Cleaner Engineering and Technology 12 (2023) 100590

Contents lists available at ScienceDirect

Cleaner Engineering and Technology

journal homepage: www.sciencedirect.com/journal/cleaner-engineering-and-technology

Supplier selection for aerospace & defense industry through

MCDM methods
Aksel Rasmussen *, Haris Sabic **, Subrata Saha, Izabela Ewa Nielsen
Department of Materials and Production, Aalborg Øst, 9220, Denmark

A R T I C L E I N F O A B S T R A C T

Keywords: This paper utilized the leverage of three Multi-Criteria Decision Making (MCDM) methods, namely AHP, TOPSIS,
MCDM and SECA, in the supplier selection process in the context of an Aerospace & Defense (A&D) company to simplify,
AHP supplement, and standardize their ongoing practice. First, we identified and recommended additional criteria,
Fuzzy-TOPSIS
the companies might include in their present supplier selection process with justification. Second, we employ the
SECA
Supplier selection
SECA and Fuzzy-TOPSIS methods; and compare the raking with and without additional criteria to explore the
critical difference and its afterword effect on the company’s performance. Third, for easy implementation and
process standardization, we utilize the functionalities of Text Shell software ESTA so that the decision makers can
understand the impact of the inclusion of criteria such as sustainability and technology within the operation and
identify the possibility for significant improvements or expected loss. Finally, a case study is considered by
benchmarking the current practice. The results reflect that adding other criteria, such as sustainability or delivery
time, can utterly change the ongoing practice. We found that the company that needs to work in close proximity
to the government and deal with the next generation of technologies can face significant barriers when they
include additional criteria in the selection process. Moreover, it is not easy to comply without a drastic change in
their whole process to ensure quality, sustainability, and financial stability simultaneously. Most importantly,
how standardizing the entire process is a critical challenge in the aerospace and defense industry.

1. Introduction unique merits and limitations. For example, researchers and policy­
makers are recommended to consider sustainability, technology, and
The supplier selection problem to assist green transition must be R&D investment, among others, as critical criteria to support the green
discussed and evaluated from a more holistic and strategic view and transition. However, direct costs, such as labor and materials, develop­
adapted based on which section the company is operating within ment and integration of new technology, and indirect costs, such as
(Hamdan and Cheaitou, 2017). An ideal solution for designing a specific quality control, lead time, and poor delivery, might prove crucial for
supplier selection process is impossible since it is complex due to the companies. In such a scenario, DMs can rely on various MCDM methods
diversity in the factors affecting the stakeholders’ perspectives. There­ in the supplier selection process. Furthermore, the decision in practice is
fore, there is no one-size-fits-all solution (Taherdoost and Brard, 2019). sometimes dichotomous. For example, a set of suppliers that satisfies the
Due to the involvement of various criteria from various stakeholders, AS9100 ISO standard should only be considered. This means that the
researchers and practitioners use Multi-Criteria Decision Making sponsor Aerospace & Defense (A&D) company only collaborates with
(MCDM) methods to resolve the conflict. However, the number of suppliers with innovative knowledge and process technologies. In that
criteria to incorporate in the MCDM methods is a complex problem since scenario, some existing suppliers might fail to comply with those stan­
the decision makers (DMs) have to consider various criteria to choose dards and face product compliance challenges in an ever-changing
the most appropriate suppliers, which might depend on context and regulatory landscape. In addition, it is paramount that the security,
internal protocols in implementing the state of art technology solutions economic, social, and environmental perspectives are well integrated
(Massa, 2022). and aligned to achieve sustainable development. Each aspect can indi­
Noticeably, the number of criteria and their characteristics have rectly impact the overall sustainable development goal, and all aspects

* Corresponding author.
** Corresponding author.
E-mail addresses: [email protected] (A. Rasmussen), [email protected] (H. Sabic).

https://doi.org/10.1016/j.clet.2022.100590
Received 22 June 2022; Received in revised form 20 November 2022; Accepted 20 November 2022
Available online 14 December 2022
2666-7908/© 2022 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-
nc-nd/4.0/).
A. Rasmussen et al. Cleaner Engineering and Technology 12 (2023) 100590

must work together to achieve it. Therefore, implementing MCDM their operation semi-autonomously. The paper is organized as follows:
methods for finding a final recommendation needs standardization. The next section presents a review of the supplier selection and corre­
Setting weight for each criterion (performance indicator) directly in­ sponding criterion. Section 3 offers an overview of MCDM techniques in
fluences the definitive ranking of the MCDM method, and those can be the form of AHP, Fuzzy TOPSIS, and SECA. In Section 4, we present the
defined objectively or subjectively. Subjective weights can be assigned results and the proposed implementation of MCDM methods, while
based on the experiences of the experts, and objective weights can be Section 5 highlights the implications and feasibility of the study. And
used through mathematical analysis based on attributes of the data set. finally, we provide concluding remarks in Section 6.
Therefore, it is not easy to support the MCDM integration without
developing a scenario analysis and problem-solving expertise. Addi­ 2. Literature review
tionally, it is impossible to ascertain which MCDM techniques are the
best in practical use since each method has its own merits and demerits. Businesses continuously seek to create competitive advantages to
The researcher recommends that more than one MCDM technique be maintain or increase their market share. Supplier selection is a well-
used to obtain a trustworthy decision (Bahrami and Rastegar, 2022). established and well-acknowledged strategic area directly affecting the
Analyzing the ranking based on multiple MCDM methods is considered company’s success (Rouyendegh et al., 2020). The process can impact
helpful insight, even though rankings are usually not concurrent. profitability by reducing costs and improving performance, thus directly
Therefore, over the years, several MCDM methods such as TOPSIS affecting competitiveness (Taherdoost and Brard, 2019). However, the
(Hwang and Yoon, 1981), VIKOR (Duckstein and Opricovic, 1980), DMs must have a wide range of criteria to select a specific supplier with
Extended VIKOR (Sayadi et al., 2009), AHP (Saaty, 1988), MARCOS certainty. Multiple criteria provide a practical framework for bench­
(Stević et al., 2020), MOORA(Brauers and Zavadskas, 2006) and marking, but the complexity for the DMs is to choose the most appro­
ELECTRE (Roy, 1968) are used in the supplier selection process. Addi­ priate supplier that meets all the requirements (Rouyendegh et al.,
tionally, several authors used fuzzy logic to consider incom­ 2020). In this context, a criterion means a parameter that impacts
plete/imprecise information, both normal fuzzy number (Sun, 2010) whether a supplier is selected among alternatives. In the following
and interval type-2 fuzzy number (Bera et al., 2020) in the context of subsection, we suggest possible supplier criteria for the sponsor com­
supplier selection. pany with additional criteria to include that will further support proper
This study uses three MCDM methods: AHP, TOPSIS, and SECA. standardization.
TOPSIS, developed by Hwang and Yoon in 1981, is one of the simple
MCDM ranking methods used extensively to solve real-world decision
problems in diverse application areas (Yoon and Hwang, 1995). We used 2.1. Criteria
a similar approach by Nkuna et al. (2022) while implementing the AHP
and TOPSIS methods. However, the SECA method was developed The analysis of identifying the best criteria to measure the perfor­
recently by Keshavarz-Ghorabaee et al. (2018), and two key advantages mance of suppliers has been the focus of many researchers and pur­
of the method are: first, there is no need to set weights separately, and chasing practitioners during the last couple of decades (Benyoucef et al.,
second, the method is developed in such a way that it can assign weights 2003). We refer to the work by Deshmukh and Chaudhari (2011), where
for each criterion and generate final ranking by solving a multiobjective the authors analyzed 49 articles published between 1992 and 2007 to
non-linear programming model to minimize bias. This method has also summarize the criterion used in the supplier selection process. In this
gained popularity in quickly solving real-life conflicting study, we update the list as shown in Table 1.
decision-making problems due to its easy application (Wang et al., Table 1 indicates that the current literature suggests a wide variety of
2020). AHP, as developed by T.L. Saaty in 1981, is a structured method potential criteria to be used for supplier selection in different industrial
in group decision-making to rank decisions based on relative impor­ contexts. However, including additional criteria brings implementation
tance. The objective of using three methods by benchmarking their and interpretation complexities and proves to be challenging from a
current practice is to ensure higher reliability and acceptance to the case data-handling perspective. Additionally, allocating weights for each
company and its operators, who is responsible for selecting suppliers for criterion is also an issue. Therefore, we introduce the SECA method to
the company. avoid such complexity and develop a software prototype for easy
We aim to develop a system that can help DMs obtain a flexible implementation. Next, we focus on the company’s current practice with
overview and provide them with an easily applicable system to stan­ four criteria and a new extended criterion list.
dardize the process and reduce their environmental footprint. Therefore,
first, we explore the company’s key criterion. Next, we integrate some 3. Methods
possible criteria, the higher management would like to incorporate in
the future and support green transition. In the literature, the compara­ 3.1. Analytic hierarchy process
tive evaluation between the existing practice and the new preference
after the inclusion of the new criterion is still missing. Consequently, we The AHP analysis follows a relative, subjective comparison of the
want to explore the possible changes and affect that can bring the in­ chosen criteria (Papathanasiou et al., 2018). The analyses are conducted
clusion of additional criteria and employ two MCDM methods. We found based on semi-structured interviews with the company representatives.
a strong correlation between the Fuzzy approach and the SECA, while The following steps are defined for AHP implementation to showcase
those are implemented for the supplier selection process based on the how the investigation is conducted.
criterion used in excising practice and criteria that would likely be in­
tegrated in the future. This finding is also supported by Bahrami and Step 1. Establishment of the pair-wise comparison matrix.
Rastegar (2022), who encountered the same when the SECA method was
evaluated with others. The analysis demonstrates that SECA is closely The first step is to define the evaluated criteria. A pair-wise com­
correlated to all fuzzy approaches the authors investigated. We utilize parison matrix (A is an n × n matrix), where n is the number of criteria
both the TOPSIS and SECA models in software to include what the au­ and the element of the matrix aij = 1∀i = j, is constructed to get the
thors expect is a big issue within companies in their supplier selection criteria priority value (CPV). This is done using the 1–9 preference scale
process, variety. We showcased in a practical setting to support flexible, (1 equal importance of both elements and nine absolute primacy of one
sustainable supplier selection through the ESTA software. Developing aspect over another). The scores illustrate the importance of each
such module software aims to provide higher explainability, insight, and alternative. The pair-wise comparison must be developed carefully and
flexibility to conduct sensitivity for the DMs, significantly improving with limited use of extremely small or enormous preference scales.

