supplier selection

2
 ACKNOWLEDGEMENT
Dr. Dao Vu Truong Son assisted and supported Group 35 as they completed the
project. He spent much time watching us, offering advice, and helping us when we
encountered problems. Group 35 wants to express our beloved speaker's genuine and heartfelt
appreciation for his priceless advice. In addition, we appreciate Dr. Truong Son for allowing
us to collaborate, learn, and do research. Since then, we have learned a lot that will be helpful
to us in the future with our applications.

We are grateful for the excitement of our friends. Group 35 could only finish this
project with their cooperation and accountability. They took the time to rehearse together
extensively and offered insightful remarks. We performed a fantastic project and a powerful
presentation as a result. Ultimately, even for a short time, it is a tremendous honour for us to
work together. We learn from our errors and use those lessons to improve how well we finish
this assignment.

3
 TABLE OF CONTENTS
ABSTRACT .............................................................................................................................. 2

ACKNOWLEDGEMENT ....................................................................................................... 3

LIST OF TABLES .................................................................................................................... 7

CHAPTER 1: INTRODUCTION ........................................................................................... 8

1.1. MOTIVATION ................................................................................................................ 8

1.2. PROBLEM STATEMENT .............................................................................................. 9

1.3. OBJECTIVES AND REQUIREMENTS ........................................................................ 9

1.3.1. OBJECTIVES ........................................................................................................... 9

1.3.2. REQUIREMENTS .................................................................................................. 10

1.4. SCOPE AND LIMITATIONS ....................................................................................... 10

1.4.1. SCOPE .................................................................................................................... 10

1.4.2. LIMITATIONS ....................................................................................................... 10

1.5. PROJECT PLAN ........................................................................................................... 11

CHAPTER 2: DESIGN CONCEPTS ................................................................................... 12

2.1. OVERVIEW .................................................................................................................. 12

2.2. CURRENT SYTSEM INVESTIGATION: ................................................................... 12

2.2.1. CURRENT SYSTEM ............................................................................................. 12

2.2.2. ADVANTAGES AND DISADVANTAGES OF THE CURRENT SYSTEM ...... 13

2.3. LITERATURE REVIEW .............................................................................................. 14

2.3.1. MULTI-CRITERIA DECISION MAKING METHOD ......................................... 14

2.3.2. FUZZY DECISION MAKING METHOD ............................................................. 14

2.3.3. PYTHAGOREAN FUZZY SETS ........................................................................... 15

2.3.4. TOPSIS AND FUZZY TOPSIS METHODOLOGY ............................................. 15

2.3.5. AHP AND FUZZY AHP METHODOLOGY ........................................................ 16

2.3.6. DEMATEL AND FUZZY DEMATEL METHODOLOGY .................................. 16

2.3.7. VIKOR AND FUZZY VIKOR METHODOLOGY ............................................... 17

2.3.8. SWARA AND FUZZY SWARA METHODOLOGY ........................................... 17

4
 2.3.9. PROMETHEE AND FUZZY PROMETHEE METHODOLOGY ........................ 18

2.4. CONSIDERED SOLUTION DIRECTIONS ................................................................ 19

2.4.1. PYTHAGOREAN FUZZY TOPSIS METHODOLOGY....................................... 19

2.4.2. PYTHAGOREAN FUZZY AHP METHODOLOGY ............................................ 19

2.4.3. PYTHAGOREAN FUZZY DEMATEL METHODOLOGY ................................. 19

2.4.4. PYTHAGOREAN FUZZY VIKOR METHODOLOGY ....................................... 20

2.4.5. PYTHAGOREAN FUZZY SWARA METHODOLOGY ..................................... 20

2.4.6. PYTHAGOREAN FUZZY PROMETHEE METHODOLOGY ............................ 20

2.4.7. PYTHAGOREAN FUZZY LINGUISTIC SCALE USED IN PFTOPSIS,

PFVIKOR, PFSWARA, PFPROMETHEE ...................................................................... 21

2.4.8. PYTHAGOREAN FUZZY LINGUISTIC SCALE USED IN PFAHP,

PFDEMATEL ................................................................................................................... 21

2.4.9. SUPPLIERS EVALUATION CRITERIA .............................................................. 22

CHAPTER 3: SYSTEM DESIGN ........................................................................................ 25

3.1. APPROACHES COMPARISON AND SELECTION .................................................. 25

3.2. SYSTEM DESIGN DESCRIPTION ............................................................................. 27

3.2.1. DESIGN STRUCTURE .......................................................................................... 27

3.2.2. KEY ADVANTAGES OF THE PROPOSED SYSTEM ....................................... 35

CHAPTER 4: PROTOTYPE DEVELOPMENT AND IMPLEMENTATION ............... 36

4.1. PROTOTYPE DEVELOPMENT .................................................................................. 36

4.2. RESULTS ...................................................................................................................... 36

4.3. ANALYSIS .................................................................................................................... 38

4.3.1. SENSITIVITY ANALYSIS .................................................................................... 38

4.3.2. FURTHER ANALYSIS .......................................................................................... 40

CHAPTER 5: CONCLUSION .............................................................................................. 41

5.1. SUMMARY ................................................................................................................... 41

5.2. RECOMMENDATION ................................................................................................. 41

REFERENCES ....................................................................................................................... 42

APPENDIX ............................................................................................................................. 45
5
A1. SURVEY FORM ........................................................................................................... 45

A2. MATLAB CODE ........................................................................................................... 47

A2.1. FUZZY AHP + TOPSIS ......................................................................................... 47

A2.2. FUZZY DEMATEL + VIKOR ............................................................................... 54

A2.3. FUZZY SWARA + PROMETHEE ........................................................................ 61

6
 LIST OF TABLES
Table 1. Timeline project.......................................................................................................... 11

Table 2. Pythagorean fuzzy linguistic scale used in PFTOPSIS, PFVIKOR, PFSWARA,

PFPROMETHEE ...................................................................................................................... 21

Table 3. Weighting scale for PAHP, PDEMATEL (Ilbahar et al. 2018) ................................. 22

Table 4. Supplier evaluation criteria......................................................................................... 23

Table 5. Advantages and Disadvantages of methods ............................................................... 25

Table 6. Weight of each criterion found by three methods ...................................................... 36

Table 7. Distance and rank by AHP + TOPSIS ........................................................................ 37

Table 8. R, S, Q and rank by DEMATEL+ VIKOR ................................................................ 37

Table [Link] and rank by SWARA + PROMETHEE ................................................................ 37

Table 10. Sensitivity analysis ................................................................................................... 38

7
 CHAPTER 1: INTRODUCTION
1.1. MOTIVATION
Vietnam is witnessing rapid growth in its electronics industry, driven by rising
domestic demand, expanding export markets, and supportive government initiatives. This
thriving sector operates within a complex network of manufacturers and suppliers. A critical
challenge emerges amid this growth: selecting the most suitable suppliers for the diverse
electronic components required.

The foundation of electronic device manufacturing depends on the quality of its

components. Substandard parts can jeopardize product integrity and harm the manufacturer's
reputation. Thus, manufacturers need a robust supplier evaluation process beyond price to
include stringent quality assurance. Reliability is also crucial. Delays in component deliveries
can disrupt production, incur penalties, and erode customer trust. Reliable suppliers ensure
smooth production and timely delivery. After-sales service is a key differentiator. Adequate
technical support addresses customer concerns and fosters loyalty. Export-oriented
manufacturers require suppliers with competent and accessible after-sales teams. Ethical
sourcing has become a core value. Unethical labour practices and social responsibility
violations can tarnish a brand's reputation. Manufacturers now seek suppliers who are
committed to ethical sourcing practices.

In this dynamic environment, a robust supplier evaluation process is indispensable.

Manufacturers must adopt methodologies beyond traditional price-centric approaches,
integrating factors like stringent quality assurance, reliable delivery schedules, and ethical
sourcing standards. This approach enables them to reduce the risks linked to inferior
components, production hold-ups, and damage to their reputation from unethical practices.

Therefore, manufacturers must make informed choices when selecting suitable

suppliers, considering price, quality, and associated services. This comprehensive approach
ensures they grasp the complex factors influencing their decisions, such as reliability, after-
sales support, and ethical sourcing practices. Making well-informed, multi-criteria-based
decisions is essential for effectively navigating the intricate supplier landscape. Furthermore,
forming strategic alliances with suppliers is essential for staying competitive and ensuring the
sustainability of operations in Vietnam's swiftly advancing electronics sector.

8
1.2. PROBLEM STATEMENT
Hanoi's growing electronics industry faces the critical challenge of selecting reliable
suppliers for electronic components. Traditional methods focused primarily on cost fail to
address the complexities of supplier evaluation, such as quality assurance, reliability, after-
sales service, and ethical sourcing.

The proposed study addresses this need by developing an integrated Multi-Criteria

Decision Making (MCDM) model tailored for evaluating and selecting electronic component
suppliers in Hanoi. By leveraging advanced methodologies such as the Analytic Hierarchy
Process (AHP), Technique for Order of Preference by Similarity to Ideal Solution (TOPSIS),
Decision-Making Trial and Evaluation Laboratory (DEMATEL), and VIseKriterijumska
Optimizacija I Kompromisno Resenje (VIKOR), Stepwise Weight Assessment Ratio Analysis
(SWARA), and Preference Ranking Organization Method for Enrichment Evaluation
(PROMETHEE), the study seeks to enhance supplier selection processes. This approach will
help Hanoi's electronics manufacturers mitigate risks associated with substandard
components, production delays, and reputational damage while fostering strategic
partnerships contributing to the industry's competitive edge and sustainability.

1.3. OBJECTIVES AND REQUIREMENTS

1.3.1. OBJECTIVES
The main objective of this study is to develop a comprehensive Multi-Criteria
Decision Making (MCDM) model to evaluate and select electronic constituent suppliers in
Hanoi. This model aims to integrate various decision-making techniques, such as AHP,
TOPSIS, DEMATEL, VIKOR, SWARA, and PROMETHEE to provide a structured
framework for supplier evaluation. By broadening the evaluation criteria to include quality
assurance, reliability, after-sales service, and ethical sourcing, the model seeks to ensure a
thorough and holistic assessment of suppliers. It aims to improve the overall decision-making
process, mitigate risks associated with substandard components, production delays, and
unethical practices, and support establishing strategic partnerships with reliable suppliers.
Ultimately, this model will contribute to Hanoi's electronics industry's sustainable growth and
competitiveness by promoting high-quality production standards and ethical sourcing
practices.

9
1.3.2. REQUIREMENTS
Several key components include the requirements for developing a Multi-Criteria
Decision Making (MCDM) model to evaluate and select electronic constituent suppliers in
Hanoi.

Firstly, it necessitates a thorough understanding of Hanoi's electronics industry,

encompassing existing supplier evaluation practices and the particular challenges encountered
by manufacturers. Detailed knowledge of MCDM techniques such as the AHP, TOPSIS,
DEMATEL, VIKOR, SWARA, and PROMETHEE is essential for integrating these methods
effectively.

Secondly, collecting and analysing relevant data on supplier performance across

multiple criteria, including quality assurance, reliability, after-sales service, and ethical
sourcing, are crucial. Collaboration with industry stakeholders is also necessary to validate the
model and ensure its practical applicability.

Finally, a robust framework for implementing the MCDM model in real-world

supplier selection processes must be established, ensuring that it can be used to make
informed and strategic supplier choices that align with the goals of Hanoi's electronics
manufacturers.

1.4. SCOPE AND LIMITATIONS

1.4.1. SCOPE
Our research focuses on evaluating the impact of quality, services, reliability, and
sustainability criteria on the selection of semiconductor material suppliers within the Hanoi -
Thai Nguyen - Bac Ninh - Bac Giang region.

1.4.2. LIMITATIONS
While the hybrid MCDM approach offers a structured framework for supplier
evaluation, these limitations highlight the need for ongoing data collection, model refinement,
and adaptability to industry changes.

Firstly, data availability poses a significant challenge. Determining suitable weights

from expert opinions can be difficult due to variations in expertise and potential biases. This
can affect the accuracy and reliability of the model.

10
 Secondly, model uncertainty is inherent in the application of MCDM methods. There
is no single best MCDM model, meaning the solutions provided may need to be more robust
and conclusive. Different models may yield varying results, which can complicate the
decision-making process.

Thirdly, inconsistency is a notable limitation of MCDM that can undermine the

reliability and validity of decision-making outcomes. Consistency measures, sensitivity
analysis, group decision-making, and proper training are essential tools in mitigating the
impact of inconsistency in MCDM processes.

Lastly, future recommendations are constrained by the static nature of the data used.
The data and criteria applied in this study are primarily unchangeable and are only valid for a
limited period. This temporal limitation makes it challenging to make reliable
recommendations about future industry trends or long-term supplier suitability.

1.5. PROJECT PLAN

Table 1. Timeline project

No. Start Date End Date Task

Submit the chosen topic for approval by the

1 08/01/2024 22/01/2024
supervising professor.

