MCDM Applications for Sustainability Assessment in Textile and Apparel Supply Chains: A Systematic Review of Methods, Contexts, and Criteria

1. Introduction

The textile industry plays a vital role in meeting basic human needs, generating employment, and supporting economic development. In recent decades, the textile supply chain has become increasingly globalized and complex, involving multiple stakeholders across raw material production, manufacturing, distribution, retail, and consumption. Alongside rapid growth in textile production and demand, concerns regarding the sustainability of the industry have intensified [1,2,3]. The textile supply chain is associated with substantial environmental impacts, including high resource and energy consumption, greenhouse gas emissions, water use, pollution from dyeing and finishing processes, microfiber release, and large volumes of textile waste [4,5,6,7,8,9]. In parallel, the industry faces persistent social challenges such as child labor, unsafe working conditions, inadequate wages, gender inequality, and weak labor protections in several production regions [10]. These sustainability issues occur across different stages of the supply chain and involve stakeholders with diverse and often conflicting priorities.

Addressing such challenges requires sustainability management approaches capable of balancing economic, environmental, and social considerations simultaneously. In this context, Multiple Criteria Decision Making (MCDM) methods have gained increasing attention as tools for supporting complex decision-making processes. MCDM methods enable the evaluation and prioritization of alternatives based on multiple quantitative and qualitative criteria, making them particularly suitable for sustainability-related problems where trade-offs between competing objectives are unavoidable [11]. Because supply chain management inherently involves multiple actors, conflicting objectives, and diverse evaluation criteria, sustainable textile supply chain management can be understood as a multi-criteria decision-making problem [12,13]. Consequently, a growing body of research has applied MCDM methods to support decisions related to sustainable supplier selection, sustainability performance assessment, material and technology selection, circular economy strategies, and identification of critical sustainability factors in textile industries [12,14,15].

Although previous reviews have examined MCDM applications in supply chain management and sustainability separately [12,15], there remains a lack of a systematic literature review focusing specifically on the intersection of sustainability assessment, textile supply chains, and MCDM methods. Existing reviews also provide limited insight into how sustainability criteria are selected and prioritized across different textile contexts and scope of studies. To address this gap, this study conducts a systematic literature review of MCDM applications in sustainability management within the textile supply chain. Specifically, the review seeks to answer the following research questions:

RQ1. Which MCDM methods have been used for sustainability assessment in the textile industry, and in which study scopes are they applied to evaluate sustainability criteria?

RQ2. What sustainability criteria and factors have been used and prioritized in MCDM-based assessments of textile industry practices?

Following the systematic literature review framework proposed by Durach, Kembro, and Wieland [16], this study categorizes the reviewed literature according to application context (scope of study), unit of analysis, methodological approach, and sustainability criteria. By doing so, the review contributes to a clearer understanding of how MCDM methods support sustainability decision-making across different settings in the textile supply chain.

2. Materials and Methods

2.1. Theoretical Framework – Textile Industry Towards Sustainability

One of the most widely adopted conceptualizations of sustainability is the Triple Bottom Line (TBL), which frames sustainability through economic, environmental, and social dimensions. However, it has been noted that the TBL concept did not emerge from a single theoretically rigorous definition, but rather evolved through interdisciplinary critiques of the economic status quo from ecological and social perspectives [17]. Therefore, sustainable development requires balancing competing and often non-comparable economic, ecological, and sociocultural objectives, including income generation, ecosystem preservation, equity, and participation [18]. These interconnected sustainability dimensions are particularly relevant to the textile industry, where economic pressures for low-cost and rapid production often conflict with environmental concerns such as resource consumption, pollution, and waste generation, as well as social issues related to labor conditions and equity across global supply chains. While this review categorizes sustainability criteria according to the traditional economic, environmental, and social dimensions, several studies extend sustainability assessment to additional domains such as governance and technology. Namely, the central role of governance and politics in enabling sustainability transitions and the importance of technology as a key factor influencing sustainable business performance have been highlighted along with their interaction with the classical sustainability dimensions [19,20].

The textile industry serves the societies by providing essential human needs, creating employment opportunities, and contributing to economic development. In an increasingly globalized economy, the textile supply chain has evolved in the form of a multi-layer and complex system with stakeholders in different parts of the world all contributing to the growing textile production. The global capita textile production increased from 5.9 kg to 13 kg per year between 1975 and 2018 [1]. World textile demand report by International Cotton Advisory Committee predicted the global demand for textile fibres to continue growing, reaching 117.3 million tons in 2025 [2]. This growth was projected after a significant decline in 2020 due to supply chain disruption, widespread lockdowns, and decrease in consumers’ spending that was followed by a strong recovery in 2021 and a slight dip in 2022 [2]. Looking into raw material production for the textile industry, Textile Exchange reports that the global textile fibre production increased by 7%, rising from 116 million tons in 2022 to 124 million tons in 2023 [3]. According to the report, if current trends continue, global fibre production is expected to reach 160 million tons by 2030. This growth is associated with significant environmental impacts including extensive resource and energy use, emissions to the environment, water and land use, etc. along with increasing concerns in the social domain such as child labor, safety at work, fair wages, among others.