2
A. Rasmussen et al. Cleaner Engineering and Technology 12 (2023) 100590

Table 1
Various criteria used in the literature.
Criteria A1 A2 A3 A4 A5 A6 A7 A8 A9 A10 A11 A12 A13 A14 A15 A16

Cost x x x x x x x x x x x x x x x x
Quality x x x x x x x x x x x x x x x x
Delivery x x x x x x x x x x x x x x x x
Worker safety and health x x x x x x x x
Technology x x x x x x x x x x x x x
Flexibility x x x x x x x x x
Environmental Affairs x x x x x x x x x
Financial Stability x x x x x x x x x x
Reliability x x x x x x x x x
Risk x x x x x
Packaging and Transport Quality x x x x x x x
Production capacity x x x x x
Location x x x x x x x x x
Communication System x x x x x x x x x
Repair Service x x x x x x x
R&D x x x x x x
Service x x x x x x x x x
Repair Service x x x x x x x x
Market Position x x x x x
Warranty/Claims x x x x
EMS x x
Green Supply Chain x x x x x
Suppliers of supplier x x x
Worker Dismissal x x
Response Speed x x
Lead Time x x x x
Past Experience x x
Reputation x x x x x
Building And Facility x x x
Relationship x x x x
Expiration Date x
Regulatory Compliance x x x
Payment Terms x x x x x
Waste Management x x x
Waste Handling x x x
Efficient Material Handling x x
Supplier Capacity x x x x
Management and Organisation x x
Attitude x x x x
Commercial Plans x
Process Improvement x
Product Development x
Professionalism x x x x
Green Manufacturing System x x x
Green Image x x x
Cooperation x x x x
Performance History x x x
Training Aids x x
Electronic Data Interchange (EDI) x x
Culture x x
Trade Restriction x x
Skill level of staff x x x
Self-audits x
IT Standards x x
Emergency orders x x
Order cycle time x x
Sales Support x

A1 = (Arabsheybani et al., 2018), A2 = (Stević, 2017), A3 = (Taherdoost and Brard, 2019), A4 = (Rouyendegh et al., 2020), A5 = (Mohammed et al., 2019), A6 =
(Weber et al., 1991), A7 = (Kar and Pani, 2014), A8 = (Ho et al., 2010), A9 = (Utama, 2021), A10 = (Stević et al., 2020), A11 = (Deng and Chan, 2011), A12 = (Wu et al.,
2022), A13 = (Hamdan and Cheaitou, 2017), A14 = (Scott et al., 2015), A15 = (Fagundes et al., 2021), A16 = (Fallahpour et al., 2017).

Step 2. Eigenvector consistent the relationship matrix. λmax equals the individual sum of the
vectors for each weight in the relationship matrix. The Consistency Ratio
This step includes computing the priority vector of criteria to identify Consistency Index (CI)
is calculated as CR = Random λmax − n
consistency index (RCI) where CI = n− 1 and RCI =
the numerical weights (w1, w2, … …, wn) of the alternatives, where 1.98(n− 2)
∑n . If CR ≤ 0.10, the
matrix is considered consistent. If not, the DMs
i=1 wi = 1.
n
will need to revise the relationship matrix.
Step 3. Consistency test
3.2. Combined F-AHP analysis
A consistency test is conducted to ensure that the calculated values
and criteria weights are consistent. The first step is to identify λmax which To ease the assessment, a Fuzzy AHP analysis is conducted. The DMs
is treated as an eigenvalue problem. The closer λmax is to n, the more will only need to relate to linguistic variables, which is beneficial in this

3
A. Rasmussen et al. Cleaner Engineering and Technology 12 (2023) 100590

case due to a lack of understanding. The linguistic terminology is seen in while maximizing the distance to the FNIS, such that: ̃
CCi =
Appendix Table B1. To aggregate the decision, we translated the deci­ ̃di ,i=1,2,…,m.
−

sion matrix into fuzzy numbers, as suggested by (Sun, 2010), we used + −

̃di +̃di
1 2 n
the following relation: ̃
aij = (̃
x ij ⊕̃
x ij ⊕… ⊕̃
x ij ). To calculate the fuzzy
1 2 n
geometric mean, we used: ̃rij = (̃ a i1 ⊕̃
a i2 ⊕… ⊕̃a in ), which is trans­ 3.4. Simultaneous evaluation of criteria and alternatives (SECA)
lated to the fuzzy weights through the following relation:
̃ i = ( ̃r i1 ⊕ ̃r i2 ⊕ … ⊕ ̃r in )− 1
w SECA method is not dependent on qualitative input in the weighing
In this study, we conduct two sets of analysis: (i) based on the present of criteria and hence remains unbiased in the weight determination. The
practice of the case company by considering the following four criteria: method evaluates the scenarios through a multiobjective non-linear
Quality (C5 ); Cost (C7 ); Relationship (C8 ); Lead Time (C9 ), and (ii) by programming model and determines the weights based on the stan­
considering the following nine criteria: Flexibility (C1 ); Financial Sta­ dard deviation and correlation in a decision matrix (Das et al., 2022).
bility (C2 ); Sustainability (C3 ); Technology (C4 ); Quality (C5 ); Delivery Step 1. We start with the decision matrix X = (xij )m×n and all the
(C6 ); Cost (C7 ); Relationship (C8 ); and Lead-Time (C9 ) based on the criteria are divided into two subcategories: beneficial criterion (BC) and
discussion with sponsor company employees, which is also supported by non-beneficial (NC). BCs have a positive effect, and growth in their
literature from Table 1. values leads to the improvement of the decision-making function,
whereas NC has a negative effect, and growth in their values has a
3.3. Fuzzy technique for order of preference by similarity to ideal solution reverse effect on the objective function. Note that all the criteria used in
(TOPSIS) our model are BC. However, the normalized decision matrix (XN) is
determined by using the following formula:
( )
A Fuzzy TOPSIS analysis is incorporated to rank alternatives based
X N = xNij ,
on the AHP analysis criteria as the supplier selection engine. The Fuzzy m×n