2 22/01/2024 24/02/2024 Problem Identification

3 24/02/2024 08/03/2024 Design Concepts Consideration

4 08/03/2024 08/04/2024 System Design

5 08/04/2024 08/05/2024 Prototype Development and Implementation

6 08/05/2024 28/05/2024 Results Analysis

7 28/05/2024 09/06/2024 Report Writing

Engage in a consultative review of the final report

8 09/06/2024 19/06/2024
with your thesis advisor.
Refine and rehearse the presentation for the
9 22/06/2024 05/07/2024
dissertation proposal defence.

11
 CHAPTER 2: DESIGN CONCEPTS
2.1. OVERVIEW
In the ever-evolving world of supply chain management, selecting the most suitable
supplier stands as a cornerstone decision (Kumari, M., Kanti De, P., Narang, P., & Shah, N.
H. 2023). A strategic supplier selection process can significantly reduce a company's bottom
line by minimizing material acquisition costs (Ecer 2022). However, the impact goes beyond
just cost savings. The right supplier can also enhance product quality and flexibility, leading
to a faster and more efficient purchasing process. Recognizing the growing importance of
environmental responsibility, supply chain management embraces a more sustainable
approach (Cheraghalipour, A., Paydar, M. M., & Hajiaghaei-Keshtelia, M. 2017).

The topic of "Multi-Criteria Decision Making Model (MCDM) to Evaluate and Select
Electronic Constituent Suppliers - A Case Study: CMC Corporation" addresses the critical
need for a sophisticated supplier selection process within Hanoi's rapidly evolving electronics
industry. We conducted our research across three regions: Hanoi, Thai Nguyen, and Bac
Ninh-Bac Giang. These areas are collectively known as the "Northern Economic
Development Triangle," characterized by substantial investments in high-quality industrial
zones and significant semiconductor component suppliers such as Intel, Samsung, and
Amkor...

This study proposes the use of an integrated MCDM approach, combining various
decision-making methods such as AHP, TOPSIS, DEMATEL, VIKOR, SWARA, and
PROMETHEE. These methods collectively facilitate a more comprehensive and nuanced
evaluation of potential suppliers, addressing factors such as quality assurance, delivery
reliability, after-sales service, and ethical sourcing. By adopting this advanced MCDM
framework, CMC Corporation Hanoi can make more informed and strategic decisions,
ultimately enhancing their competitiveness and sustainability in the global electronics market.

2.2. CURRENT SYTSEM INVESTIGATION:

2.2.1. CURRENT SYSTEM
The current system evaluates suppliers primarily based on the competitiveness of their
electronic component pricing and adherence to delivery schedules to ensure timely production
processes. While quality is considered, the system needs a comprehensive approach to
assessing suppliers' quality assurance practices. Suppliers must meet essential legal and
12
regulatory standards, but assessing ethical sourcing and social responsibility is limited.
Additionally, there is minimal emphasis on suppliers' capabilities in providing after-sales
service to address post-purchase issues and maintain customer satisfaction.

The current system operates on a straightforward framework where suppliers submit

their proposals, which are evaluated primarily based on cost and delivery performance. The
evaluation process involves reviewing submitted pricing and comparing it against market
standards, monitoring suppliers' delivery timelines to ensure they meet agreed schedules,
conducting basic quality checks on received components to ensure they meet minimum
standards, and ensuring suppliers comply with basic legal and regulatory requirements.

2.2.2. ADVANTAGES AND DISADVANTAGES OF THE CURRENT

SYSTEM
Advantages: The current system offers several advantages. Its simplicity makes it
easy to understand and implement, ensuring a straightforward supplier evaluation process. By
emphasizing pricing, the system helps control costs, which is critical in competitive markets.
Additionally, regular tracking of delivery performance ensures that production schedules are
maintained, reducing the risk of delays and ensuring timely production processes.

Disadvantages:
- Limited Quality Assessment: The system must thoroughly evaluate suppliers'
quality assurance practices, potentially leading to substandard components.
- Inadequate Ethical Considerations: There needs to be more emphasis on ethical
sourcing and social responsibility, which are increasingly important in the global
market.
- Minimal After-Sales Service Evaluation: Limited focus on after-sales service
capabilities can lead to unresolved post-purchase issues, affecting customer
satisfaction and brand loyalty.
- Narrow Criteria: The current system focuses heavily on price and delivery,
missing out on other critical factors such as reliability, quality, and service, which
are essential for long-term success.

13
2.3. LITERATURE REVIEW
2.3.1. MULTI-CRITERIA DECISION MAKING METHOD
Choosing the right supplier is crucial in supply chain management (SCM) because it
impacts a company's ability to avoid disruptions and shortages. It is a complex task due to the
many factors and options involved.

This method often involves a Multi-Criteria Decision-Making Method (MCDM) to

determine the importance of different supplier qualities (e.g., price, quality, environmental
impact). The company's decision-makers then use qualitative and quantitative data to assess
potential suppliers based on these criteria. Ultimately, company experts define the criteria for
identifying "ideal" suppliers.

These ideal suppliers can provide good quality products at fair prices, adapt to
changing demand quickly, and maintain safety and environmental responsibility. Defining the
criteria for such suppliers, especially in uncertain times, is critical).

2.3.2. FUZZY DECISION MAKING METHOD

Fuzzy decision-making (FDM) emerges as a robust approach for navigating scenarios
characterized by uncertainty, incompleteness, or ambiguity in the underlying information
(Pınar, A., Erdebilli, B. D. R. B., & ¨Ozdemir, Y. S., 2021). It represents a significant
departure from traditional decision-making paradigms that rely on crisp, binary logic. In
FDM, fuzzy sets, as opposed to crisp sets, are employed to characterize both input and output
variables. These fuzzy sets enable the representation of partial membership, allowing a
variable to possess a degree of membership between 0 and 1 rather than being strictly
classified as belonging to or excluded from a set entirely (Ayyildiz, E., & Taskin Gumus, A.,
2021). This framework inherently facilitates incorporating various levels of uncertainty and
imprecision within the decision-making process. The FDM process itself typically
encompasses three key stages: fuzzification, rule evaluation, and defuzzification (Boix-Cots,
D., Pardo-Bosch, F., & Pujadas, P., 2023). The culmination of this process is a membership
function, which assigns a degree of membership to each potential decision alternative. By
virtue of its versatility, FDM has garnered widespread application across diverse fields,
including finance, engineering, and control systems.

14
2.3.3. PYTHAGOREAN FUZZY SETS
Intuitionistic fuzzy sets (IFSs), introduced by Atanassov (1986), have gained traction
in various research fields as a tool to manage uncertainty. Membership functions, non-
membership functions, and hesitancy degrees characterize these sets. However, IFSs exhibit
limitations in specific scenarios where the sum of membership and non-membership grades
surpass unity. Consequently, these sets need help to capture such situations effectively
(Yager, 2014).

To address this shortcoming, (Yager, 2014) proposed Pythagorean fuzzy sets (PFSs).
PFSs leverage the Pythagorean theorem to define set membership. The membership degree of
an element in a PFS is determined by its distance from the set's centre, calculated using the
Pythagorean theorem (Büyük¨ozkan, G., & G¨oçer, F., 2021). PFSs generalize IFSs, offering
a solution in situations where IFSs fall short in uncertainty representation. PFSs possess
several advantages over traditional fuzzy sets. Notably, they provide a geometric
interpretation and facilitate a more natural representation of uncertain information (Yu, C.,
Shao, Y., Wang, K., & Zhang, L., 2019). The applicability of PFSs extends to various
domains, including pattern recognition, image processing, decision-making, and control
systems (Han, Q., Li, W., Xu, Q., Song, Y., Fan, C., & Zhao, M., 2022). Their ability to
handle uncertainty beyond the limitations of IFSs renders PFSs a more powerful and flexible
tool (Ak, M. F., & Gul, M., 2019; Ilbahar, E., Karas¸an, A., Cebi, S., & Kahraman, C., 2018).
It is crucial to note that, unlike IFSs, PFSs permit the sum of membership and non-
membership degrees to exceed 1, with the constraint that the sum of their squares remains less
than or equal to 1 (Ilbahar, E., Karas¸an, A., Cebi, S., & Kahraman, C., 2018; Zhang, X., &
Xu, Z., 2014; Zeng S, 2016).

2.3.4. TOPSIS AND FUZZY TOPSIS METHODOLOGY

The Technique for Order Preference by Similarity to the Ideal Solution (TOPSIS), a
prominent Multi-Criteria Decision-Making (MCDM) method categorized under multi-quality
decision-making approaches, was introduced by Hwang and Yoon in 1981 (Hwang et al.; K.,
1981). The core principle of the TOPSIS algorithm lies in identifying the alternative closest to
a hypothetical "positive ideal solution" and the alternative furthest from a hypothetical
"negative ideal solution" (Jin, 2023).

The Fuzzy TOPSIS method extends the traditional TOPSIS by incorporating fuzzy
logic, which addresses the uncertainty and vagueness inherent in decision-making processes.

15
Introduced by Chen in 2000, Fuzzy TOPSIS applies fuzzy set theory to represent uncertain
data more effectively. This method follows steps similar to TOPSIS but incorporates fuzzy
numbers, usually triangular or trapezoidal fuzzy numbers, to handle imprecise information.

2.3.5. AHP AND FUZZY AHP METHODOLOGY

The Analytic Hierarchy Process (AHP), developed by Thomas L. Saaty in the 1970s,
is a renowned method for structured decision-making involving multiple criteria (Saaty,
1990). It decomposes complex problems into a hierarchical structure of goals, criteria, and
alternatives (Ishizaka, 2011). AHP integrates qualitative and quantitative aspects through
pairwise comparisons, assessing the relative importance of criteria and alternatives. These
comparisons are quantified to calculate weights that reflect each criterion's significance (Saaty
& Vargas, 2012). Alternatives are evaluated based on these criteria, and the consistency of
comparisons ensures reliable results (Ishizaka, 2011). Finally, criteria weights and alternative
performance scores are synthesized to determine an overall priority ranking. AHP's structured
approach and ability to handle complex multi-criteria decisions make it valuable across
various fields.

To address this limitation, the Fuzzy Analytic Hierarchy Process (F-AHP)

methodology emerges as a powerful extension. F-AHP builds on AHP's core principles but
incorporates fuzzy set theory to handle decision-making vagueness (Wang, 2016). This theory
allows for partial membership in sets, offering flexibility. In F-AHP, pairwise comparisons
use fuzzy linguistic terms like "slightly more important" instead of precise numerical values.
These terms convert into fuzzy numbers, representing the uncertainty in comparisons. Fuzzy
logic techniques then perform calculations, including weight derivation and consistency
checks, resulting in a ranking of alternatives that accounts for decision-making fuzziness.
Consequently, F-AHP provides a more realistic and robust decision-making approach in
complex and uncertain environments.

2.3.6. DEMATEL AND FUZZY DEMATEL METHODOLOGY

The Decision-Making Trial and Evaluation Laboratory (DEMATEL) method,
developed by the Battelle Memorial Institute, is widely used for analyzing and visualizing
complex causal relationships among factors in decision-making processes (Fontela & Gabus,
1976). It is beneficial when understanding the interdependencies and influence among criteria
is crucial. DEMATEL transforms these interdependencies into a structured model, typically
represented by a directed graph, where nodes represent criteria and directed edges indicate the

16
influence strength between them. This method helps decision-makers identify critical factors
and understand the feedback and causality in complex systems (Tzeng, 2007).

To improve DEMATEL's ability to manage uncertainty and imprecision, the Fuzzy

DEMATEL methodology was developed. This method incorporates fuzzy set theory, which is
ideal for representing the ambiguous and uncertain information often encountered in real-
world decision-making scenarios (Tseng, 2009). Decision-makers use fuzzy linguistic terms
to describe the degree of influence between criteria. These terms are then converted into fuzzy
numbers and used to create a fuzzy influence matrix. This matrix undergoes fuzzy arithmetic
operations to produce a fuzzy total influence matrix, capturing the overall impact and
interdependencies among criteria with greater accuracy.

2.3.7. VIKOR AND FUZZY VIKOR METHODOLOGY

The VIKOR methodology, developed by Opricovic and Tzeng in 2004, is a multi-
criteria decision-making (MCDM) approach designed to rank and select from a set of
alternatives while considering conflicting criteria. VIKOR emphasizes finding a compromise
solution closest to the ideal and anti-ideal solutions, which is particularly useful when
decision-makers cannot agree on the relative importance of different criteria (Opricovic &
Tzeng, 2004). It involves constructing a performance matrix, calculating each alternative's
utility and regret measures, and determining a compromise ranking by combining these
measures.

To enhance VIKOR's ability to handle the inherent uncertainty and vagueness in

decision-making, the Fuzzy VIKOR methodology was introduced. This method incorporates
fuzzy set theory to better represent uncertain and imprecise information (Opricovic & Tzeng,
2006). In Fuzzy VIKOR, decision-makers express their preferences using fuzzy linguistic
terms, which are then converted into fuzzy numbers. These fuzzy numbers are used to
construct a fuzzy performance matrix, where the utility and regret measures are calculated
using fuzzy arithmetic operations. The result is a fuzzy compromise ranking that accounts for
the uncertainty and imprecision of the decision-makers' judgments, providing a more realistic
and robust decision-making process (Opricovic & Tzeng, 2006; Chen & Tzeng, 2004).