Sustainability assessments are essential for evaluating impacts, determining progress toward targets, identifying the hotspots in the textile supply chain and strategizing towards SDGs. Sustainability assessment methods including Environmental and Social Life Cycle Assessment (e-LCA and s-LCA), standards and reporting systems like ISO 14001 series and GRI have been widely used by researchers and stakeholders in the textile supply chain for evaluating of the impacts, reporting and improving sustainability, scoring materials, processes, and supply sources [21].

As reported in multiple sources [4,5,6,7,8], and summarized in [9], the industry generates 8-10% of global CO₂ emissions, equivalent to 4-5 billion tons annually, consumes approximately 79 trillion litters of water per year and contributes to about 20% of industrial water pollution from textile treatment and dyeing. The sector contributes around 35% of primary micro-plastic pollution in the oceans, releasing an estimated 190,000 tons annually and also produces more than 92 million tons of textile waste each year, much of which ends up in landfills, is incinerated, or discarded as unsold products [9]. Polyester and cotton are dominant fibre materials in the textile sector contributing to 77% of global textile fibre consumption [3]. Therefore, comparing the environmental impacts of these two textile raw materials illustrates the complexity of environmental impact assessments. Respective life cycle impacts within the production phase for 264 g of fibre shows that while cotton has a slightly higher carbon footprint (1.2 kg CO2-eq vs. 1.1 kg CO2-eq for polyester), it requires significantly less non-renewable energy (9.1 MJ-eq vs. 24.5 MJ-eq) but its water footprint is significantly higher than polyester (50 m³ vs. <1 m³) [22]. This example demonstrates the conflicting nature of criteria for environmental impact assessment. Moreover, sustainability is a multi-pillar concept in environmental, economic and social contexts each with a pool of different criteria and influencing factors. The concept also has spatial and temporal aspects that expand all around the globe from the past to the far future. Unlike isolated environmental impact assessments, sustainability assessments must integrate social and environmental aspects that introduce significant complexity due to the interwoven nature of socio-ecological systems [23].

Each material and process combination within the textile supply chain presents trade-offs across various environmental and social impacts. The environmental impact of processes used to produce textile products excluding wet processing (as it is common in almost all textile materials) and energy use (due to data limitations) was analysed in [24]. The study highlighted the broad environmental footprint of textiles including water and land use, resource depletion, recyclability, and biodegradability. Furthermore, the trade-offs among natural, synthetic, and regenerated fibres in terms of their environmental impact were identified [24]. While impact assessment methods such as Life Cycle Assessments (LCAs) have shown significant utility for sustainability evaluations, they also include a challenging step of normalization and weighting of impact categories that are influenced by decision-maker preferences. A case study [25] addressed the challenges of interpreting and comparing environmental life cycle assessment (e-LCA) results of t-shirts made from cotton, polyester, and viscose across various environmental impacts such as global warming potential, ozone depletion potential, acidification potential, water consumption, etc. The study demonstrated the utility of MCDM methods to evaluate and select optimal alternatives [25].

On the social front, results of a case study based on Social Life Cycle Assessment (s-LCA) methodology using the Swedish clothing consumption data highlighted important hotspots in the supply chain based on 11 social indicators chosen by consumers [10]. The study flagged sectors like commerce and business services in Bangladesh and China as hotspots for multiple indicators of concern including risk for child labor, non-ratification of ILO conventions, non-fair wages among others. Additionally, key sectors in the textile supply chain, such as plant-based fibre production, textile materials processing, and apparel production, were highlighted as significant negative social hotspots in multiple locations including India, Uzbekistan and China with indicators such as risk of wages falling below the country’s minimum wage, gender inequality and, non-adequate labor laws [10]. In the light of such data, for the textile supply chain to be aligned with the Sustainable Development Goals (SDGs), those issues must be addressed considering the viewpoints of worldwide stakeholders, including farmers, workers, manufacturers, retailers, brand owners, consumers, and communities who do not necessarily share similar interests, perspectives or, priorities – like in Supply Chain Management (SCM).

SCM differs from traditional logistics operations in that it extends beyond the internal processes of a single organization. Instead, SCM is characterized as a coordinated network of multiple companies working collaboratively to deliver finished products to end consumers [12]. SCM is defined as the Decision Making (DM) in strategic, tactical, and operational levels to optimize the performance of supply chain [13]. Because supply chain activities are carried out by various actors within a network, each with distinct and often conflicting criteria, SCM can be characterized as a MCDM problem [12]. Indeed, literature reviews on the application of MCDM methods in sustainable engineering have identified sustainable supply chain management as a key area of application e.g., [14].