TOPSIS analysis makes sense as it compares the alternatives by inte­ ⎧ xij

grating incomplete and uncertain information and using qualitative ⎪ if j ∈ BC
⎨ maxk xkj
⎪
assessments in a matter where the ideal solution is considered. In sup­ where, xNij =
plier selection, everything is a trade-off, and the ability to create weights ⎪ mink xkj
⎪
⎩ if j ∈ BC
that challenges the ideal solutions seems a great fit (Altintas and Utlu, xkj
2021). Similar to F-AHP, each criterion is evaluated in a linguistic, The elements of the weighted normalized matrix for the SECA
qualitative manner, conducted through interviews. According to lin­ method are presented in Tables C1 and C.2.
guistic terminology, the fuzzy set is presented in Appendix Table B2.
This paper follows the methodology from Sun (2010) to interlink AHP Step 2: Determine the correlation between each pair of criteria (πjk )
with TOPSIS. Such integration of F-AHP methods with other MCDM and the standard deviation (σj , j = 1, 2, …, n) for each criterion to
methods such as MOORA is also familiar in the literature (Singh et al., obtain the variation information within and in between criteria (see
2022). Some steps are aggregated for simplicity, but the overall Tables C3 for correlation coefficient and Tables C4 for standard
approach remains identical. The decision matrix is the resulting Fuzzy deviations).
ratings from a workshop where the purpose was to establish the current Step 3: Compute the conflict between each criterion against other
∑
performance of suppliers. The steps are as follows: criteria (π j ), where π j = nj=1 (1 − πjk ). Note that an increase in the
variation within the criterion intensifies the objective importance of
1. Determine the weighting of criteria is where DMs assess the that criterion.
weighting of evaluation criteria in relation to the linguistic Step 4: Normalized the σ and πj as the reference points by using the
terminology. following relations:
2. Construct decision matrix is where the decision matrix is deter­
⎧ N ∑σ j
mined based on the linguistic, triangular set and aggregates the an­ ⎪ σj = n
⎪ σk
swers from different DMs.xij= xij1 ⊕ ̃
̃ 1k (̃ xijk ⊕ …xij̃K ), where ̃
xijk equals ⎨
(1)
k=1

the performance rating for the k th expert for the jc criteria as ̃

xijk = ⎪
⎪ π
⎩ πNj = ∑n j
(lij mij, uij .
̃
k,
̃ k
̃ k) k=1
πk
3. Normalization of the decision matrix is the max/min normaliza­ We refer to Tables C5 and C.6 for the detail.
tion of criteria ratings. From the decision matrix (R), we determine
the normalized decision matrix as: ̃R = [̃rij ]m×n , i = 1, 2…, n; j = 1, 2. Step 5: Finally, the weights (wj) are determined by solving the
( )
⃒ following non-linear optimization problem
j = maxi {uij i = 1,2,…,n}. For the
.., n where rij = uy+ , u+y , u+y and u+
l m u ⃒
j j j ⎧
⎪
weighted, normalized decision matrix ( ̃V), the formula is: ̃V = ⎪
⎪
⎪
⎪
⎪
vij ]m×n , i = 1, 2, ..., m; j = 1, 2, …, n where ̃
[̃ vij = ̃rij ⊕ w
̃i. ⎪
⎪
⎪
Max Z = λa − β(λb + λc )
⎪
4. The fuzzy positive ideal solution (FPIS) and fuzzy negative ideal ⎪
⎪
⎪
⎪ s.t.
⎪
solution (FNIS) is the calculation of the best possible outcome for ⎪
⎪
⎪
∑n
⎪
⎪ λa ≤ Si , where Si = ωj xNij ∀∀i ∈ {1, 2, …, m}
each (FPIS A+) and the negative counterpart (FNIS A− ). We define ⎪
⎪
⎨
j=1

them as: A+ = (̃ vl , ̃
+ +
vj , …, ̃
vn ) and A− = (̃
+
vl , ̃
−
vj , …, ̃
−
vn ).
− ∑n ( )2
(2)
⎪
⎪ λb = ωj − σnj ∀i ∈ {1, 2, …, n}
5. Distance to ideal solution calculates the distance from each crite­ ⎪ j=1
⎪
⎪
⎪
⎪ ∑n ( )2
+ − ) ⎪
rion to the best and worst outcomes. These distances (̃ di and ̃ di
n
⎪
⎪
⎪
λc = j=1
ωj − πj ∀i ∈ {1, 2, …, n}
∑ ⎪
⎪
⎪
di = nj=1 d(̃
+ ∑n
from each alternative is calculated as: ̃ vj ), i = 1, 2, …
vij ,̃ ⎪
+
⎪
⎪ ωj = 1, ε ≤ ωj ≤ 1, ∀j ∈ {1, 2, …n}
∑ ⎪
⎪ j=1
− n ⎪
, m; j = 1, 2, …, n and ̃
di = j=1 d(̃ vj ),i = 1, 2,…,m; j = 1, 2,…,n
vij ,̃ ⎪
−
⎪
⎩
6. Obtain Closeness Coefficient (CC) is a measure from which the
ranking can be derived when minimizing the distance to the FPIS In Equation (2), the objective is to maximize the performance of each

4
A. Rasmussen et al. Cleaner Engineering and Technology 12 (2023) 100590

criterion by considering the effects of the overall performance score of 4.1. Practical implementation
each alternative criterion (λa ), and variation within and between criteria
through (λb ) and (λc ), respectively. Through the process, the coefficient While MCDM and supplier selection is a much-visited territory in
for aggregation β (β ≥ 0) for all three measures (λa , λb , and λc ) is used. academia, the practical implementations of such are not. Authors seem
Note that the constraint introduced for weights ensures that the sum of to infer that existing analyses of MCDM application for supplier selection
weights should be equal to a unit, and a lower non-negative bound (ε = are relatively less concerned about how companies might utilize these
0.001) is used for a lower limit of each criterion. We determine the methods in a practical setting. Definitely, the supplier selection process
optimal value of β through sensitivity analysis, as shown in Tables C7. is context-dependent, and extracting structured data is a real problem.
More specifically, defining sustainability and quantifying such measures
4. Results is still missing in the A&D sectors. Therefore, we emphasize how DMs
could use MCDM in the supplier selection process with little to no
Currently, the criteria used at the company are Cost, Relationship, knowledge about the process through prototype software. The func­
Lead-Time, and Quality. The relationship, defined as the individual tionalities will be showcased through a constraint-based software called
preference of a purchaser as a result of their relationship with any given Expert System Shell for Text Automation, or ESTA (He et al., 2019). The
supplier, is deemed the most important. This is because the DMs will intent of ESTA is to aid DMs by establishing a knowledge base and an
disregard a small cost, quality, and lead time difference in selecting the expert system. The constraint-based nature of the software allows for
preferred supplier. The relationship matrix is derived from company easy implementation of intended functionalities with other software.
interviews and presented in Table A1. This relationship matrix yields a The system operators are sent in any appropriate direction based on
CI of 0.014 < 0.1, i.e., the relationship is consistent. The final weights their previous selections due to the current situation. This is why ESTA
are obtained as follows: Lead Time- 9.8%, Quality-16.5%, Cost- 33.7%; makes sense in this application, as MCDM requires a knowledge base
and Relationship- 40.0%. The corresponding ranking is presented in that ESTA can mimic. An assumption in MCDM methods, weight
Table 2. Similarly, we compute the ranking for the supplier based on establishment, is that the weights are static, which is not the case in a
Fuzzy-TOPSIS and SECA methods based on nine criteria. We refer to real-world environment. Purchasers might encounter various situations
Appendices B and C, respectively, for the step-wise detail of obtaining where they must deviate from the standard set of weights, creating the
the final ranking for both methods. Note that while we determine need for dynamic weight determination. An example could be a rush
weights, both methods lead to different outcomes, as presented in order, where the lead time is to be prioritized. Notably, the fuzzy
Table 2. approach allows much freedom in the designation of weights, thus
Clearly, MCDM selects a better supplier. This is primarily due to S2 allowing for the development and alignment of weights for individual
performing much better in what the literature suggests to be essential scenarios. For rush orders, the DMs should be able to select a model
criteria. It is found that the rankings change depending on the weights, where the weight of lead time is high, at the trade-off of other criteria.
but not enough to disregard performance. This indicates a healthy The intended scenarios, weights, and results are depicted in Table 4.
sensitivity in the model(s). S2 would be the preferred supplier of both When comparing the result in Table 4 to the decision matrix in
MCDM models, while the company’s preferred supplier would be S8. Table B6, it is apparent that the manipulation of weights works as
intended and thus can be included in the proposed software to account
for sensitivity analysis of highlighting possible options in a dynamic
setting. The intention is to create a database that interconnects suppliers
and their ability to deliver specific part numbers with performance data,
to select the best supplier. This database should be connected to front-
Table 2 end software, which should be developed. The decision tree and pro­
The CCi and final ranks for each method. posed software functionality through ESTA can be seen in Fig. 1.
Fuzzy-TOPSIS SECA AS-IS As an illustration of the functionalities, a decision tree is constructed
CCi Rank CCi Rank CCi Rank
to the left in Fig. 1. The decision tree represents the functionalities of the
software. Each step represents a choice in ESTA software, which is
S1 0.641 4 0.609 3 0.672 2
showcased to the right (Fig. 1b). The system recognizes suppliers’ parts,
S2 0.792 1 0.711 1 0.589 4
S3 0.728 3 0.676 2 0.396 10 scenarios, and performance to calculate the CC and ranks suppliers
S4 0.595 7 0.495 8 0.671 3 accordingly. This allows DMs to utilize the system while not having to
S5 0.618 5 0.512 7 0.429 9 consider MCDM methods in detail and thus will enable managers to
S6 0.562 8 0.542 6 0.485 7 control how the process is conducted.
S7 0.752 2 0.558 5 0.473 8
S8 0.614 6 0.600 4 0.687 1
S9 0.553 9 0.426 10 0.554 5
S10 0.470 10 0.433 9 0.530 6 4.2. Managerial implications
Note that we compute the Spearman correlation coefficient (SCC) to investigate
the difference among ranks (Akoglu, 2018). The correlation coefficient between While the previous section regards a possible solution, it is also
the TOPSIS and SECA is obtained as ρ{Fuzzy, SECA} = 0.84. Therefore, it indicates a important to investigate how such developments would need to be
strong correlation. This seems like a positive effect of working with performance implemented. As mentioned, several prerequisites exist for successfully
data and only changing weights, as the decision matrix remains a potent and implementing and using these methods. This implementation will be
decisive factor. Considering the same performance data, the mismatch between split into system prerequisites and human interaction. System pre­
the analyses and the company AHP looks profound. It is found that ρ{TOPSIS, requisites in this context mean what needs to be done for the proposed
Company} = − 0.139 and ρ{SECA, Company} = 0.055. This means that even though it is system to work properly. The authors suggest a step-wise implementa­
applied to the same performance data, the company selects significantly
tion to test the waters and then scale to the rest of the suppliers. These
different than the MCDM methods recommended. Therefore, this supports the
considerations will be based on Karandikar and Nidamarthi (2007),
study hypothesis that the company’s current setup/practice needs to be evalu­
ated further. For both the TOPSIS and SECA methods, S2 gets a higher prefer­ where the authors investigated how to implement standardization ini­
ence, whereas, for the sponsored company, it is S8. From Table 3, presented tiatives successfully. Our view on how to adapt the line of thinking to the
below, we quantify the performance of suppliers with respect to the criterion sponsor company is presented below:
before applying the MCDM methods; then, we can see the importance of weight
setting for the newly included criterion. 1. Create consensus on internal benefits and customer value