2.3.8. SWARA AND FUZZY SWARA METHODOLOGY

The Stepwise Weight Assessment Ratio Analysis (SWARA) methodology, developed
by Kersuliene et al. in 2010, is an MCDM technique used to determine the relative importance
of various criteria in decision-making. SWARA relies on a structured approach where experts
17
assess the importance of each criterion step-by-step, allowing for a systematic derivation of
weights (Kersuliene et al., 2010). The process begins by listing criteria in descending order of
importance, followed by experts evaluating the relative significance of each criterion
compared to the previous one. This evaluation results in the calculation of a ratio reflecting
the significance of each criterion, which is then used to determine the final weights.

The Fuzzy SWARA methodology was introduced to address the limitations of

SWARA in handling uncertainty and imprecision. Fuzzy SWARA integrates fuzzy set theory
to represent the vagueness inherent in expert judgments (Zavadskas et al., 2014). In Fuzzy
SWARA, experts use fuzzy linguistic terms, such as "very important" or "slightly less
important," to express the relative importance of criteria. These linguistic terms are
subsequently transformed into fuzzy numbers, which are utilized to carry out fuzzy arithmetic
operations during the SWARA process. The outcome is a set of fuzzy weights that better
capture the uncertainty and imprecision of the experts' assessments, resulting in a more robust
and realistic evaluation of criteria importance (Zavadskas et al., 2014).

2.3.9. PROMETHEE AND FUZZY PROMETHEE METHODOLOGY

The PROMETHEE (Preference et al. Method for Enrichment Evaluation)
methodology, developed by Brans and Vincke in the 1980s, is a prominent Multi-Criteria
Decision Making (MCDM) tool used to evaluate and rank a set of alternatives based on
multiple criteria (Brans & Vincke, 1985). PROMETHEE I provides a partial ranking of other
options, indicating which alternatives are better or worse without necessarily ranking all of
them. PROMETHEE II, on the other hand, offers a complete ranking by computing net
outranking flows, which are the difference between positive and negative outranking flows for
each alternative (Brans et al., 1986). The methodology involves several steps: defining criteria
and alternatives, determining weights for each criterion, and calculating preference functions
for pairwise comparisons between alternatives. The preference functions express the degree of
preference for one alternative over another on a particular criterion. The results from these
comparisons are aggregated to produce the outranking flows, which are then used to rank the
other options.

Fuzzy PROMETHEE integrates fuzzy set theory into the standard PROMETHEE
approach to better manage the vagueness inherent in decision-making processes (Ruan, 2007).
In Fuzzy PROMETHEE, decision-makers express their preferences using fuzzy linguistic
terms, such as "slightly preferred" or "very preferred," instead of exact numerical values.
These fuzzy linguistic terms are converted into fuzzy numbers, allowing for a more nuanced
18
representation of uncertainty in the preference functions. Fuzzy arithmetic operations are then
applied to the fuzzy preference values to compute fuzzy outranking flows. The resulting fuzzy
net flows provide a more robust and realistic ranking of alternatives by capturing the
ambiguity and imprecision of real-world decision-making scenarios. By incorporating fuzzy
set theory, Fuzzy PROMETHEE enhances the traditional PROMETHEE methodology's
ability to deal with complex and uncertain environments, offering a more comprehensive
decision-support tool for multi-criteria evaluations (Mareschal & Ruan, 2007).

2.4. CONSIDERED SOLUTION DIRECTIONS

2.4.1. PYTHAGOREAN FUZZY TOPSIS METHODOLOGY
The Pythagorean fuzzy TOPSIS (PF-TOPSIS) method emerges as a prominent Multi-
Criteria Decision-Making (MCDM) technique. It leverages the concept of Pythagorean fuzzy
sets to identify the optimal solution within a set of alternatives (Cevik Onar, S., Oztaysi, B., &
Kahraman, C., 2018). This method aligns with the core principle of TOPSIS, where the
chosen alternative exhibits the shortest distance from a hypothetical "positive ideal solution"
and the most significant distance from a hypothetical "negative ideal solution." By
incorporating Pythagorean fuzzy sets, PF-TOPSIS facilitates the representation of uncertainty
inherent in decision-making processes.

2.4.2. PYTHAGOREAN FUZZY AHP METHODOLOGY

Intuitionistic fuzzy sets (IFSs), introduced by (Atanassov, 1986), have served as a
foundation for various extensions, including neutrosophic sets (Smarandache, 1995),
Pythagorean fuzzy sets (Yager R. A., 2013), and orthopaedic fuzzy sets (Yager R. A., 2017).
These extensions aim to better capture and manage inherent uncertainties within decision-
making environments. This study focuses on Pythagorean fuzzy sets due to their ability to
address vagueness and imprecision. We will explore the core principles and basic operations
of Pythagorean fuzzy sets, followed by a detailed examination of the Pythagorean fuzzy AHP
methodology.

2.4.3. PYTHAGOREAN FUZZY DEMATEL METHODOLOGY

To enhance DEMATEL's capability in handling such uncertainty, the Pythagorean
Fuzzy DEMATEL methodology was introduced. This approach integrates Pythagorean fuzzy
sets (PFS) into the traditional DEMATEL framework to better represent and manage
ambiguous information (Zadeh, 1965). PFS extend the concept of fuzzy sets by allowing the
sum of the squares of membership and non-membership degrees to be less than or equal to
19
one, offering greater flexibility and precision in expressing uncertainty (Yager, 2014). In the
Pythagorean Fuzzy DEMATEL methodology, decision-makers use Pythagorean fuzzy
numbers to express the degree of influence among criteria. These fuzzy numbers are then used
to construct a fuzzy direct-relation matrix, capturing the initial influence relationships more
accurately. Fuzzy arithmetic operations are applied to this matrix to derive a fuzzy total-
relation matrix, which reveals the overall impact and interdependencies among criteria.

2.4.4. PYTHAGOREAN FUZZY VIKOR METHODOLOGY

VIKOR is particularly useful for decision-making problems where conflicting criteria
must be balanced to achieve the best overall outcome. To enhance VIKOR's ability to handle
uncertainty, the Pythagorean Fuzzy VIKOR methodology incorporates Pythagorean fuzzy sets
(PFS). PFS extend the traditional fuzzy set theory by allowing the sum of the squares of
membership and non-membership degrees to be less than or equal to one, providing a more
flexible representation of uncertainty (Yager, 2014). In the Pythagorean Fuzzy VIKOR
methodology, decision-makers use Pythagorean fuzzy numbers to express the importance of
alternatives' performance ratings and the criteria. This approach captures the inherent
vagueness in subjective assessments more effectively than traditional crisp values.

2.4.5. PYTHAGOREAN FUZZY SWARA METHODOLOGY

Like many traditional decision-making tools, SWARA can need help with the
uncertainty and imprecision inherent in human judgments. The Pythagorean Fuzzy SWARA
methodology integrates Pythagorean fuzzy sets (PFS) into the SWARA framework to address
these limitations. PFS, an extension of fuzzy sets, provide greater flexibility in representing
uncertainty by allowing the sum of the squares of membership and non-membership degrees
to be less than or equal to one (Yager, 2014). This enables a more accurate and nuanced
handling of ambiguity in decision-making. In the Pythagorean Fuzzy SWARA methodology,
decision-makers use Pythagorean fuzzy numbers to express the importance of each criterion
relative to others. These fuzzy numbers capture the inherent vagueness in subjective
assessments more effectively than traditional crisp values.

2.4.6. PYTHAGOREAN FUZZY PROMETHEE METHODOLOGY

The Preference Ranking Organization Method for Enrichment PROMETHEE is
particularly effective in scenarios where decision-makers must evaluate a finite set of
alternatives concerning several, often conflicting, criteria. However, the traditional
PROMETHEE method can be limited by the precision of the data and the clarity of the

20
criteria weights provided by decision-makers. To address these limitations, the Pythagorean
Fuzzy PROMETHEE methodology incorporates Pythagorean fuzzy sets (PFS), which extend
traditional fuzzy set theory by allowing the square sum of the membership and non-
membership degrees to be less than or equal to one. This extension offers a more flexible and
accurate representation of uncertainty and imprecision in decision-making (Yager, 2014).

2.4.7. PYTHAGOREAN FUZZY LINGUISTIC SCALE USED IN

PFTOPSIS, PFVIKOR, PFSWARA, PFPROMETHEE
When evaluating suppliers using the PFTOPSIS, PFVIKOR, PFSWARA,
PFPROMETHEE methodology, criteria assessments are conducted based on the Pythagorean
fuzzy linguistic scale established by (Pe´rez-Domı´nguez L, 2018). This scale defines a set of
linguistic variables that can be represented mathematically using Pythagorean fuzzy numbers.
The scale itself comprises nine points, offering a nuanced representation of decision-maker
judgments (Pe´rez-Domı´nguez L, 2018). Notably, the supplier prioritization process within
this study utilizes a five-member decision panel, and each member employs those above
Pythagorean fuzzy linguistic scale for their evaluations.

Table 2. Pythagorean fuzzy linguistic scale used in PFTOPSIS, PFVIKOR, PFSWARA,

PFPROMETHEE

Linguistic term Corresponding Pythagorean fuzzy number (u,v)

Not Important (0.10, 0.8)

Slight Important (0.25, 0.6)

Moderately Important (0.5, 0.4)

Important (0.75, 0.2)

Very Important (0.9, 0.05)

2.4.8. PYTHAGOREAN FUZZY LINGUISTIC SCALE USED IN PFAHP,

PFDEMATEL
The evaluation of both primary and sub-criteria during supplier selection via pairwise
comparison within the PFAHP, PFDEMATEL leverages the scale established by (Ilbahar, E.,
Karas¸an, A., Cebi, S., & Kahraman, C., 2018). This scale is designed explicitly for interval-
valued Pythagorean fuzzy numbers, as detailed in Table 3.

21
 Table 3. Weighting scale for PAHP, PDEMATEL (Ilbahar et al. 2018)

Linguistic term Corresponding Pythagorean fuzzy number (u,v)

Certainly Low Important (0, 0, 0.90, 1)

Very Low Important (0.10, 0.20, 0.80, 0.90)

Low Important (0.20, 0.35, 0.65, 0.8)

Below Average Important (0.35, 0.45, 0.55, 0.65)

Average Important (0.45, 0.55, 0.45, 0.55)

Above Average Important (0.55, 0.65, 0.35, 0.45)

High Important (0.65, 0.80, 0.20, 0.35)

Very High Important (0.80, 0.90, 0.10, 0.20)

Certainly High Important (0.90, 1, 0, 0)

Exactly Equal (0.1965, 0.1965, 0.1965, 0.1965)

2.4.9. SUPPLIERS EVALUATION CRITERIA

In this article, we consider 5 main factors: cost, quality, reliability, service, and sustainability.
Within each element there will be sub-elements as follows:
A) Cost:
- Raw material prices: selling prices, discounts and incentives from suppliers.
- Shipping costs: shipping costs and policies of the supplier.
- Hidden fees: additional costs such as loading and unloading goods, loss during
transportation...
- Payment terms.
B) Quality:
- Consistency in product specifications: the product must ensure consistency in
technical specifications or necessary requirements.
- Product defect rate: High defect rates can lead to material waste, increased labor
costs due to replacements, and can cause delays.
- Inventory management: suppliers must have appropriate inventory policies to
avoid shortages of goods.
- Factory performance: factory performance, number of finished products...
C) Reliability:
22
 - On-time delivery: the supplier must ensure the receipt and return of orders
according to the agreement.
- Ability to meet demand for goods: can the supplier meet demand for goods during
peak periods? Is there a clear cargo handling process that can handle supply
fluctuations?
- Experience and reliability from other partners: does the supplier receive good
reviews from their partners? What is your experience in providing this item?
D) Service:
- Response rate: Level of willingness to address partners' questions and concerns as
well as cooperate in overcoming arising problems.
- Compensation terms: When risks or uncertainties occur, does the supplier offer
reasonable compensation and loss settlement terms? (technical support, return -
refund...
E) Sustainability:
- Human resources: suppliers use workers in accordance with the law to ensure their
basic rights are met.
- Work safety and labor health: ensure labor safety regulations.
- Financial stability: the supplier has stable finances and development speed.
- Environmental responsibility: waste, water, and dust treatment systems must
comply with environmental regulations.
- Social responsibility: the supplier has positive activities for the locality.

Table 4. Supplier evaluation criteria

Main criteria Sub-criteria

C11 Raw material prices

C12 Shipping costs

C1 Cost
C13 Hidden fees

C14 Payment terms

C21 Consistency

C2 Quality C22 Product defect rates

C23 Inventory management

23
 C24 Factory performance

C31 On-time delivery

C3 Reliability C32 Ability to meet production demands

C33 Experiences

C41 Responsiveness
C4 Service
C42 Warranty terms

C51 Human resources

C52 Work safety & labor health

C5 Sustainability C53 Financial stability

C54 Environmental responsibility

C55 Social responsibility

S1 Industrial Zone Hoa Lac (Ha Noi)

Suppliers S2 Industrial zone Yen Binh (Thai Nguyen)

(Alternatives)
Industrial Zone Yen Phong – Van Trung
S3
(Bac Ninh – Bac Giang)

24
 CHAPTER 3: SYSTEM DESIGN
3.1. APPROACHES COMPARISON AND SELECTION
Based on the previous part, the MCDM can be solved by various Pythagorean Fuzzy
method. In this study, one approach including two types of method, one for calculating the
criteria weight while the other is used to ranking the alternatives.