5
A. Rasmussen et al. Cleaner Engineering and Technology 12 (2023) 100590

Table 3
Linguistic ratings of supplier performance.
C1 C2 C3 C4 C5 C6 C7 C8 C9

S1 Goo Poor Good Good Fair Fair Good Good Good

S2 Fair Good Good Fair Good Very Good Very Good Good Fair
S3 Good Good Good Very Good Good Good Fair Very Good Poor
S4 Good Very Poor Poor Poor Very Good Fair Fair Good Good
S5 Good Poor Fair Good Very Good Fair Poor Good Fair
S6 Very Good Very Poor Poor Fair Poor Very Good Good Fair Good
S7 Good Good Fair Very Good Very Good Very Poor Very Good Fair Fair
S8 Very Poor Very Good Very Good Good Good Fair Poor Good Very Good
S9 Very Poor Fair Fair Fair Poor Fair Good Very Good Fair
S10 Fair Very Poor Poor Very Poor Fair Very Good Poor Good Good

training, and the Case Company would probably need to initiate change
Table 4
management initiatives to make the purchasers own the new process.
The scenarios, weights, and results. The results are shown as CCi (Rank).
Cameron and Green (2015) suggested that solid leadership communi­
Criteria Lead-Time Exceptional Quality Low Cost Sustainability cation will help the purchasers own the changes and early preparation
C1 5.0% 5.0% 5.0% 5.0% and incorporation in the change.
C2 5.0% 5.0% 5.0% 60.0%
C3 5.0% 5.0% 5.0% 5.0%
5. Discussion
C4 5.0% 15.0% 5.0% 5.0%
C5 5.0% 50.0% 5.0% 5.0%
C6 25.0% 5.0% 5.0% 5.0% The literature on the association between sustainability and the de­
C7 5.0% 5.0% 60.0% 5.0% fense industry is often neglected and sometimes not compatible with the
C8 5.0% 5.0% 5.0% 5.0% Environmental, Social, and Governance criteria used to define sustain­
C9 40.0% 5.0% 5.0% 5.0%
ability in the context of other sectors. The technology integration or
Alternatives Lead-Time Exceptional Quality Low Cost Sustainability production of defense material represents a great responsibility and
S1 0.630 (5) 0.533 (7) 0.685 (3) 0.369 (6) investment; and is subject to strict regulations and financial constraints.
S2 0.665 (3) 0.698 (5) 0.861 (1) 0.759 (2) It is phenomenal that many financial institutions are reluctant to support
S3 0.497 (7) 0.757 (2) 0.527 (6) 0.759 (3) defense activities; they sometimes implement their own internal policies
S4 0.561 (6) 0.657 (6) 0.412 (7) 0.187 (9)
that limit cooperation with the defense industry. Therefore, the debate
S5 0.468 (8) 0.757 (2) 0.242 (9) 0.332 (7)
S6 0.688 (2) 0.278 (9) 0.624 (4) 0.198 (8) about sustainable finance often lumps together with some suppliers’
S7 0.399 (10) 0.828 (1) 0.824 (2) 0.727 (4) categories, severely affecting competitiveness and financial viability
S8 0.717 (1) 0.698 (4) 0.291 (8) 0.845 (1) (Massa, 2022). Note that supplier selection criteria and Total Cost of
S9 0.422 (9) 0.249 (10) 0.594 (5) 0.465 (5)
Ownership (TCO) are interlinked. TCO provides many benefits such as it
S10 0.630 (4) 0.320 (8) 0.164 (10) 0.144 (10)
helps clarify and define supplier performance expectations both in the
manufacturer and supplier over time (Ellram, 1995). The perspective of
This paper seeks to validate that the current process is inefficient and selecting the proper supplier is to mitigate the total direct and indirect
that utilizing MCDM as a flexible standardization tool will reduce hidden costs associated with purchasing a part rather than just the acquisition
costs. The next step is for the case company to adapt to this thinking line cost (Dogan and Aydin, 2011). As we found in Table 4, there are
and establish a business case. Therefore, the first step would be under­ somehow links between the ranking of SECA, TOPSIS, and TCO, and the
standing the value and verifying the potential based on the data. assessment of whether an increase happens from selecting the proposed
supplier is possible. For example, one key factor that affects defense bids
2. Agree on guiding principles is the relationship (Emmanuel-Ebikake et al., 2014). Results also reflect
the effect.
These guiding principles could roughly be translated to the models, In addition, a standardized approach to sustainable supplier selec­
criteria, and weights in MCDM. Managers within the case company must tion is achieved by incorporating flexibility in weights to account for
agree and align the criteria, weights, performance data, and models that real-life variety (Fallahpour et al., 2017). The proposed software func­
the proposed revamped process should use. tionalities will enable the purchasers to utilize MCDM methods in
practice without having to conduct a new MCDM analysis at each pur­
3. Create sales strategy chasing scenario. Managerial implications of this implementation boil
down to system prerequisites and human interaction to aid the case
Standardization is primarily an internal benefit. However, the case company in proper, verified implementation. However, there are some
company started early in the process, and the standardization would risks in utilizing this approach. For the Fuzzy TOPSIS, the inherent risk is
enable them to create documentation on the process that could ulti­ the information bias in assigning weights. For SECA, an inherent risk is
mately become an order winner in the A&D industry. Therefore, they data and lack of flexibility. The inherent risk for the sponsor company to
should investigate how implementing this standardization could benefit keep its current process is the lack of a standardized and high TCO. The
them from a sales point of view. point is that either way, sustainable supplier selection is a process of
high risk and possibly high reward (Arabsheybani et al., 2018). There­
4. Technical implementation fore, a standardized, data-driven approach is beneficial. The risks of not
aligning and standardizing, meaning keeping the current process by
This is where the sponsor company implements the proposed soft­ excluding the criterion such as Sustainability (C1) or Financial Stability
ware prototype and integrates it into its present system before defining (C2), are much riskier than the one of utilizing a unified strategy with an
functionalities clearly to the DMs. additional amount of criterion. In this direction, since there is no
The other aspect is humans. The DMs that currently select the sup­ possible way to calculate the actual savings, it adds additional risk as the
pliers would need to adapt to a new system. To do so, they will need solutions might be redundant (Ershadi et al., 2021). In addition, the

6
A. Rasmussen et al. Cleaner Engineering and Technology 12 (2023) 100590

Fig. 1. Implementation of ESTA

overall impact of implementation across all suppliers and part numbers developing new systems and technologies and supporting the stan­
remains unquantifiable with respect to the sponsor company. Therefore, dardization of tasks by paying attention to the factors such as the digi­
we conduct a qualitative analysis based on the present situation at the talization of paper works, the design of energy consumption
company and the ranking recommended by two MCDM methods. The management policies, and others. To address this chllanging task, the
results reflect that our recommendation also brings direct benefit to in Defense industry should divert a significant number of resources to
perspective of TCO. “green R&D”. Technological innovation is the keyway to making the
Optimizing supplier selection allows the selection of the greener green transition possible, not only in the military but in all industrial
supplier and optimizes the selection process. A better process requires areas sectors (Massa, 2022).
minimum resources; thus, companies will gain the optimum outcome As a result of the current global situation, sustainability will inevi­
with fewer resources. This is also a sustainability improvement, which tably become a deciding factor in tomorrow’s supply chains. That is,
inevitably becomes more achievable by incorporating sustainable sustainability might be the deciding variable in whether companies
MCDM supplier selection. The reason is that MCDM supplier selection thrive or diminish. The Case Company is in the A&D sector, closely
allows for blending sustainability measures and criteria without linked to governments and their sustainability goals. Thus, it is even
compromising business objectives or profitability more than absolutely more critical. But how should companies cope? The authors would argue
required. This should translate into companies being more willing to that sustainable supplier selection is an ideal place to start, as every
adapt to the green agenda in the future, with the added benefit of being decision regarding product supply significantly impacts emissions. For
front-runners in the worldwide sustainability race and the advantages the Case Company, this means that they will have a direct negative
that follow, at a cost minimum. impact on the buyer’s Sustainable Development Goals (SDGs) if they do
In a recent report by Bowcott et al. (2021), it is estimated that the not take action. This does not mean that companies would need to focus
armed forces, and more specifically, the defense departments, are only on sustainability, but it should be included in the decision to secure
generating almost 50% of the greenhouse gas emissions of the public future endeavors. MCDM is an ideal process to balance sustainability
sector. Therefore, the industry cannot be excluded from the green with the organization’s internal goals, which is what the proposed so­
transition. In this study, our key focus is sustainable supplier selection. lutions seek to do (Stević et al., 2020). The barriers to the adaptation of
The components suppliers for the defense industry also need to ensure sustainability in the context of the A&D industry are many. McKinsey
precise emissions standards in producing and shipment of such com­ (Bowcott et al., 2021) argue that the priority of having mission-critical
ponents. The sponsored company must incorporate those measures to capabilities, long equipment life cycles, and increased focus on
comply with its strategies’ green and sustainable policies, standards, and high-emission niches will prove significant challenges in future A&D
procedures. Historically, the defense industry remains one of the key sustainability endeavors. When operating in the defense industry, there
contributors to technological innovation (Bellais, 2013). However, the are lives at stake. Revamping processes to favor sustainability might
dynamic nature of quality standard, contracts and past performance are jeopardize a fully functional system that might cost lives. Therefore, a
the key in this industry (Emmanuel-Ebikake et al., 2014). As included in vast barrier is a need to reduce emissions without making unacceptable
the UN Sustainable Development Goal (SDG) number 16: peace, secu­ trade-offs.
rity, and strong institutions are the key to the prosperity of our countries
and societies and so implicitly, thanks to the defense industry. Now, the
defense industry can enable the green transition mainly through

7
A. Rasmussen et al. Cleaner Engineering and Technology 12 (2023) 100590

6. Conclusions incorporate criteria such as sustainability or technology, but the inclu­

sion can significantly affect the current practice. In fact, during our
Supplier selection is generally a highly complex problem due to the discussion also, the authorities are much more concerned with such
sheer number of alternatives and criteria. This paper proposes a method changes and their effect on the total cost of ownership and building long-
to identify a preference for alternative suppliers through two MCDM term resilience. To continue the green transition, a lot of resources are
methods. The sponsor company’s AS-IS selection is based on four criteria needed for R&D to develop new systems and technologies that will
considered as a benchmark to compare the performance of the two shape the future shift toward a greener and Yazdani more sustainable
methods. To investigate the impact of utilizing this approach on the world. In this regard, the prototype we developed can support the
supplier selection process, a supplier base of 10 individual suppliers is evaluation of alternatives and be a concrete technological solution for
established from the sponsor company. The rankings of these suppliers supplier selection. The authors found for the initial assessment that
are identified through the application of the AHP weights in a TOPSIS implementing such models at the sponsor company would significantly
engine based on a supplier performance-based decision matrix. These impact the current system.
rankings are supported by the SECA method, which is incorporated to
validate the use of MCDM models concerning the sponsor company. The Declaration of competing interest
result shows a powerful correlation between the two MCDM models and
little-to-no correlation between the MCDM models and the current state. The authors declare that they have no known competing financial
Clearly, this indicates a potential opportunity for process optimization. interests or personal relationships that could have appeared to influence
Considerations of enabling a dynamic environment to account for real- the work reported in this paper.
life variety are established, and a prototype system (ESTA) is con­
structed to showcase the proposed functionalities of practical imple­ Data availability
mentation. The results reflect that the relationship remains a critical
criterion in the defense industry. The sponsor company is willing to No data was used for the research described in the article.

A. Present practice for the company

As discussed earlier, the company presently relies on the four criteria. Note that one of the key barriers in A&D is that the industry faces unique
challenges related to regulatory compliance and strict security protocols. Therefore, too many criteria are still not crucial like in other sectors. Thus,
the initial relationship matrix used in this study is presented in Table A1.

Table A.1
The relationship matrix derived from company interviews.

Criterion Lead Time Quality Cost Relationship

Lead Time 1 1/2 1/3 1/4

Quality 2 1 1/2 1/3
Cost 3 2 1 1
Relationship 4 3 1 1

B. Fuzzy integration in AHP and TOPSIS

In this section, we present the detail of the Fuzzy-TOPSIS method. Note that the method is based on the nine criteria instead of the four criteria. As
shown in Table 1, researchers proposed various criteria to be used in the supplier selection process till we used those additional criteria, which
appeared critical during our semi-structured interview. We use the following relationship matrix among criteria as shown Table B1 as our departure
point.

Table B.1
The linguistic relationship matrix with additional criterion

C1 C2 C3 C4 C5 C6 C7 C8 C9

C1 1/I 1/I 1/VI 1/AI 1/VI 1/AI 1/VI 1/MI

C2 MI EI 1/I 1/VI 1/I 1/VI 1/I EI
C3 I EI 1/MI 1/I 1/MI 1/I 1/MI MI
C4 VI I MI 1/MI EI 1/MI EI I
C5 AI VI I MI MI EI MI VI
C6 VI I MI EI 1/MI 1/MI EI I
C7 AI VI MI MI EI MI MI VI
C8 VI MI MI EI 1/MI EI 1/MI I
C9 MI 1/MI 1/MI 1/I 1/VI 1/I 1/VI 1/I
Similarly, we use the following triangular fuzzy number (Sun, 2010) for linguistic analysis among the performance of suppliers based on different criteria as shown in
Tables B2 and B.3.

8
A. Rasmussen et al. Cleaner Engineering and Technology 12 (2023) 100590

Table B.2
The triangular set and linguistic terminology for
assessment

Fuzzy ratings for linguistic variables

Fuzzy Number Criteria Assessments

(1, 1, 3) Very Poor (VP)
(1, 3, 5) Poor (P)
(3, 5, 7) Fair (F)
(5, 7, 9) Good (G)
(7, 9, 9) Very Good (VG)

Table B.3
Positive/negative linguistic scale for fuzzy number (Fu et al., 2020)

Statement Positive Rating Positive Fuzzy Set Negative Rating Negative Fuzzy set
)
Equally Important 1
̃ (1, 1, 3) 1 (1 1 1
, ,
1
̃ 3 1 1
)
Moderately Important 3
̃ (1, 3, 5) 1 (1 1 1
, ,
3
̃ 5 3 1
)
Important 5
̃ (3, 5, 7) 1 (1 1 1
, ,
5
̃ 7 5 3
)
Very Important 7
̃ (5, 7, 9) 1 (1 1 1
, ,
7
̃ 9 7 5
)
Absolutely Important 9
̃ (7, 9, 9) 1 (1 1 1
, ,
9
̃ 9 9 7
Next, we define the fuzzy relationship matrix to be used for weight computation. Using Table B1, we obtain the following relationship matrix as presented in
Table B4.

Table B.4
The fuzzy triangular relationship matrix for AHP.