Six methods that are compared a selected. The characteristics of them are listed in the
table below:

Table 5. Advantages and Disadvantages of methods

Methods Advantages Disadvantages

- Because of simple technique,

- TOPSIS is the best method to find TOPSIS is based on the assumption
the most suitable alternative with of monotonically increasing or
strong confidence. decreasing the criteria that will hide
- Easy to adapt with simple the importance of them.
TOPSIS
calculation and meaning. - Correlation among attributes is not
- Able to make comparisons defined strongly though Euclidean
between alternatives. distance is used.
- Not able to determine the weight of
criteria.
- Ability to solve complex real-life
- Time consuming when dealing with
problem.
big problems due to big number of
- Help the decision maker make the
pairwise comparisons.
comparison between attributes to
AHP - Relied strongly on the data
be more confident with the result.
collected.
- Can be integrated with modern
- Uncertainties are not considered.
theory like data analysis, game
theory,
- Most suitable method in dealing - Not suitable to find the importance
with complex problems that are of relative factors.
DEMATEL
hardly to understand. - Hardly and inefficiency when
- Making decision maker dealing with Fuzzy number so the

25
 understand more about the numbers have to be deffuzyficated.
interdependent among attributes to
find their transformation trends.
- Best used in comparative
strength.
- Able to deal with problems that
- The appearance of S and R make
have various unit of measurement.
VIKOR the acceptable range of solution
- Ranking the alternatives by the
decrease.
ability to satisfy criteria of them.
- Best accurate in analyze the
expert opinion to find the rank of
- Only available with the first type
criteria.
when only provide the initial decision
- A more accurate result when it
matrix.
SWARA provides the link between experts.
- Easily to be wrong with the
- Make the policy for decision
unprofitable criteria.
maker to choose the suitable
- Not show much information.
criteria.

- Can be deal with problem that

- No formal guideline for finding the
contain number of units of
weight of criteria so it needs to be
measure.
combined with the other.
PROMETHEE - Ability analyzes the alternative
- When new alternatives are
that hardly to compare.
considered, PROMETHEE is hardly
- Does not need to repeat the
to solve base on the old data.
pairwise comparison.

With the information listed above, three iterated methods are chosen. That are:

- Pythagorean Fuzzy AHP+ Pythagorean Fuzzy TOPSIS: The hybrid method that
makes the AHP method simpler when only weight criteria that are clustered while the
complex ranking alternative will be deal with the simple techniques of TOPSIS.

- Pythagorean Fuzzy DEMATEL+ Pythagorean Fuzzy VIKOR: the method that is

best fit in the uncertain and complexity problem when DEMATEL is used to find the rank of
criteria with it positive of evaluate the comparative strength while VIKOR will choose the
best fit alternative.

26
 - Pythagorean Fuzzy SWARA + Pythagorean Fuzzy PROMETHEE: The last
method is combined based on two method that only used for specific role. If SWARA only
can weight criteria, PROMETHEE still ranking alternative without considered the accurate of
weight of criteria.

3.2. SYSTEM DESIGN DESCRIPTION

3.2.1. DESIGN STRUCTURE
[Link]. FUZZY AHP + FUZZY TOPSIS
[Link].1. FUZZY AHP
The Fuzzy AHP method will be applied to find the relative weight between the main
criteria as well as between the sub criterion inside that main criteria. The steps on this
subsection will be based on Yucesan (2019):

Step 1:

From the linguistic collected in the first stage of step 1. Create the compromised
(𝑘) 𝑝(𝑖)𝑘
matrix Ak=(𝑎𝑖𝑗 )𝑚∗𝑚 (1) with 𝑎𝑖𝑗 𝑘 = (2) where m represent the number of criteria,
p(j)𝑘

k is the set of expert and p(i) is the point that expert k give for the criteria i. Beacause for
each p we have 9 case which aij will have total of 81 numbers, we first classify them into 9
term in table (3) before transferring it into the fuzzy form. Finally, we will calculate the
aggregate compromised matrix A=∑𝑘 𝐴𝑘 ⨂ 𝑤 𝑘 .

Step 2:

Calculate the Difference matrix 𝐷 = (𝑑𝑖𝑘 )𝑚∗𝑚 (3) between the lower and upper
value of membership and non-membership function.

2 2
𝑑𝑖𝑘𝐿 = 𝜇𝑖𝑘𝐿 − 𝑣𝑖𝑘𝑈 (4)
2 2
𝑑𝑖𝑘𝑈 = 𝜇𝑖𝑘𝑈 − 𝑣𝑖𝑘𝐿 (5)

Step 3:

Determine the Interval Multiplicated Matrix S=(𝑠𝑖𝑘 )𝑚∗𝑚 (6)

𝑆𝑖𝑘𝐿 = √1000𝑑𝑖𝑘𝐿 (7)

𝑆𝑖𝑘𝑈 = √1000𝑑𝑖𝑘𝑈 (8)

27
Step 4:

Calculate the determinacy value (𝜏𝑖𝑗 )𝑚∗𝑚 (8)

2 2 2 2
𝜏𝑖𝑘 = 1 − (𝜇𝑖𝑘𝑈 − 𝜇𝑖𝑘𝐿 ) − (𝑣𝑖𝑘𝑈 − 𝑣𝑖𝑘𝐿 ) (9)

Step 5:

The weight matrix before normalization is found by multiplying the τ with the average
of two matrix.

𝑠𝑖𝑘𝐿 + 𝑠𝑖𝑘𝑈
𝑡𝑖𝑘 = ∗ 𝜏𝑖𝑘 (10)
2
Step 6:

The normalized priority weights are computed

∑𝑚𝑘=1 𝑡𝑖𝑘
𝑤𝑖 = 𝑚 𝑚 (11)
∑𝑖=1 ∑𝑘=1 𝑡𝑖𝑘
Step 7:

Lastly, the global weight of each sub criteria is calculated by multiply the weight of
sub criteria on the previous step by the main weight.

[Link].2. FUZZY TOPSIS

The PF-TOPSIS method is used to rank the alternative. In this study, we base on the
model built by Mostafa Hajiaghaei-Keshteli (2023).

Step 1:

From the linguistic data, transfer it into the fuzzy number form using table (2)

Step 2:

Calculate the weight level of expert among other experts by the formular below

𝜇𝑘
𝜇𝑘 + 𝜋𝑘 ∗ (𝜇 )
𝑘 +𝜗𝑘
𝜎𝑘 = 𝜇𝑘
(12)
∑𝑛𝑘=1(𝜇𝑘 + 𝜋𝑘 ∗ ( ))
𝜇 𝑘 +𝜗𝑘

28
Step 3:

Calculate the Combined Pythagorean Fuzzy Decision Matrix using the Pythagorean
Fuzzy Weighted Average operator (PFWA) by take product between the Pythagorean fuzzy
decision matrix and the weight level of expert:

𝑛 𝜎𝑘
(𝑘) 2
𝑥𝑖𝑗 = √1 − ∏ (1 − (𝜇𝑖𝑗 ) ) ,
𝑘=1
(
(13)
𝑛 𝑛 𝜎𝑘 𝑛
(𝑘) 𝜎𝑘 (𝑘) 2 (𝑘) 𝜎𝑘
∏(𝜗𝑖𝑗 ) , √1 − ∏ (1 − (𝜇𝑖𝑗 ) ) − ∏((𝜗𝑖𝑗 ) )2
𝑘=1 𝑘=1 𝑘=1
)
Step 4:

Calculating a Combined Weighted Pythagorean Fuzzy Decision Matrix: X′ = (𝑥′𝑖𝑗 )𝑙𝑚

by multiplying the combined Pythagorean decision matrix and the weight matrix of the
criteria calculated by AHP method above.

′
𝑥𝑖𝑗 = 𝑥𝑖𝑗 ∗ 𝑤𝑗 (14)

Step 5:

Identifying the Positive Ideal Solution and the Negative Ideal Solution: The
calculation process of the Pythagorean fuzzy positive ideal (PFPIS) and negative ideal
solution (PFNIS) is given in the equations below.

𝑥 + = {< (𝐶1 , 𝑃(𝜇1+ , 𝜗1+ )) >, < (𝐶2 , 𝑃(𝜇2+ , 𝜗2+ )) >, … , < (𝐶𝑚 , 𝑃(𝜇𝑚
+ +
, 𝜗𝑚 )) >} (15)

𝑥 − = {< (𝐶1 , 𝑃(𝜇1− , 𝜗1− )) >, < (𝐶2 , 𝑃(𝜇2− , 𝜗2− )) >, … , < (𝐶𝑚 , 𝑃(𝜇𝑚
− −
, 𝜗𝑚 )) >} (16)

Where s is a score function that compare two Pythagorean Fuzzy Number:

𝑃(𝜇𝑖 , 𝜗𝑖 ) = 𝜇(𝑖)2 − 𝜈(𝑖)2 (17)

Step 6:

For each alternative, determine the distance to the PFPIS and PFNIS using Euclidean
distance which D(xi,x+) is the distance from current alternative I to the PFPIS and D(xi,x-)
from PFNIS shown below:

29
 𝑚
1 2 2 2
𝐷(𝑥𝑖 , 𝑥 + ) = √ ∗ ∑ 𝑤𝑗 (|(𝜇𝑖𝑗 ) − (𝜇 + )2 | + |(𝜗𝑖𝑗 ) − (𝜗 + )2 | + |(𝜋𝑖𝑗 ) − (𝜋 + )2 |) (18)
2
𝑗=1

𝑚
−)
1 2 2 2
𝐷(𝑥𝑖 , 𝑥 = √ ∗ ∑ 𝑤𝑗 (|(𝜇𝑖𝑗 ) − (𝜇 − )2 | + |(𝜗𝑖𝑗 ) − (𝜗 − )2 | + |(𝜋𝑖𝑗 ) − (𝜋 − )2 |) (19)
2
𝑗=1

Step 7:

After calculating the two distances, the Index of Relative Closeness of Alternative: the
Index determine how close is that alternative to the positive ideal solution and how far is it
from the negative ideal solution, is found in the given formular:

𝐷(𝑥𝑖 , 𝑥 − )
𝜚(𝑥𝑖 ) = ( ) (20)
𝐷(𝑥𝑖 , 𝑥 − ) + 𝐷(𝑥𝑖 , 𝑥 + )
Step 8:

Choosing the most suitable supplier. The alternative which the largest Index of
Relative Closeness of Alternative is the most suitable that experts choose for our case study.

[Link]. FUZZY DEMATEL + VIKOR

[Link].1. FUZZY DEMATEL
First, the weight of criteria will be computed by DEMATEL, the procedure is
proposed by [Link]

Step 1:

Define the linguistic term and fuzzification using table (1)

Step 2:

From the linguistics collected in the first stage of step 1. Create the compromised
𝑝(𝑖)𝑘
matrix Ak=(𝑎𝑖𝑗 )𝑚∗𝑚 (𝑘) (21) with 𝑎𝑖𝑗 𝑘 = (22) where m represent the number of
p(j)𝑘

criteria, k is the set of expert and p(i) is the point that expert k give for the criteria i. Beacause
for each p we have 9 case which aij will have total of 81 numbers, we first classify them into
9 term in table (3) before transferring it into the fuzzy form. Finally, we will calculate the
aggregate compromised matrix A=∑𝑘 𝐴𝑘 ⨂𝑤 𝑘 (23)

30
Step 3:

Defuzzification the aggregate matrix using score function.

𝑍 = (𝑧𝑖𝑗 )𝑚∗𝑚 (24)

𝑧𝑖𝑗 = 𝜇𝑎2𝑖𝑗 − 𝑣𝑎2𝑖𝑗 (25)

Step 4:

Construct the normalized crisp matrix:

𝑋 = 𝑠𝑍 (26)
1
𝑠= (27)
𝑚𝑎𝑥1≤𝑖≤𝑚 ∑𝑚
𝑗=1 𝑧𝑖𝑗

Step 5:

Compute total-relation matrix:

𝑇 = 𝑋(𝐼 − 𝑋)−1 (28)

Step 6:

Calculate the sum of row and sum of column:

𝑟𝑖 = ∑ 𝑡𝑖𝑗 (29)
𝑗=1
𝑚

𝑐𝑗 = ∑ 𝑡𝑖𝑗 (30)
𝑖=1

Step 7:

√(𝑟𝑖 + 𝑐𝑖 )2 + (𝑟𝑖 − 𝑐𝑖 )2
𝑊𝑖 = (31)
∑𝑚 2
𝑖=1 √(𝑟𝑖 + 𝑐𝑖 ) + (𝑟𝑖 − 𝑐𝑖 )
2

31
[Link].2. FUZZY VIKOR
Step 1:

Because the limit of VIKOR is only deal with crisp value, so after aggregate the result,
the score function will be used to compute the crisp matrix 𝑓𝑖𝑗(𝑚∗𝐽) (32) with J is number of
alternatives.