C1 C2 C3 C4 C5 C6 C7 C8 C9

C1 1.0 1.0 1.0 0.2 0.3 1.0 0.1 0.2 0.3 0.1 0.1 0.2 0.1 0.1 0.1 0.1 0.1 0.2 0.1 0.1 0.1 0.1 0.1 0.2 0.2 0.3 1.0
C2 1.0 3.0 5.0 1.0 1.0 1.0 0.2 0.3 1.0 0.1 0.2 0.3 0.1 0.1 0.2 0.1 0.2 0.3 0.1 0.1 0.2 0.1 0.2 0.3 1.0 1.0 1.0
C3 3.0 5.0 7.0 1.0 3.0 5.0 1.0 1.0 1.0 0.2 0.3 1.0 0.1 0.2 0.3 0.2 0.3 1.0 0.1 0.2 0.3 0.2 0.3 1.0 1.0 3.0 5.0
C4 5.0 7.0 9.0 3.0 5.0 7.0 1.0 3.0 5.0 1.0 1.0 1.0 0.2 0.3 1.0 1.0 1.0 1.0 0.2 0.3 1.0 1.0 1.0 1.0 3.0 5.0 7.0
C5 7.0 9.0 9.0 5.0 7.0 9.0 3.0 5.0 7.0 1.0 3.0 5.0 1.0 1.0 1.0 1.0 3.0 5.0 1.0 1.0 1.0 1.0 3.0 5.0 5.0 7.0 9.0
C6 5.0 7.0 9.0 3.0 5.0 7.0 1.0 3.0 5.0 1.0 1.0 1.0 0.2 0.3 1.0 1.0 1.0 1.0 0.2 0.3 1.0 1.0 1.0 1.0 3.0 5.0 7.0
C7 7.0 9.0 9.0 5.0 7.0 9.0 3.0 5.0 7.0 1.0 3.0 5.0 1.0 1.0 1.0 1.0 3.0 5.0 1.0 1.0 1.0 1.0 3.0 5.0 5.0 7.0 9.0
C8 5.0 7.0 9.0 3.0 5.0 7.0 1.0 3.0 5.0 1.0 1.0 1.0 0.2 0.3 1.0 1.0 1.0 1.0 0.2 0.3 1.0 1.0 1.0 1.0 3.0 5.0 7.0
C9 1.0 3.0 5.0 1.0 1.0 1.0 0.2 0.3 1.0 0.1 0.2 0.3 0.1 0.1 0.2 0.1 0.1 0.3 0.1 0.1 0.3 0.1 0.2 0.3 1.0 1.0 1.0
Once again, applying the similar procedure as presented in Section 3, we obtain the weights for each criterion through the AHP method, as shown in Table B5.

Table B.5
The resulting weights of the F-AHP

Criterion wi

Flexibility 0.020 0.016 0.018

S & EA 0.031 0.029 0.031
Financial Stab. 0.051 0.056 0.074
Technology 0.128 0.119 0.125
Quality 0.241 0.257 0.235
Delivery 0.128 0.119 0.125
Cost 0.241 0.257 0.235
Relationship 0.128 0.119 0.125
Lead-Time 0.032 0.029 0.031
Next, we analyze the final ranking obtained through F-TOPSIS. We present linguistic ratings of each supplier
based on the nine criteria in Table 3 and corresponding numerical representation in the following Table B6.

Table B.6
A translation of the linguistic decision matrix into triangular sets.

C1 C2 C3 C4 C5 C6 C7 C8 C9

S1 579 135 579 579 357 3 57 579 579 579

S2 357 579 579 357 579 7 99 799 579 357
S3 579 579 579 799 579 5 79 357 799 135
S4 579 113 135 135 799 3 57 357 579 579
(continued on next page)

9
A. Rasmussen et al. Cleaner Engineering and Technology 12 (2023) 100590

Table B.6 (continued )

C1 C2 C3 C4 C5 C6 C7 C8 C9

S5 579 135 357 579 799 3 57 135 579 357

S6 799 113 135 357 135 7 99 579 357 579
S7 579 579 357 799 799 1 13 799 357 357
S8 113 799 799 579 579 3 57 135 579 799
S9 113 357 357 357 135 3 57 579 799 357
S10 357 113 135 113 357 7 99 135 579 579
Therefore, the normalized decision matrix is obtained as follows.

Table B.7
The normalized decision matrix for supplier performance.

C1 C2 C3 C4 C5 C6 C7 C8 C9

S1 0.56 0.78 1.00 0.11 0.33 0.56 0.56 0.78 1.00 0.56 0.78 1.00 0.33 0.56 0.78 0.33 0.56 0.78 0.56 0.78 1.00 0.56 0.78 1.00 0.56 0.78 1.00
S2 0.33 0.56 0.78 0.56 0.78 1.00 0.56 0.78 1.00 0.33 0.56 0.78 0.56 0.78 1.00 0.78 1.00 1.00 0.78 1.00 1.00 0.56 0.78 1.00 0.33 0.56 0.78
S3 0.56 0.78 1.00 0.56 0.78 1.00 0.56 0.78 1.00 0.78 1.00 1.00 0.56 0.78 1.00 0.56 0.78 1.00 0.33 0.56 0.78 0.78 1.00 1.00 0.11 0.33 0.56
S4 0.56 0.78 1.00 0.11 0.11 0.33 0.11 0.33 0.56 0.11 0.33 0.56 0.78 1.00 1.00 0.33 0.56 0.78 0.33 0.56 0.78 0.56 0.78 1.00 0.56 0.78 1.00
S5 0.56 0.78 1.00 0.11 0.33 0.56 0.33 0.56 0.78 0.56 0.78 1.00 0.78 1.00 1.00 0.33 0.56 0.78 0.11 0.33 0.56 0.56 0.78 1.00 0.33 0.56 0.78
S6 0.78 1.00 1.00 0.11 0.11 0.33 0.11 0.33 0.56 0.33 0.56 0.78 0.11 0.33 0.56 0.78 1.00 1.00 0.56 0.78 1.00 0.33 0.56 0.78 0.56 0.78 1.00
S7 0.56 0.78 1.00 0.56 0.78 1.00 0.33 0.56 0.78 0.78 1.00 1.00 0.78 1.00 1.00 0.11 0.11 0.33 0.78 1.00 1.00 0.33 0.56 0.78 0.33 0.56 0.78
S8 0.11 0.11 0.33 0.78 1.00 1.00 0.78 1.00 1.00 0.56 0.78 1.00 0.56 0.78 1.00 0.33 0.56 0.78 0.11 0.33 0.56 0.56 0.78 1.00 0.78 1.00 1.00
S9 0.11 0.11 0.33 0.33 0.56 0.78 0.33 0.56 0.78 0.33 0.56 0.78 0.11 0.33 0.56 0.33 0.56 0.78 0.56 0.78 1.00 0.78 1.00 1.00 0.33 0.56 0.78
S10 0.33 0.56 0.78 0.11 0.11 0.33 0.11 0.33 0.56 0.11 0.11 0.33 0.33 0.56 0.78 0.78 1.00 1.00 0.11 0.33 0.56 0.56 0.78 1.00 0.56 0.78 1.00
Finally, using the methodology as presented in Subsection 3.3, we determine the Fuzzy positive ideal solution (FPIS) and fuzzy negative ideal solution (FNIS) as
presented in Table B8 and B.9, respectively.

Table B.8
The FPIS scores for each alternative in relation to each criterion.

Distance from FPIS

C1 C2 C3 C4 C5 C6 C7 C8 C9 d+
i

S1 0.003 0.018 0.010 0.022 0.095 0.048 0.045 0.022 0.006 0.270
S2 0.007 0.005 0.010 0.048 0.045 0.000 0.000 0.022 0.012 0.149
S3 0.003 0.005 0.010 0.000 0.045 0.022 0.095 0.000 0.018 0.200
S4 0.003 0.022 0.035 0.075 0.000 0.048 0.095 0.022 0.006 0.306
S5 0.003 0.018 0.022 0.022 0.000 0.048 0.148 0.022 0.012 0.296
S6 0.000 0.022 0.035 0.048 0.148 0.000 0.045 0.048 0.006 0.352
S7 0.003 0.005 0.022 0.000 0.000 0.092 0.000 0.048 0.012 0.182
S8 0.013 0.000 0.000 0.022 0.045 0.048 0.148 0.022 0.000 0.299
S9 0.013 0.012 0.022 0.048 0.148 0.048 0.045 0.000 0.012 0.347
S10 0.007 0.022 0.035 0.092 0.095 0.000 0.148 0.022 0.006 0.428

Table B.9
The FNIS scores for each alternative in relation to each criterion.

Distance from FNIS

C1 C2 C3 C4 C5 C6 C7 C8 C9 d−i

S1 0.011 0.005 0.039 0.081 0.092 0.057 0.132 0.047 0.016 0.481
S2 0.007 0.018 0.039 0.057 0.132 0.095 0.162 0.047 0.012 0.569
S3 0.011 0.018 0.039 0.095 0.132 0.081 0.092 0.056 0.010 0.533
S4 0.011 0.000 0.025 0.039 0.162 0.057 0.092 0.047 0.016 0.450
S5 0.011 0.005 0.030 0.081 0.162 0.057 0.074 0.047 0.012 0.478
S6 0.013 0.000 0.025 0.057 0.074 0.095 0.132 0.039 0.016 0.451
S7 0.011 0.018 0.030 0.095 0.162 0.023 0.162 0.039 0.012 0.550
S8 0.000 0.022 0.043 0.081 0.132 0.057 0.074 0.047 0.020 0.476
S9 0.000 0.012 0.030 0.057 0.074 0.057 0.132 0.056 0.012 0.429
S10 0.007 0.000 0.025 0.023 0.092 0.095 0.074 0.047 0.016 0.379
Finally, the CCi for each alternative at any given criteria is determined, and the final ranking is presented in Table 2.

C. Step-wise computation for the SECA method

The following step-wise analysis is executed to obtain the final ranking by SECA. The first step is establishing a decision matrix, which is the same
as Table B7 from the F-TOPSIS. To assign non-fuzzy weights, each fuzzy number is aggregated into a single crisp value, and the detail is presented in
Table C1.

10
A. Rasmussen et al. Cleaner Engineering and Technology 12 (2023) 100590

Table C.1
The aggregated decision matrix.

C1 C2 C3 C4 C5 C6 C7 C8 C9

S1 0.78 0.33 0.78 0.78 0.56 0.56 0.78 0.78 0.78

S2 0.56 0.78 0.78 0.56 0.78 0.93 0.93 0.78 0.56
S3 0.78 0.78 0.78 0.93 0.78 0.78 0.56 0.93 0.33
S4 0.78 0.19 0.33 0.33 0.93 0.56 0.56 0.78 0.78
S5 0.78 0.33 0.56 0.78 0.93 0.56 0.33 0.78 0.56
S6 0.93 0.19 0.33 0.56 0.33 0.93 0.78 0.56 0.78
S7 0.78 0.78 0.56 0.93 0.93 0.19 0.93 0.56 0.56
S8 0.19 0.93 0.93 0.78 0.78 0.56 0.33 0.78 0.93
S9 0.19 0.56 0.56 0.56 0.33 0.56 0.78 0.93 0.56
S10 0.56 0.19 0.33 0.19 0.56 0.93 0.33 0.78 0.78
Next, we normalized the decision matrix by a similar approach as presented in Section 3.4, and the results are shown in Table C2.

Table C.2
The aggregated normalized decision matrix.

C1 C2 C3 C4 C5 C6 C7 C8 C9

S1 0.84 0.36 0.84 0.84 0.60 0.60 0.84 0.84 0.84

S2 0.60 0.84 0.84 0.60 0.84 1.00 1.00 0.84 0.60
S3 0.84 0.84 0.84 1.00 0.84 0.84 0.60 1.00 0.36
S4 0.84 0.20 0.36 0.36 1.00 0.60 0.60 0.84 0.84
S5 0.84 0.36 0.60 0.84 1.00 0.60 0.36 0.84 0.60
S6 1.00 0.20 0.36 0.60 0.36 1.00 0.84 0.60 0.84
S7 0.84 0.84 0.60 1.00 1.00 0.20 1.00 0.60 0.60
S8 0.20 1.00 1.00 0.84 0.84 0.60 0.36 0.84 1.00
S9 0.20 0.60 0.60 0.60 0.36 0.60 0.84 1.00 0.60
S10 0.60 0.20 0.36 0.20 0.60 1.00 0.36 0.84 0.84
Correlation coefficients are determined to capture the information from the between-criteria in the normalized decision matrix (Keshavarz-Ghorabaee et al., 2018),
and the results are shown in Table C3.

Table C.3
Correlation matrix obtained from Table C2

C1 C2 C3 C4 C5 C6 C7 C8 C9

C1 1 − 0.471 − 0.396 0.121 0.213 0.062 0.200 − 0.491 − 0.066

C2 − 0.471 1 0.795 0.630 0.315 − 0.279 0.192 0.178 − 0.436
C3 − 0.396 0.795 1 0.666 0.221 − 0.157 0.048 0.361 − 0.172
C4 0.121 0.630 0.666 1 0.306 − 0.514 0.210 − 0.053 − 0.277
C5 0.213 0.315 0.221 0.306 1 − 0.439 − 0.220 − 0.053 0.041
C6 0.062 − 0.279 − 0.157 − 0.514 − 0.439 1 − 0.139 0.175 − 0.052
C7 0.200 0.192 0.048 0.210 − 0.220 − 0.139 1 − 0.320 − 0.536
C8 − 0.491 0.178 0.361 − 0.053 − 0.053 0.175 − 0.320 1 − 0.297
C9 − 0.066 − 0.436 − 0.172 − 0.277 0.041 − 0.052 − 0.536 − 0.297 1
Next, the Standard Deviation (σ) is calculated to measure the spread of values for the ten alternative suppliers. This is shown in Table C4.

Table C.4
The Standard Deviation and their normalized values

C C1 C2 C3 C4 C5 C6 C7 C8 C9

σ 0.28 0.31 0.23 0.26 0.25 0.25 0.26 0.13 0.20

Nσ 0.13 0.14 0.11 0.12 0.11 0.12 0.12 0.06 0.09
∑m
From Table C3, the Aggregated decision matrix is determined (π j = l=1 (1 − rjl ))

Table C.5
Transformed matrix obtained by following Step 3

C1 C2 C3 C4 C5 C6 C7 C8 C9

C1 0.000 1.471 1.396 0.879 0.787 0.938 0.800 1.491 1.066

C2 1.471 0.000 0.205 0.370 0.685 1.279 0.808 0.822 1.436
C3 1.396 0.205 0.000 0.334 0.779 1.157 0.952 0.639 1.172
C4 0.879 0.370 0.334 0.000 0.694 1.514 0.790 1.053 1.277
C5 0.787 0.685 0.779 0.694 0.000 1.439 1.220 1.053 0.959
C6 0.938 1.279 1.157 1.514 1.439 0.000 1.139 0.825 1.052
C7 0.800 0.808 0.952 0.790 1.220 1.139 0.000 1.320 1.536
C8 1.491 0.822 0.639 1.053 1.053 0.825 1.320 0.000 1.297
C9 1.066 1.436 1.172 1.277 0.959 1.052 1.536 1.297 0.000
The π values are then summarized from the transformed model and normalized as Nπ illustrated in C.6.

11
A. Rasmussen et al. Cleaner Engineering and Technology 12 (2023) 100590

Table C.6
The standard deviation for conflict between each criterion against other criteria (π and Nπ) values

C1 C2 C3 C4 C5 C6 C7 C8 C9

8.83 7.08 6.63 6.91 7.61 9.34 8.57 8.50 9.80

0.12 0.10 0.09 0.09 0.10 0.13 0.12 0.12 0.13

The results presented in Table C2 are used to construct the optimization problem as presented in Step 5. we use Wolfram Mathematica to solve the
optimization problem. The weights obtained for various β values are presented in Table C7. As reported by (Keshavarz-Ghorabaee et al., 2018), the
weights should be convergent; our results are also consistent with the observation. Note that weights are used (β = 5) for the final ranking, as shown in
Table 2.

Table C.7
Various β values and corresponding weights for each criterion

β w1 w2 w3 w4 w5 w6 w7 w8 w9

0.1 0.3 0.06 0.00 0.00 0.00 0.14 0.07 0.18 0.32 0.22
0.5 0.06 0.02 0.00 0.05 0.12 0.15 0.17 0.26 0.18
1 0.06 0.05 0.04 0.05 0.09 0.17 0.14 0.22 0.16
2 0.08 0.08 0.07 0.06 0.08 0.18 0.12 0.17 0.15
3 0.11 0.09 0.08 0.08 0.10 0.16 0.11 0.13 0.14
5 0.12 0.10 0.09 0.08 0.11 0.15 0.11 0.11 0.13
0.13 0.10 0.09 0.09 0.11 0.14 0.11 0.11 0.13