Step 2:

Find the best and worst value:

𝑓𝑖+ = 𝑚𝑎𝑥𝑗 𝑓𝑖𝑗 (33)

𝑓𝑖− = 𝑚𝑖𝑛𝑗 𝑓𝑖𝑗 (34)

Step 3:

Calculate the maximum group utility Sj and the minimum individual regret of the
opponent Rj:

𝑆𝑗 = 𝐿1,𝑗 = ∑ 𝑤𝑗 ∗ ( 𝑓𝑖+ − 𝑓𝑖𝑗 )/(𝑓𝑖+ − 𝑓𝑖− ) (34)

𝑖=1

𝑅𝑗 = max (𝑓𝑖+ − 𝑓𝑖𝑗 )/(𝑓𝑖+ − 𝑓𝑖− ) (35)

Step 4:

Compute Qj

𝑣(𝑆𝑗 − 𝑆 − ) (1 − 𝑣)(𝑅𝑗 − 𝑅 − )
𝑄𝑗 = + + (36)
𝑆 − 𝑆− 𝑅+ − 𝑅−
Step 5:

The alternative with lowest Q is chosen. However, two conditions need to be checked:

1
C1: Where a” is the second lowest alternative and 𝐷𝑄 = 𝐽−1

𝑄(𝑎′′ ) − 𝑄(𝑎) ≥ 𝐷𝑄 (37)

C2: Is R(a’) and S(a’) is also the lowest one.

The compromise of solution is proposed when only one of them is satisfied:

32
 - Two alternative a’ and a” if only C2 is satisfied.

- Set of alternative a’..am with Q(am) is the lowest value that satisfied:

Recheck C1:

𝑄(𝑎′′ ) − 𝑄(𝑎𝑚 ) ≥ 𝐷𝑄 (38)

[Link]. FUZZY SWARA + PROMETHEE

[Link].1. FUZZY SWARA
The SWARA technique is computed based on [Link]
899X/1090/1/012045

Step 1:

The Criteria after being defuzzification is arrange based on their importance to make
the matrix C

Step 2:

𝐶𝑗−1 − 𝐶𝑗
𝑆𝑗 = ∨ 𝑗𝜖2: 𝑚 (39)
𝑟
Step 3:

𝐾1 = 1; 𝐾𝐽 = 𝑆𝑗 + 1 (40)

Step 4:

𝑊𝑗−1
𝑊𝑗 = ; 𝑊1 = 1 (41)
𝐾𝑗
Step 5:

The total weight of criteria is computed by normalized W.

[Link].2. FUZZY PROMETHEE

The PROMETHEE method is based on pairwise comparison of each alternative with
other alternatives for the predefined set of criteria. There are two variants of PROMETHEE
generally used. PROMETHEE I is used for partial ordering while complete ordering is done
through PROMETHEE II, i.e. it provides ranking for all the alternatives. The procedure
discussed in this section is PROMETHEE II which is used to obtain complete ranking of
alternatives.
33
 The procedural steps involved in PROMETHEE II method are enlisted as below:

Step 1:

Decision matrix is normalized using the following equation:

𝑥𝑖𝑗 − min(𝑥𝑖𝑗 )
𝑅𝑖𝑗 = (42)
max(𝑥𝑖𝑗 ) − min (𝑥𝑖𝑗 )

where xij is the performance measure of ith alternative with respect to jth criterion.

In case of non-beneficial criteria the equation can be modified as below:

max(𝑥𝑖𝑗 ) − 𝑥𝑖𝑗
𝑅𝑖𝑗 ′ = (43)
max(𝑥𝑖𝑗 ) − min (𝑥𝑖𝑗 )
Step 2:

Pairwise calculation of differences is done for all alternatives based on their criteria
values to obtain the evaluative differences for every alternative in comparison with other
alternatives.

Step 3:

Calculate the preference function, Pj(i,i')

While as per Brans [3] there exist six types of preference functions, it is usually
difficult for the decision maker to ascertain which preference function is suitable for each
factor. A simplified preference function is given as below:

𝑃𝑗(𝑖,𝑖 ′ ) = 0 𝑖𝑓 𝑅𝑖𝑗 < 𝑅𝑖′𝑗 (44)

𝑃𝑗(𝑖,𝑖 ′ ) = (𝑅𝑖𝑗 − 𝑅𝑖 ′ 𝑗 ) 𝑖𝑓 𝑅𝑖𝑗 ≥ 𝑅𝑖′𝑗 (45)

Step 4:

Calculate the aggregated preference function considering the weights assigned to each
criteria. Aggregated preference function:

∑𝑚
𝑗=1 𝑤𝑗 ∗ 𝑃𝑗(𝑖,𝑖 ′ )
′)
𝜋(𝑖, 𝑖 = (46)
∑𝑚
𝑗=1 𝑤𝑗

where wj is the relative importance (weight) of jth criterion.

34
Step 5:

Calculate the outranking flow/PROMETHEE II complete ranking

𝑛
+ (𝑖)
1
𝜑 = ∑ 𝜋(𝑖 ′ , 𝑖) (47)
𝑛−1 ′
𝑖 =1
𝑛
1
𝜑 − (𝑖) = ∑ 𝜋(𝑖 ′ , 𝑖) (48)
𝑛−1 ′
𝑖 =1

where “n” is the number of alternatives.

Step 6:

Calculate the net outranking flow/PROMETHEE II complete ranking

𝜑(𝑖) = 𝜑 + (𝑖) − 𝜑 − (𝑖) (49)

Step 7:

Values of ϕ(i) are then used to rank all the alternatives. Higher value of ϕ(i) indicates
that the alternative is better than an alternative with lower ϕ(i) value. The best alternative is
thus the one which has the highest value of ϕ(i).

3.2.2. KEY ADVANTAGES OF THE PROPOSED SYSTEM

The propose iterated method make the advantage of filling the gap of advantage of
each method that is propose in previous part. Moreover, using various methods make the data
stronger to be proved.

35
 CHAPTER 4: PROTOTYPE DEVELOPMENT
AND IMPLEMENTATION
4.1. PROTOTYPE DEVELOPMENT
To gather data from potential suppliers, we will use a survey form (Appendix) that
captures information based on the identified parameters.

4.2. RESULTS
Base on the data collected, the result done by each method is listed in the tables below:

Table 6. Weight of each criterion found by three methods

Criteria TOPSIS + AHP DEMATEL + VIKOR SWARA + PROMETHEE

C11 0.057988 0.047753 0.054878

C12 0.036432 0.047754 0.041726

C13 0.038652 0.047756 0.044218

C14 0.057943 0.047752 0.050193

C21 0.057453 0.04909 0.057463

C22 0.051497 0.049073 0.053742

C23 0.044555 0.049077 0.041988

C24 0.042817 0.049082 0.043128

C31 0.073195 0.070037 0.081183

C32 0.067787 0.070048 0.074251

C33 0.069135 0.070032 0.054682

C41 0.111426 0.111426 0.114054

36
C42 0.111426 0.111426 0.108797

C51 0.03554 0.035938 0.035982

C52 0.037491 0.035941 0.043515

C53 0.037779 0.035939 0.033976

C54 0.034168 0.035939 0.032141

C55 0.034718 0.035939 0.034081

Table 7. Distance and rank by AHP + TOPSIS

Alternative D+ D- q Rank

1 2.4956 4.8576 0.6606 1

2 3.9580 3.9798 0.5014 2

3 5.1352 2.1048 0.2907 3

Table 8. R, S, Q and rank by DEMATEL+ VIKOR

Alternative R S Q Rank

1 0.0491 0.3608 0 1

2 0.1114 0.4639 0.6285 2

3 0.0833 0.7616 0.7748 3

Table [Link] and rank by SWARA + PROMETHEE

Alternative Φ Rank

1 0.3602 1

2 0.2413 2

3 -0.0398 3

Based on the tables, three methods present nearly the same result. In the weight of
criteria, we can see that only the two criteria C41 and C42 which are related to customer

37
service, have a proportion larger than 10 percent. The next important main criterion is the
trustfulness of the supplier. The other main criteria, although conclude more sub criteria, have
been pointed less important so their weight is not high, about 3 percent. The alternative 1 is
highly recommended by all method, followed by is the Thai Nguyen region and the worst one
which is far unsuitable than the others is the third alternative.

4.3. ANALYSIS
4.3.1. SENSITIVITY ANALYSIS
The current result is under the same proportion of weight of expert. In the real case,
the first five people in the list work directly on the procurement process, the rest two is only
the sale managers of garment industry so they do not really know clearly about the
semiconductor supplier. So, their choice will affect the final result. That is why some analysis
had to be done.

Besides that, the weight of the main criteria is another factor that needs to be
considered for change. The world is changed day to day so the important of these criteria is
only relative with current time. In the future, these numbers will be changed so the sensitivity
will help us to know what the comparative advantages of each company in our study is. With
the trend of sustainability, the last three main criteria are chosen to be tested.

The factors, changing points and the new results are listed below:

Table 10. Sensitivity analysis

Type Factor Changing Result

AHP DEMATEL SWARA

Ranking
+ TOPSIS +VIKOR +PROMETHEE

1 1 1(*) 2

Expert 6 Eliminate 2 2 2 1

Expert 3 3 3 3

weight 1 1 1(*) 2

Expert 2 Eliminate 2 2 2 1

3 3 3 3

38
 1 1 2,1 2

Expert 6,7 0.2,0.65 2 2 - 1

3 3 3 3

1 1 1(*) 3

Expert 3,7 Eliminate 2 2 2 2

3 3 3 1

1 1 1(*) 1

0.3 2 3 3 2

3 2 2 3
Criteria 3
1 1 2(*) 2

0.4 2 2 3 1

3 3 1 3

1 1 1,2 1

0.3 2 2 - 2

3 3 3 3
Main Criteria 4
1 1 1,2 1
Criteria
0.4 2 2 - 2

3 3 3 3

1 1 1(*) 1

0.25 2 2 2 2

3 3 3 3

Criteria 5 1 1 1(*) 1

0.35 2 2 2 2

3 3 3 3

0.40 1 1 1,2 1

39
 2 2 - 2

3 3 3 3

The table showed that in most cases there is a race between alternative 1 and 2,
especially by the strongly recognize by the procurement expert. On the main criteria side,
with the increase in proportion, the alternative 1 follows the sustainability strategy while the
trustfulness and the customer services sides are not as good as the second company. For the
third supplier, the only thing that is concluded is that they focus much on gaining the trust of
customers.

4.3.2. FURTHER ANALYSIS

Economic, environmental, and social factors all play a role in supplier selection. While
environmental and social impacts may not directly influence a company's decision, they
remain important considerations (as discussed later). Economic factors and adherence to
standards like ISO, however, carry significant weight due to their direct impact on material
costs, product quality, and supplier reliability.

When selecting suppliers, it's crucial to consider environmental, social, and economic
impacts alongside traditional parameters like costs, quality, reliability, service, and
sustainability. Environmental impacts include resource utilization, emissions, and compliance
with standards like ISO 14001. Social impacts encompass labor practices, community
engagement, health and safety, and ethical conduct, ensuring suppliers adhere to ISO 26000
and OSHA standards. Economic impacts involve cost efficiency, financial stability,
innovation, and economic contribution. By incorporating these factors into the Multi-Criteria
Decision Making (MCDM) process, organizations can achieve a balanced and responsible
supplier selection, ensuring long-term sustainability and compliance with relevant standards
while optimizing costs and quality.

40
 CHAPTER 5: CONCLUSION
5.1. SUMMARY
Based on the collected data, we applied it to the model to determine the most suitable
criteria for the case. Given the absence of random variables in our models, we ran only one
replication to obtain the results. Interestingly, the outcomes, influenced by expert opinions,
varied significantly across the different MCDM methods. AHP+TOPSIS recommended the
third alternative, DEMATEL+VIKOR did not provide a definitive result, and SWARA-
PROMETHEE suggested the first alternative. This divergence highlights the importance of
considering multiple methods in supplier selection processes and suggests further
investigation into the criteria and methodologies used to ensure robust decision-making.

5.2. RECOMMENDATION
To enhance the supplier selection process based on the findings, several
recommendations are proposed. Firstly, adopting a multi-methodology approach is crucial to
validate results across different MCDM techniques, ensuring robustness and reliability in
decision-making. Secondly, refining the criteria ranges and specifications is essential to
address observed fuzziness and inconsistencies in outcomes. Regularly incorporating expert
feedback will also help maintain the relevance and accuracy of criteria weighting. Improving
data collection practices to encompass a broader spectrum of supplier performance metrics
will provide a more comprehensive basis for analysis. Expanding sensitivity analyses to
include diverse variables and scenarios will further strengthen the model's resilience and
adaptability. Future research should focus on advanced fuzzy logic techniques to better
manage uncertainty in supplier selection criteria. Lastly, establishing a continuous
improvement framework will ensure the models remain effective and aligned with evolving
business needs.

41
 REFERENCES
Ak, M. F., & Gul, M. 2019. "AHP–TOPSIS integration extended with Pythagorean fuzzy sets
for information security risk analysis." Complex & Intelligent Systems 5 (2).
[Link]
Atanassov, K.T. 1986. "Intuitionistic fuzzy sets." Fuzzy Sets Syst. 20 (1): 87–96.
Ayyildiz, E., & Taskin Gumus, A. 2021. "Pythagorean fuzzy AHP based risk assessment
methodology for hazardous material transportation: An application in Istanbul."
Environmental Science and Pollution Research 28 (27): 35798–35810.
[Link]
Boix-Cots, D., Pardo-Bosch, F., & Pujadas, P. 2023. "A hierarchical integration method under
social constraints to maximize satisfaction in multiple criteria group decision making
systems." Expert Systems with Applications 216.
[Link]
Büyük¨ozkan, G., & G¨oçer, F. 2021. "Evaluation of software development projects based on
integrated Pythagorean fuzzy methodology." Expert Systems with Applications 183.
[Link]
Cevik Onar, S., Oztaysi, B., & Kahraman, C. 2018. "Multicriteria evaluation of cloud service
providers using pythagorean fuzzy TOPSIS." Journal of Multiple-Valued Logic 30 (2-
3).
Cheraghalipour, A., Paydar, M. M., & Hajiaghaei-Keshtelia, M. 2017. "An integrated
approach for collection center selection in reverse logistics." International Journal of
Engineering, Transactions A: Basics 30 (7). [Link]
Ecer, F. 2022. "Multi-criteria decision making for green supplier selection using interval type-
2 fuzzy AHP: A case study of a home appliance manufacturer." Operational 22 (1).
[Link]
Gündo˘gdu, F. K., & Kahraman, C. 2019. "Spherical fuzzy sets and spherical fuzzy TOPSIS
method." Journal of Intelligent and Fuzzy Systems. [Link]
181401.
Han, Q., Li, W., Xu, Q., Song, Y., Fan, C., & Zhao, M. 2022. "Novel measures for linguistic
hesitant Pythagorean fuzzy sets and improved TOPSIS method with application to
contributions of system-of-systems." Expert Systems with Applications 199.
[Link]
Hwang, C. L., & Yoon, K. 1981. Multiple attribute decision making: Methods and
applications. Springer-Verlag.

42
Ilbahar, E., Karas¸an, A., Cebi, S., & Kahraman, C. 2018. "A novel approach to risk
assessment for occupational health and safety using Pythagorean fuzzy AHP & fuzzy
inference system." Safety Science 103: 124–136.
[Link]
Ishizaka, A., Balkenborg, D., & Kaplan, T. 2011. "Influenceofaggregation
andmeasurementscaleonranking a compromise alternative in AHP." Journal of the
Operational Research Society 62 (4): 700–710.
Jin, G. 2023. "Selection of virtual team members for smart port development projects through
the application of the direct and indirect uncertain TOPSIS method." Expert Systems
with Applications 217. [Link]
Kumar, S., & Barman, A. G. 2021. "Fuzzy TOPSIS and fuzzy VIKOR in selecting green
suppliers for sponge iron and steel manufacturing." Soft Computing 25 (8).
[Link]
Kumari, M., Kanti De, P., Narang, P., & Shah, N. H. 2023. "Integrated optimization of
inventory, replenishment, and vehicle routing for a sustainable supply chain utilizing a
novel hybrid algorithm with carbon emission regulation." Expert Systems.
[Link]
Pe´rez-Domı´nguez L, Rodrı´guez-Pico´n LA, Alvarado-Iniesta A, Luviano Cruz D, Xu Z.
2018. "MOORA under Pythagorean fuzzy set for multiple criteria decision making."
2018: 1-10.
Pınar, A., Erdebilli, B. D. R. B., & ¨Ozdemir, Y. S. 2021. "Q-rung orthopair fuzzy topsis
method for green supplier selection problem." Sustainability (Switzerland) 13 (2): 985.
[Link]
Smarandache, F. 1995. "Neutrosophic logic and set. mss."
TL, Saaty. 1990. How to make a decision: the analytic hierarchy process. Vol. 48. 1 vols. Eur
J Oper Res.
Wang, Q., Wang, H., Qi, Z. 2016. "An application of nonlinear fuzzy analytic hierarchy
process in safety evaluation of coal mine." Safety Science 86: 78–87.
Yager, R. R. 2014. "Pythagorean membership grades in multicriteria decision making." IEEE
Transactions on Fuzzy Systems 22 (4): 958–965.
[Link]
Yager, R.R., Abbasov, A.M.,. 2013. "Pythagorean membership grades, complex numbers, and
decision making." International Journal of Intelligent Systems 28 (5): 436–452.
Yager, R.R., Alajlan, N. 2017. "Approximate reasoning with generalized orthopair fuzzy
sets." Information Fusion 38: 65–73.
43
Yu, C., Shao, Y., Wang, K., & Zhang, L. 2019. "A group decision making sustainable
supplier selection approach using extended TOPSIS under interval-valued." Expert
Systems with Applications 121: 1-17. [Link]
Zeng S, Chen J, Li X. 2016. "A hybrid method for pythagorean fuzzy multiple-criteria
decision making." International Journal of Information Technology & Decision
Making 15 (02): 403–422.
Zhang, X., & Xu, Z. 2014. "Extension of TOPSIS to multiple criteria decision making with
pythagorean fuzzy sets." International Journal of Intelligent Systems 29 (12).
[Link]

44
 APPENDIX
A1. SURVEY FORM

MAIN CRITERIA

How are RAW MATERIAL PRICES in the following areas evaluated?

How are SHIPPING COSTS in the following areas evaluated?

How are HIDDEN FEES in the following areas evaluated?

How are PAYMENT TERMS in the following areas evaluated?

How are CONSISTENCY in the following areas evaluated?

How are PRODUCT DEFECT RATES in the following areas evaluated?

How are INVENTORY MANAGEMENT in the following areas evaluated?

How are FACTORY PERFORMANCE in the following areas evaluated?

How are ON-TIME DELIVERY in the following areas evaluated?

How are ABILITY TO MEET PRODUCTION DEMANDS in the following areas evaluated?

How are EXPERIENCES in the following areas evaluated?

How are RESPONSIVENESS in the following areas evaluated?

How are WARRANTY TERMS in the following areas evaluated?

How are HUMAN RESOURCES in the following areas evaluated?

How are WORK SAFETY & LABOR HEALTH in the following areas evaluated?

How are FINANCIAL STABILITY in the following areas evaluated?

How are ENVIRONMENTAL RESPONSIBILITY in the following areas evaluated?

How are SOCIAL RESPONSIBILITY in the following areas evaluated?

FOLLOWING AREAS WEIGHTED SCORE

INDUSTRIAL ZONE HOA LAC (HA NOI) No Difference

INDUSTRIAL ZONE YEN BINH (THAI NGUYEN) Slight Difference

45
INDUSTRIAL ZONE YEN PHONG – VAN TRUNG
Difference
(BAC NINH – BAC GIANG)

Big Difference

Huge Difference

46
 A2. MATLAB CODE

A2.1. FUZZY AHP + TOPSIS

Main code
clear all
clc
c=18;% number of criterias
a=3;%number of alternatives
e=7; %number of expert
for i=1:e
for j=1:a
xi(:,:,j,i)=readmatrix('[Link]','Sheet','Result','Range',[3+c*(
i-1),2+j,3+c*(i-1)+17,2+j]);
for k=1:c
if xi(k,:,j,i)==1
x(k,:,j,i)=readmatrix('[Link]','Sheet','Fuzzyfication','Range',
[2,2,2,4]);
elseif xi(k,:,j,i)==2
x(k,:,j,i)=readmatrix('[Link]','Sheet','Fuzzyfication','Range',
[3,2,3,4]);
elseif xi(k,:,j,i)==3
x(k,:,j,i)=readmatrix('[Link]','Sheet','Fuzzyfication','Range',
[4,2,4,4]);
elseif xi(k,:,j,i)==4
x(k,:,j,i)=readmatrix('[Link]','Sheet','Fuzzyfication','Range',
[5,2,5,4]);
else
x(k,:,j,i)=readmatrix('[Link]','Sheet','Fuzzyfication','Range',
[6,2,6,4]);
end
end
end
end

47
P=readmatrix('[Link]','Sheet','Expert
weight','Range',[3,2,2+e,4]);
ro=calro(P,e);

AHP
for i=1:e
Xtest(:,1,:,i)=(1-x(:,1,:,i).^2).^ro(1,i);
Xtest(:,2,:,i)=x(:,2,:,i).^ro(1,i);
end
for i=1:a
X(:,1,i)=sqrt(1-prod(Xtest(:,1,i,:),4));
X(:,2,i)=prod(Xtest(:,2,i,:),4);
X(:,3,i)=sqrt(prod(Xtest(:,1,i,:),4)-
(prod(Xtest(:,2,i,:),4)).^2);
end
Xend(:,1,:)=X(:,1,:).^2-X(:,2,:).^2;
% Xend(:,1,:)=sqrt(1-(1-X(:,1,:).^2).^(w'));
% Xend(:,2,:)=X(:,2,:).^(w');
% Xend(:,3,:)=sqrt(X(:,1,:).^2+X(:,2,:).^2);
s(:,1,:)=Xend(:,1,:).*(w');
[x,y]=max(s,[],3);
[x1,y1]=min(s,[],3);
for i=1:c
Xpos(i,:)=Xend(i,:,y(i,1));
Xneg(i,:)=Xend(i,:,y1(i,1));
end
[Dpos,Dneg]=Discal(Xend,Xpos,Xneg,w);
for i=1:a
q(i)=Dneg(:,:,i)./(Dpos(:,:,i)+Dneg(:,:,i));
end
[~,position]=max(q);
fprintf('The Best Alternative is: %f',position);

48
AHP SCRIPT
nummain=5;
numsub1=4;
numsub2=4;
numsub3=3;
numsub4=2;
numsub5=5;
scale=readmatrix('[Link]','Sheet','Fuzzyfication','Range',[3
,2,12,5]);
for i=1:e
main(:,:,i)=readmatrix('[Link]','Sheet','Result','Range',[i+
1,2,i+1,nummain+2-1]);
sub1(:,:,i)=readmatrix('[Link]','Sheet','Result','Range',[i+
10,2,i+10,numsub1+2-1]);
sub2(:,:,i)=readmatrix('[Link]','Sheet','Result','Range',[i+
19,2,i+19,numsub2+2-1]);
sub3(:,:,i)=readmatrix('[Link]','Sheet','Result','Range',[i+
28,2,i+28,numsub3+2-1]);
sub4(:,:,i)=readmatrix('[Link]','Sheet','Result','Range',[i+
37,2,i+37,numsub4+2-1]);
sub5(:,:,i)=readmatrix('[Link]','Sheet','Result','Range',[i+
46,2,i+46,numsub5+2-1]);
fuzzymain(:,:,:,i)=paircal(main(:,:,i),scale,ro(i));
fuzzysub1(:,:,:,i)=paircal(sub1(:,:,i),scale,ro(i));
fuzzysub2(:,:,:,i)=paircal(sub2(:,:,i),scale,ro(i));
fuzzysub3(:,:,:,i)=paircal(sub3(:,:,i),scale,ro(i));
fuzzysub4(:,:,:,i)=paircal(sub4(:,:,i),scale,ro(i));
fuzzysub5(:,:,:,i)=paircal(sub5(:,:,i),scale,ro(i));
end
pairwisemain=fuzzymain(:,:,:,1);
pairwisesub1=fuzzysub1(:,:,:,1);
pairwisesub2=fuzzysub2(:,:,:,1);
pairwisesub3=fuzzysub3(:,:,:,1);
pairwisesub4=fuzzysub4(:,:,:,1);
49
pairwisesub5=fuzzysub5(:,:,:,1);
for z=2:e
for zz=1:2
pairwisemain(:,:,2*(zz-1)+1)=sqrt(pairwisemain(:,:,2*(zz-
1)+1).^2+fuzzymain(:,:,2*(zz-1)+1,z).^2-pairwisemain(:,:,2*(zz-
1)+1).^2.*fuzzymain(:,:,2*(zz-1)+1,z).^2);
pairwisemain(:,:,2*(zz-1)+2)=pairwisemain(:,:,2*(zz-
1)+2).*fuzzymain(:,:,2*(zz-1)+1,z);
pairwisesub1(:,:,2*(zz-1)+1)=sqrt(pairwisesub1(:,:,2*(zz-
1)+1).^2+fuzzysub1(:,:,2*(zz-1)+1,z).^2-pairwisesub1(:,:,2*(zz-
1)+1).^2.*fuzzysub1(:,:,2*(zz-1)+1,z).^2);
pairwisesub1(:,:,2*(zz-1)+2)=pairwisesub1(:,:,2*(zz-
1)+2).*fuzzysub1(:,:,2*(zz-1)+1,z);
pairwisesub2(:,:,2*(zz-1)+1)=sqrt(pairwisesub2(:,:,2*(zz-
1)+1).^2+fuzzysub2(:,:,2*(zz-1)+1,z).^2-pairwisesub2(:,:,2*(zz-
1)+1).^2.*fuzzysub2(:,:,2*(zz-1)+1,z).^2);
pairwisesub2(:,:,2*(zz-1)+2)=pairwisesub2(:,:,2*(zz-
1)+2).*fuzzysub2(:,:,2*(zz-1)+1,z);
pairwisesub3(:,:,2*(zz-1)+1)=sqrt(pairwisesub3(:,:,2*(zz-
1)+1).^2+fuzzysub3(:,:,2*(zz-1)+1,z).^2-pairwisesub3(:,:,2*(zz-
1)+1).^2.*fuzzysub3(:,:,2*(zz-1)+1,z).^2);
pairwisesub3(:,:,2*(zz-1)+2)=pairwisesub3(:,:,2*(zz-
1)+2).*fuzzysub3(:,:,2*(zz-1)+1,z);
pairwisesub4(:,:,2*(zz-1)+1)=sqrt(pairwisesub4(:,:,2*(zz-
1)+1).^2+fuzzysub4(:,:,2*(zz-1)+1,z).^2-pairwisesub4(:,:,2*(zz-
1)+1).^2.*fuzzysub4(:,:,2*(zz-1)+1,z).^2);
pairwisesub4(:,:,2*(zz-1)+2)=pairwisesub4(:,:,2*(zz-
1)+2).*fuzzysub4(:,:,2*(zz-1)+1,z);
pairwisesub5(:,:,2*(zz-1)+1)=sqrt(pairwisesub5(:,:,2*(zz-
1)+1).^2+fuzzysub5(:,:,2*(zz-1)+1,z).^2-pairwisesub5(:,:,2*(zz-
1)+1).^2.*fuzzysub5(:,:,2*(zz-1)+1,z).^2);
pairwisesub5(:,:,2*(zz-1)+2)=pairwisesub5(:,:,2*(zz-
1)+2).*fuzzysub5(:,:,2*(zz-1)+1,z);
end