References Fallahpour, A., Olugu, E.U., Musa, S.N., Wong, K.Y., Noori, S., 2017. A decision support
model for sustainable supplier selection in sustainable supply chain management.
Comput. Ind. Eng. 105, 391–410.
Akoglu, H., 2018. User’s guide to correlation coefficients. Turkish Journal of Emergency
Fu, H.H., Chen, Y.Y., Wang, G.J., 2020. Using a fuzzy analytic hierarchy process to
Medicine 18, 91–93. https://doi.org/10.1016/j.tjem.2018.08.001. URL:
formulate an effectual tea assessment system. Sustainability. https://doi.org/
https://www.sciencedirect.com/science/article/pii/S2452247318 302164.
10.3390/su12156131.
Altintas, E., Utlu, Z., 2021. Planning energy usage in electricity production sector
Hamdan, S., Cheaitou, A., 2017. Supplier selection and order allocation with green
considering environmental impacts with Fuzzy TOPSIS method & Game Theory.
criteria: an mcdm and multiobjective optimization approach. Comput. Oper. Res. 81,
Cleaner Engineering and Technology 5, 100283.
282–304.
Arabsheybani, A., Paydar, M.M., Safaei, A.S., 2018. An integrated fuzzy MOORA method
He, F., Zeng, L., Li, D., Ren, Z., 2019. Study of LED array fill light based on parallel
and FMEA technique for sustainable supplier selection considering quantity
particle swarm optimization in greenhouse planting. Information processing in
discounts and supplier’s risk. J. Clean. Prod. 190, 577–591.
agriculture 6 (1), 73–80.
Bahrami, S., Rastegar, M., 2022. Security-based critical power distribution feeder
Ho, W., Xu, X., Dey, P.K., 2010. Multi-criteria decision making approaches for supplier
identification: application of fuzzy BWM-VIKOR and SECA. Int. J. Electr. Power
evaluation and selection: a literature review. Eur. J. Oper. Res. 202, 16–24.
Energy Syst. 134, 107395.
Hwang, C.L., Yoon, K., 1981. Methods for multiple attribute decision making. In:
Bellais, R., 2013. Technology and the defense industry: real threats, bad habits, or new
Multiple Attribute Decision Making. Springer, Berlin, Heidelberg, pp. 58–191.
(market) opportunities? Journal of Innovation Economics Management 12, 59–78.
Kar, A.K., Pani, A.K., 2014. Exploring the importance of different supplier selection
Benyoucef, L., Ding, H., Xie, X., 2003. Supplier selection problem: selection criteria and
criteria. Management Research Review.
methods. INRIA. Ph.D. thesis.
Karandikar, H., Nidamarthi, S., 2007. Implementing a platform strategy for a systems
Bera, A.K., Jana, D.K., Banerjee, D., Nandy, T., 2020. Supplier selection using extended
business via standardization. J. Manuf. Technol. Manag.
IT2 fuzzy TOPSIS and IT2 fuzzy MOORA considering subjective and objective
Keshavarz-Ghorabaee, M., Amiri, M., Zavadskas, E.K., Turskis, Z., Antucheviciene, J.,
factors. Soft Comput. 24, 8899–8915.
2018. Simultaneous evaluation of criteria and alternatives (seca) for multi-criteria
Bowcott, H., Gatto, G., Hamilton, A., Sullivan, E., 2021. Decarbonizing Defense:
decision-making. Informatica 29, 265–280.
Imperative and Opportunity. McKinsey. https://www.mckinsey.com/industrie
Massa, T., 2022. The possible contribution of the defence industry to the green transition.
s/aerospace-and-defense/our-insights/decarbonizing-defenseimperative-and
In: Innovative Technologies and Renewed Policies for Achieving a Greener Defence.
-opportunity.
Springer, pp. 85–94.
Brauers, W.K., Zavadskas, E.K., 2006. The moora method and its application to
Mohammed, A., Harris, I., Govindan, K., 2019. A hybrid mcdm-fmoo approach for
privatization in a transition economy. Control Cybern. 35, 445–469.
sustainable supplier selection and order allocation. Int. J. Prod. Econ. 217, 171–184.
Cameron, E., Green, M., 2015. Making Sense of Change Management. Kogan Page
Nkuna, S.G., Olwal, T.O., Chowdhury, S.D., 2022. Assessment of thermochemical
Limited. Number ISBN: 978 0 7494 7258 0 in Book.
technologies for wastewater sludge-to-energy: an advance MCDM model. Cleaner
Das, R., Saleh, S., Nielsen, I., Kaviraj, A., Sharma, P., Dey, K., Saha, S., 2022. Performance
Engineering and Technology, 100519.
analysis of machine learning algorithms and screening formulae for β–thalassemia
Papathanasiou, J., et al., 2018. Multiple criteria decision aid. In: 0 in Book. Springer.
trait screening of Indian antenatal women. Int. J. Med. Inf., 104866
Number ISBN 978-3-319-91646-.
Deng, Y., Chan, F.T., 2011. A new fuzzy dempster mcdm method and its application in
Rouyendegh, B.D., Yildizbasi, A., Üstünyer, P., 2020. Intuitionistic fuzzy TOPSIS method
supplier selection. Expert Syst. Appl. 38, 9854–9861.
for green supplier selection problem. Soft Comput. 24 (3), 2215–2228.
Deshmukh, A.J., Chaudhari, A.A., 2011. A review for supplier selection criteria and
Roy, B., 1968. Classement et choix en présence de points de vue multiples. Revue
methods. In: Technology Systems and Management. Springer, pp. 283–291.
française d’informatique et de recherche opérationnelle 2, 57–75.
Dogan, I., Aydin, N., 2011. Combining bayesian networks and total cost of ownership
Saaty, T.L., 1988. What is the analytic hierarchy process? In: Mitra, G., Greenberg, H.J.,
method for supplier selection analysis. Comput. Ind. Eng. 61, 1072–1085.
Lootsma, F.A., Rijkaert, M.J., Zimmermann, H.J. (Eds.), Mathematical Models for
Duckstein, L., Opricovic, S., 1980. Multiobjective optimization in river basin
Decision Support. Springer Berlin Heidelberg, Berlin, Heidelberg, pp. 109–121.
development. Water Resour. Res. 16, 14–20.
Sayadi, M.K., Heydari, M., Shahanaghi, K., 2009. Extension of vikor method for decision
Ellram, L.M., 1995. Total cost of ownership: an analysis approach for purchasing. Int. J.
making problem with interval numbers. Appl. Math. Model. 33, 2257–2262.
Phys. Distrib. Logist. Manag.
Scott, J., Ho, W., Dey, P.K., Talluri, S., 2015. A decision support system for supplier
Emmanuel-Ebikake, O., Roy, R., Shehab, E., 2014. Supplier sustainability assessment for
selection and order allocation in stochastic, multi-stakeholder and multi-criteria
the UK defence industry. Int. J. Prod. Perform. Manag.
environments. Int. J. Prod. Econ. 166, 226–237.
Ershadi, M., Jefferies, M., Davis, P., Mojtahedi, M., 2021. Barriers to achieving
Singh, S., Kawade, S., Dhar, A., Powar, S., 2022. Analysis of mango drying methods and
sustainable construction project procurement in the private sector. Cleaner
effect of blanching process based on energy consumption, drying time using multi-
Engineering and Technology 3, 100125.
criteria decision-making. Cleaner Engineering and Technology 8, 100500.
Fagundes, M.V.C., Keler, Á.C., Teles, E.O., de Melo, S.A.B.V., Freires, F.G.M., 2021.
Stević, Ž., 2017. Criteria for supplier selection: a literature review. International Journal
Multicriteria decision-making system for supplier selection considering risk: a
of Engineering, Business and Enterprise Applications 19, 23–27.
computational fuzzy ahp-based approach. IEEE Latin America Transactions 19,
Stević, Ž., Pamučar, D., Puška, A., Chatterjee, P., 2020. Sustainable supplier selection in
1564–1572.
healthcare industries using a new mcdm method: measurement of alternatives and
ranking according to compromise solution (marcos). Comput. Ind. Eng. 140, 106231.

12
A. Rasmussen et al. Cleaner Engineering and Technology 12 (2023) 100590

Sun, C.C., 2010. A performance evaluation model by integrating fuzzy ahp and fuzzy Weber, C.A., Current, J.R., Benton, W., 1991. Vendor selection criteria and methods. Eur.
topsis methods. Expert Syst. Appl. 37, 7745–7754. J. Oper. Res. 50, 2–18.
Taherdoost, H., Brard, A., 2019. Analyzing the Process of Supplier Selection Criteria and Wu, C., Gao, J., Barnes, D., 2022. Sustainable partner selection and order allocation for
Methods. Procedia Manufacturing. strategic items: an integrated multi-stage decision-making model. Int. J. Prod. Res.
Utama, D., 2021. AHP and topsis integration for green supplier selection: a case study in 1–25.
Indonesia. J. Phys. Conf., 012015. IOP Publishing. Yoon, K.P., Hwang, C.L., 1995. Multiple Attribute Decision Making: an Introduction.
Wang, C.N., Tsai, H.T., Ho, T.P., Nguyen, V.T., Huang, Y.F., 2020. Multi-criteria decision Sage publications.
making (mcdm) model for supplier evaluation and selection for oil production
projects in vietnam. Processes 8, 134.

13