50
end
mainw=weightcalAHP(pairwisemain,nummain);
mainw=readmatrix('[Link]','Sheet','Result','Range',[3,9,3,8+
nummain]);
sub1w=weightcalAHP(pairwisesub1,numsub1);
sub2w=weightcalAHP(pairwisesub2,numsub2);
sub3w=weightcalAHP(pairwisesub3,numsub3);
sub4w=weightcalAHP(pairwisesub4,numsub4);
sub5w=weightcalAHP(pairwisesub5,numsub5);
w=[];
w=[w sub1w.*mainw(1,1);];
w=[w sub2w.*mainw(1,2);];
w=[w sub3w.*mainw(1,3);];
w=[w sub4w.*mainw(1,4);];
w=[w sub5w.*mainw(1,5);];

FUNCTION TO CALCULATE THE COMPROMISE WEIGHT

function ro=calro(P,e)
for k=1:e
dro(k,1)=P(k,1)+P(k,3)*P(k,1)/(P(k,1)+P(k,2));
end
for k=1:e
ro(1,k)=dro(k,1)/sum(dro);
end
end

FUNCTION TO CALCULATE THE DISTANCE

function [Dposs,Dnegg]=Discal(Xend,Xpos,Xneg,w)
Pdis=abs(Xend.^2-Xpos.^2);
Pneg=abs(Xend.^2-Xneg.^2);
Dpos=sqrt(0.5*sum(sum(Pdis,2).*w));
Dneg=sqrt(0.5*sum(sum(Pneg,2).*w));
Dposs=sum(Dpos);
Dnegg=sum(Dneg);
End

51
FUNCTION TO CALCULATE THE PAIRWISE COMPARISION
function fuzzypair=paircal(X,scale,ro)
le=length(X);
for i=1:le
for j=1:le
crisppair(i,j)=X(1,i)/X(1,j);
if crisppair(i,j)<=0.25
for k=1:4
fuzzypair(i,j,k)=scale(1,k);
end
elseif crisppair(i,j)<=0.4
for k=1:4
fuzzypair(i,j,k)=scale(2,k);
end
elseif crisppair(i,j)<=0.5
for k=1:4
fuzzypair(i,j,k)=scale(3,k);
end
elseif crisppair(i,j)<=2/3
for k=1:4
fuzzypair(i,j,k)=scale(4,k);
end
elseif crisppair(i,j)<=4/3 && crisppair(i,j)~=1
for k=1:4
fuzzypair(i,j,k)=scale(5,k);
end
elseif crisppair(i,j)<=5/3&&crisppair(i,j)~=1
for k=1:4
fuzzypair(i,j,k)=scale(6,k);
end
elseif crisppair(i,j)<=2&&crisppair(i,j)~=1
for k=1:4
fuzzypair(i,j,k)=scale(7,k);
end
52
 elseif crisppair(i,j)<=3&&crisppair(i,j)~=1
for k=1:4
fuzzypair(i,j,k)=scale(8,k);
end
elseif crisppair(i,j)<=5&&crisppair(i,j)~=1
for k=1:4
fuzzypair(i,j,k)=scale(9,k);
end
elseif crisppair(i,j)==1
for k=1:4
fuzzypair(i,j,k)=scale(10,k);
end
end
end
end
for j=1:2
fuzzypair(:,:,2*(j-1)+1)=sqrt(1-((1-fuzzypair(:,:,2*(j-
1)+1).^2).^ro));
fuzzypair(:,:,2*(j-1)+2)=fuzzypair(:,:,2*(j-1)+2).^ro;
end

FUNCTION TO CALCULATE THE WEIGHT OF CRITERIA

function w=weightcalAHP(X,le)
for i=1:le
for j=1:le
d(i,j,1)=X(i,j,1)^2-X(i,j,4)^2;
d(i,j,2)=X(i,j,2)^2-X(i,j,3)^2;
s(i,j,1)=sqrt(1000^d(i,j,1));
s(i,j,2)=sqrt(1000^d(i,j,2));
to(i,j)=1-(X(i,j,2)^2-X(i,j,1)^2)-(X(i,j,4)^2-X(i,j,3)^2);
t(i,j)=(s(i,j,1)+s(i,j,2))/2*to(i,j);
end
end
for i=1:le
w(1,i)=sum(t(i,:))/sum(sum(t));
53
end

A2.2. FUZZY DEMATEL + VIKOR

MAIN SCRIPT
clear all
clc
c=18;% number of criterias
a=3;%number of alternatives
e=7; %number of expert
nummain=5;%number of main criteria
numsub(1)=4;%number of sub criteria1
numsub(2)=4;%number of sub criteria2
numsub(3)=3;%number of sub criteria3
numsub(4)=2;%number of sub criteria4
numsub(5)=5;%number of sub criteria5
expertmatrix=zeros([e,3]);
v=0.5;
%import data from excel and fuzzyfication
scale=readmatrix('[Link]','Sheet','Fuzzyfication','Range',[3
,2,12,5]);
for i =1:e

[Link](i).Weight=readmatrix('[Link]','Sheet','Expert
weight','Range',[i+2,2,i+2,4]);
expertmatrix(i,:)= [Link](i).Weight;
end
weigthofexpert=calro(expertmatrix,e);
for i =1:e
for j=1:c
for k=1:a
[Link](i).Criteria(j).Alternative(k).Point=readmatrix('Rating.
xlsx','Sheet','Result','Range',[c*(i-1)+j+2, k+2,c*(i-1)+j+2,k+2]);
if [Link](i).Criteria(j).Alternative(k).Point==1

54
[Link](i).Criteria(j).Alternative(k).Fuzzypoint=readmatrix('Ra
[Link]','Sheet','Fuzzyfication','Range',[2,2,2,3]);
elseif
[Link](i).Criteria(j).Alternative(k).Point==2
[Link](i).Criteria(j).Alternative(k).Fuzzypoint=readmatrix('Ra
[Link]','Sheet','Fuzzyfication','Range',[3,2,3,3]);
elseif
[Link](i).Criteria(j).Alternative(k).Point==3
[Link](i).Criteria(j).Alternative(k).Fuzzypoint=readmatrix('Ra
[Link]','Sheet','Fuzzyfication','Range',[4,2,4,3]);
elseif
[Link](i).Criteria(j).Alternative(k).Point==4
[Link](i).Criteria(j).Alternative(k).Fuzzypoint=readmatrix('Ra
[Link]','Sheet','Fuzzyfication','Range',[5,2,5,3]);
else
[Link](i).Criteria(j).Alternative(k).Fuzzypoint=readmatrix('Ra
[Link]','Sheet','Fuzzyfication','Range',[6,2,6,3]);
end
end
end
[Link](i).[Link]=readmatrix('[Link]','Sheet','Re
sult','Range',[1+i,2,1+i,2+nummain-1]);
[Link](i).[Link]=paircal([[Link](i).[Link]
int],scale,weigthofexpert(i));
for j=1:nummain
[Link](i).Sub(j).Point=readmatrix('[Link]','Sheet','
Result','Range',[1+9*j+i,2,1+9*j+i,2+numsub(j)-1]);
[Link](i).Sub(j).Fuzzypoint=paircal([[Link](i).Sub(j
).Point],scale,weigthofexpert(i));
end
end
%DEMATEL
[Link](:,:,1)=([[Link](1).[Link](:,:
,1)]+[[Link](1).[Link](:,:,2)])/2;

55
[Link](:,:,2)=([[Link](1).[Link](:,:
,3)]+[[Link](1).[Link](:,:,4)])/2;
for i=1:nummain
[Link](i).Pairwise(:,:,1)=([[Link](1).Sub(i).Fuzzypoint
(:,:,1)]+[[Link](1).Sub(i).Fuzzypoint(:,:,2)])/2;

[Link](i).Pairwise(:,:,2)=([[Link](1).Sub(i).Fuzzypoint
(:,:,3)]+[[Link](1).Sub(i).Fuzzypoint(:,:,4)])/2;
end
for i=2:e
[Link](:,:,1)=sqrt([[Link](:,:,1)].^2+
([[Link](i).[Link](:,:,1)]+[[Link](i).Main.
Fuzzypoint(:,:,2)]).^2-
(([[Link](i).[Link](:,:,1)]+[[Link](i).Main
.Fuzzypoint(:,:,2)]).^2.*[[Link](:,:,1)].^2));
[Link](:,:,2)=[[Link](:,:,2)].*([DEMAT
[Link](i).[Link](:,:,3)]+[[Link](i).[Link]
int(:,:,4)])/2;
for j=1:nummain
[Link](j).Pairwise(:,:,1)=sqrt([[Link](j).Pairwise(:,:,1)]
.^2+([[Link](i).Sub(j).Fuzzypoint(:,:,1)]+[[Link](i)
.Sub(j).Fuzzypoint(:,:,2)]).^2-
(([[Link](i).Sub(j).Fuzzypoint(:,:,1)]+[[Link](i).Su
b(j).Fuzzypoint(:,:,2)]).^2.*[[Link](j).Pairwise(:,:,1)].^2));
[Link](j).Pairwise(:,:,2)=[[Link](j).Pairwise(:,:,2)].*([D
[Link](i).Sub(j).Fuzzypoint(:,:,3)]+[[Link](i).Sub(j)
.Fuzzypoint(:,:,4)])/2;
end
end
[Link]=[[Link](:,:,1)].^2-
[[Link](:,:,2)].^2;
[Link]=weightcalDEMATEL([Link],nummain);
[Link]=readmatrix('[Link]','Sheet','Result','Ra
nge',[3,9,3,8+nummain]);

56
for i=1:nummain
[Link](i).Crisp=[[Link](i).Pairwise(:,:,1)].^2-
[[Link](i).Pairwise(:,:,2)].^2;
[Link](i).Weigth=weightcalDEMATEL([Link](i).Crisp,numsub(i
));
end
weigth=[];
for i=1:nummain
weigth=[weigth [Link](i).Weigth.*[Link](i)];
end
%VIKOR
for i=1:a
for j=1:c
[Link](i).Criteria(j).Aggregate(:,1)=sqrt(1-(1-
[Link](1).Criteria(j).Alternative(i).Fuzzypoint(:,1)^2)^weigth
ofexpert(1));
[Link](i).Criteria(j).Aggregate(:,2)=[Link](1).Crit
eria(j).Alternative(i).Fuzzypoint(:,2).^weigthofexpert(1);
for k=2:e
[Link](i).Criteria(j).Aggregate(1)=sqrt([[Link]
e(i).Criteria(j).Aggregate(1)].^2+1-(1-
[Link](k).Criteria(j).Alternative(i).Fuzzypoint(:,1)^2)^weigth
ofexpert(k)-((1-(1-
[Link](k).Criteria(j).Alternative(i).Fuzzypoint(:,1)^2)^weigth
ofexpert(k)).*[[Link](i).Criteria(j).Aggregate(1)].^2));
[Link](i).Criteria(j).Aggregate(2)=[[Link](i).
Criteria(j).Aggregate(2)].*([[Link](1).Criteria(j).Alternative
(i).Fuzzypoint(2)]^weigthofexpert(k));
end
[Link](i).Criteria(j).Crisp=[Link](i).Criteria
(j).Aggregate(1).^2-
[Link](i).Criteria(j).Aggregate(2).^2;
end
Crispmatrix(i,:)= [[Link](i).[Link]];

57
end
[S,R,Q]=findvikor(Crispmatrix,weigth,a,v);
DQ=1/(a-1);
[valueq ,posq]=sort(Q);
[values ,poss]=sort(S);
[valuer ,posr]=sort(R);
if Q(posq(2))-Q(posq(1))>=DQ
if S(posq(1))==S(poss(1))&&R(posq(1))==R(posq(1))
fprintf('The Best Alternative is: %f ',posq(1));
else
fprintf('The Best Alternative are: %f and
%f',posq(1),posq(2));
end
else
if posq(1)==poss(1)&&posq(1)==posr(1)
i=3;
while i<=length(posq)
if Q(posq(i))-Q(posq(1))<=DQ
i=i+1;
else
break
end
end
out=posq(1:i-1);
fprintf('The best alternative are %f ',posq(1));
for kk=2:length(out)
fprintf('and %f',posq(kk));
end
else
fprintf('no alternative is strongly suggested by this
method');
end
end

58
FUNCTION TO CALCULATE THE PAIRWISE COMPARISION
function fuzzypair=paircal(X,scale,ro)
le=length(X);
for i=1:le
for j=1:le
crisppair(i,j)=X(1,i)/X(1,j);
if crisppair(i,j)<=0.25
for k=1:4
fuzzypair(i,j,k)=scale(1,k);
end
elseif crisppair(i,j)<=0.4
for k=1:4
fuzzypair(i,j,k)=scale(2,k);
end
elseif crisppair(i,j)<=0.5
for k=1:4
fuzzypair(i,j,k)=scale(3,k);
end
elseif crisppair(i,j)<=2/3
for k=1:4
fuzzypair(i,j,k)=scale(4,k);
end
elseif crisppair(i,j)<=4/3 && crisppair(i,j)~=1
for k=1:4
fuzzypair(i,j,k)=scale(5,k);
end
elseif crisppair(i,j)<=5/3&&crisppair(i,j)~=1
for k=1:4
fuzzypair(i,j,k)=scale(6,k);
end
elseif crisppair(i,j)<=2&&crisppair(i,j)~=1
for k=1:4
fuzzypair(i,j,k)=scale(7,k);
end
59
 elseif crisppair(i,j)<=3&&crisppair(i,j)~=1
for k=1:4
fuzzypair(i,j,k)=scale(8,k);
end
elseif crisppair(i,j)<=5&&crisppair(i,j)~=1
for k=1:4
fuzzypair(i,j,k)=scale(9,k);
end
elseif crisppair(i,j)==1
for k=1:4
fuzzypair(i,j,k)=scale(10,k);
end
end
end
end
for j=1:2
fuzzypair(:,:,2*(j-1)+1)=sqrt(1-((1-fuzzypair(:,:,2*(j-
1)+1).^2).^ro));
fuzzypair(:,:,2*(j-1)+2)=fuzzypair(:,:,2*(j-1)+2).^ro;
end

FUNCTION TO CALCULATE WEIGHT OF CRITERIA

function Weigth=weightcalDEMATEL(A,f)
c=sum(A,2);
d=1/max(c);
e=A.*d;
a=e*(eye(f)-e)^-1;
R=sum(a,2);
C=sum(a,1);
W=(R+C').^2+(R-C').^2;
for i=1:f
Weigth(i)=W(i)./sum(W);
end

60
FUNCTION TO RANK ALTERNATIVE
function [S,R,Q]=findvikor(Crispmatrix,weigth,a,v)
fmax=max(Crispmatrix);
fmin=min(Crispmatrix);
for i=1:a
S(i)=sum((fmax-Crispmatrix(i,:))./(fmax-fmin).*weigth);
R(i)=max((fmax-Crispmatrix(i,:))./(fmax-fmin).*weigth);
end
for i=1:a
Q(i)=v*(S(i)-min(S))/(max(S)-min(S))+(1-v)*(R(i)-
min(R))/(max(R)-min(R));
end

A2.3. FUZZY SWARA + PROMETHEE

MAIN SCRIPT
clear all
clc
c=18;% number of criterias
a=3;%number of alternatives
e=7; %number of expert
nummain=5;%number of main criteria
numsub(1)=4;%number of sub criteria1
numsub(2)=4;%number of sub criteria2
numsub(3)=3;%number of sub criteria3
numsub(4)=2;%number of sub criteria4
numsub(5)=5;%number of sub criteria5
expertmatrix=zeros([e,3]);
v=0.5;
%import data from excel and fuzzyfication
scale=readmatrix('[Link]','Sheet','Fuzzyfication','Range',[3
,2,12,5]);
for i =1:e
[Link](i).Weight=readmatrix('[Link]','Sheet','Expert
weight','Range',[i+2,2,i+2,4]);
61
 expertmatrix(i,:)= [Link](i).Weight;
end
weigthofexpert=calro(expertmatrix,e);
for i =1:e
[Link](i).[Link]=readmatrix('[Link]','Sheet','Resu
lt','Range',[1+i,2,1+i,2+nummain-1]);
for j=1:nummain

[[Link](i).[Link](j,:)]=fuzzyfication([Link](i)
.[Link](j));

[Link](i).Sub(j).Point=readmatrix('[Link]','Sheet','Re
sult','Range',[1+9*j+i,2,1+9*j+i,2+numsub(j)-1]);
for k=1:numsub(j)

[[Link](i).Sub(j).Fuzzypoint(k,:)]=fuzzyfication([[Link]
(i).Sub(j).Point(k)]);
end
end
end
[Link](:,1)=sqrt(1-(1-
[[Link](1).[Link](:,1)].^2).^weigthofexpert(1));
[Link](:,2)=[[Link](1).[Link](:,2)].^we
igthofexpert(1);
for i=1:nummain
[Link](i).Aggregate(:,1)=sqrt(1-(1-
[[Link](1).Sub(i).Fuzzypoint(:,1)].^2).^weigthofexpert(1));

[Link](i).Aggregate(:,2)=[[Link](1).Sub(i).Fuzzypoint(:,2)]
.^weigthofexpert(1);
end
for i=2:e
[Link](:,1)=sqrt([[Link](:,1)].^2+1-
(1-[[Link](i).[Link](:,1)].^2).^weigthofexpert(i)-(1-

62
(1-
[[Link](i).[Link](:,1)].^2).^weigthofexpert(i)).*[SWA
[Link](:,1)].^2);

[Link](:,2)=[[Link](:,2)].*([[Link]
t(i).[Link](:,2)].^weigthofexpert(i));
for j=1:nummain

[Link](j).Aggregate(:,1)=sqrt([[Link](j).Aggregate(:,1)].^2+1-
(1-[[Link](i).Sub(j).Fuzzypoint(:,1)].^2).^weigthofexpert(i)-
(1-(1-
[[Link](i).Sub(j).Fuzzypoint(:,1)].^2).^weigthofexpert(i)).*[S
[Link](j).Aggregate(:,1)].^2);

[Link](j).Aggregate(:,2)=[[Link](j).Aggregate(:,2)].*([SWARA.E
xpert(i).Sub(j).Fuzzypoint(:,2)].^weigthofexpert(i));
end
end
[Link]=[Link](:,1).^2-
[Link](:,2).^2;
[Link]=SWARAWeigth([[Link]],nummain);
[Link]=readmatrix('[Link]','Sheet','Result','Rang
e',[3,9,3,8+nummain]);
weigth=[];
for i=1:nummain
[Link](i).Crisp=[Link](i).Aggregate(:,1).^2-
[Link](i).Aggregate(:,2).^2;
[Link](i).Weigth=SWARAWeigth([[Link](i).Crisp],numsub(i));
weigth=[weigth [Link](i).Weigth.*[Link](i)];
end
%PROMETHEE
for i =1:e
for j=1:c
for k=1:a

63
[Link](i).Criteria(j).Alternative(k).Point=readmatrix('Rat
[Link]','Sheet','Result','Range',[c*(i-1)+j+2, k+2,c*(i-
1)+j+2,k+2]);
if
[Link](i).Criteria(j).Alternative(k).Point==1
[Link](i).Criteria(j).Alternative(k).Fuzzypoint=readmatrix
('[Link]','Sheet','Fuzzyfication','Range',[2,2,2,3]);
elseif
[Link](i).Criteria(j).Alternative(k).Point==2
[Link](i).Criteria(j).Alternative(k).Fuzzypoint=readmatrix
('[Link]','Sheet','Fuzzyfication','Range',[3,2,3,3]);
elseif
[Link](i).Criteria(j).Alternative(k).Point==3
[Link](i).Criteria(j).Alternative(k).Fuzzypoint=readmatrix
('[Link]','Sheet','Fuzzyfication','Range',[4,2,4,3]);
elseif
[Link](i).Criteria(j).Alternative(k).Point==4
[Link](i).Criteria(j).Alternative(k).Fuzzypoint=readmatrix
('[Link]','Sheet','Fuzzyfication','Range',[5,2,5,3]);
else
[Link](i).Criteria(j).Alternative(k).Fuzzypoint=readmatrix
('[Link]','Sheet','Fuzzyfication','Range',[6,2,6,3]);
end
end
end
end
for i=1:a
for j=1:c
[Link](i).Criteria(j).Aggregate(1)=sqrt(1-(1-
([Link](1).Criteria(j).Alternative(i).Fuzzypoint(1))^2).^w
eigthofexpert(1));
[Link](i).Criteria(j).Aggregate(2)=([Link](
1).Criteria(j).Alternative(i).Fuzzypoint(2)).^weigthofexpert(1);
for k=2:e

64
[Link](i).Criteria(j).Aggregate(1)=sqrt([[Link]
ternative(i).Criteria(j).Aggregate(1)].^2+1-(1-
([Link](k).Criteria(j).Alternative(i).Fuzzypoint(1))^2).^w
eigthofexpert(k)-(1-(1-
([Link](k).Criteria(j).Alternative(i).Fuzzypoint(1))^2).^w
eigthofexpert(k)).*[[Link](i).Criteria(j).Aggregate(1
)].^2);
[Link](i).Criteria(j).Aggregate(2)=[[Link]
tive(i).Criteria(j).Aggregate(2)].*([[Link](1).Criteria(j)
.Alternative(i).Fuzzypoint(2)]^weigthofexpert(k));
end
[Link](i).Criteria(j).Crisp=[Link](i).
Criteria(j).Aggregate(1)^2-
[Link](i).Criteria(j).Aggregate(2)^2;
end
Crispmatrix(i,:)= [[Link](i).[Link]];
end
Phi=findprom(Crispmatrix,a,c,weigth);
[~,position]=max(Phi);
fprintf('The Best Alternative is: %f',position);

FUZZYFICATION FUNCTION
function b=fuzzyfication(a)
if a==1
b=readmatrix('[Link]','Sheet','Fuzzyfication','Range',[2,2,2,3]
);
elseif a==2
b=readmatrix('[Link]','Sheet','Fuzzyfication','Range',[3,2,3,3]
);
elseif a==3
b=readmatrix('[Link]','Sheet','Fuzzyfication','Range',[4,2,4,3]
);
elseif a==4
b=readmatrix('[Link]','Sheet','Fuzzyfication','Range',[5,2,5,3]
);
65
else
b=readmatrix('[Link]','Sheet','Fuzzyfication','Range',[6,2,6,3]
);
end

FUNCTION TO WEIGHTING CRITERIA

function out=SWARAWeigth(Crisp,num)
out=zeros(1,num);
[a ,b]=sort(Crisp,'descend');
s(1)=0;
k(1)=1;w(1)=1;
for i=2:num
s(i)=a(i-1)-a(i);
k(i)=s(i)+1;
w(i)=w(i-1)/k(i);
end
for i=1:num
out(b(i))=w(i)/sum(w);
end

FUNCTION TO RANK ALTERNATIVE

function out=findprom(crisp,a,c,weigth)
R=zeros([a,c]);
for i =1:a
R(i,:)=((crisp(i,:)-min(crisp))./(max(crisp)-min(crisp)));
end
for i=1:a
for j=1:a
if j~=i
[Link]. Altera(i).Alterb(j).Criteria=R(i,:)-
R(j,:);
[Link](i).Alterb(j).Value=0;
for k=1:c
if [Link]. Altera(i).Alterb(j).Criteria(k)>=0

66
 [Link].
Altera(i).Alterb(j).Criteria(k)=[Link].
Altera(i).Alterb(j).Criteria(k);
else
[Link].
Altera(i).Alterb(j).Criteria(k)=0;
end
[Link](i).Alterb(j).Value=[Link](i).Alterb(j).Value+
[Link]. Altera(i).Alterb(j).Criteria(k).*weigth(k);
end
end
end
end
for i=1:a
phiplus(i)=sum([[Link](i).[Link]])/(a-1);
end
phiminus=zeros([a,1]);
for i=1:a
for j=1:a
if j~=i

phiminus(i)=phiminus(i)+[Link](j).Alterb(i).Value;
end
phiminus(i)=phiminus(i)/(a-1);
end
end
out=phiplus'-phiminus;

67