2.2. Multiple Criteria Decision

In recent decades, mathematical Multiple Criteria Decision Making (MCDM) methods have been developed to deal with such situations. As the name implies, MCDM assists decision-makers in addressing (solving or organizing) problems that involve multiple criteria (quantitative or qualitative) by logically determining the weights and interrelationships between those criteria. These methods are applied where there is no single optimal solution, so decision-makers’ preferences are used to distinguish between the available options [11]. Solving can mean selecting the "best" alternative, which is the most preferred by the decision-maker(s), or identifying a small set of good alternatives and grouping them into different preference sets – as needed in sustainability management in the textile supply chain engaging with a diversity of potentially competing criteria in holistic approaches that allow for selecting the optimal courses of action (e.g., adopting sustainable initiatives, materials, processes, or partners).

A systematic literature review [26] on MCDM methodologies and their applications from 1999 to 2009 identified 12 major application areas in scholarly research. Transportation and logistics emerged as the most prevalent topic, followed by business and financial management. Notably, they found a total of 32 studies applying MCDM methods in environmental management between 1999 to 2009. Water management, manufacturing and assembly, energy management, agricultural and forestry management, managerial and strategic planning, project management and evaluation, social service, and military service account for the remaining distinct application areas for MCDM methods [26].

While MCDM methods are diverse, they share certain aspects [27]:

Alternatives: These are the various courses of action available to the decision-maker, which will be screened, prioritized, and ultimately ranked.

Multiple criteria: These represent the various dimensions from which the alternatives are evaluated.

Conflicting criteria and different measurement units: Since the criteria reflect different aspects of the alternatives, they may have different units of measurement and could conflict with one another.

Decision weights: Many MCDM methods require assigning weights to the criteria based on their importance, and typically normalized during the process.

Decision matrix: This is the mathematical representation of the problem for calculating the relative performance of the alternatives and different criteria.

MCDM methods handle quantitative, qualitative, or mixed data, and typically use either crisp numerical values or fuzzy linguistic terms. These are mathematical models, there are plenty of them, and the field is still growing. There are also different ways of categorizing MCDM methods. One broad categorization of MCDM methods is differentiation based on the type of data that they handle, i.e., crisp and fuzzy methods. Fuzzy approaches are preferred when uncertainty and incomplete data require expert judgment. In some cases, initially crisp data are converted into linguistic terms for fuzzy set analysis. This conversion occurs when crisp data fail to capture system vagueness or when numerical values do not inherently indicate cost or benefit effect in decision making, leaving the interpretation to expert judgment [28]. Alternative approaches to categorizing MCDM methods are also documented in the literature. In a survey on MCDM methods and concepts [29], key differences of these methods were identified based on several factors: the complexity of the underlying algorithms, the criteria weighting schemes, the representation of a preference evaluation criteria, the handling of uncertainty in data, and the type of data aggregation employed. There are three general steps in a typical MCDM procedure including (1) identification and selection of criteria, (2) determination of resources weights, and (3) ranking the resources using a proper MCDM method [29]. Table 1 presents a list of those MCDM methods identified and repeated within our review sample with their primary function and original source of model development. The article combines weighting methods, ranking methods, causal-structuring approaches, optimization models, efficiency methods, and supporting frameworks under a broad MCDM label. It is important to recognize that while some MCDM methods have outranking function (e.g., AHP and BWM), others approach data differently, focusing on categorizing criteria into levels of importance (e.g., ISM) or examining the cause-effect relationships among them (e.g., DEMATEL).

Growing consumer awareness and increasing sustainability regulations, such as the European Green Deal and Circular Economy Action Plan (CEAP), Corporate Sustainability Reporting Directive (CSRD) in Europe and similarly, Corporate Sustainability Disclosure Standards (CSDR) of China have driven companies to integrate sustainability into their operations. Supplier selection by prioritizing environmentally and socially responsible suppliers while maintaining economic viability plays a key role in competitiveness and public image of companies. A review on the MCDM studies for sustainable supplier selection [52] identified 101 papers that were published between 2018 and 2022. Results supported the previous studies suggesting that DEA, TOPSIS, AHP and their variations are the most frequent MCDM methods for sustainable supplier selection [52]. While literature reviews have examined MCDM in supply chain management and sustainable development separately e.g., in [12,15], we have identified a lack of a systematic literature review (SLR) that explores the intersection of “sustainability assessment” within the “textile supply chain” using “MCDM” methods. Moreover, no existing systematic review thoroughly examines the scopes and context- specific selection of sustainability criteria in MCDM analyses of textile industries to track the trends, significance, and prevalence of criteria and influencing factors.

The aim of this study is to synthesize existing literature on the utility of MCDM methods for sustainability assessment and management in the textile supply chain. Theoretical background for this assumption has derived from the idea that sustainability criteria in the textile supply chain are diverse due to the variability in raw materials, manufacturing processes, locations, logistics, and product use phase. This diversity complicates interpreting sustainability assessment results and MCDM methods provide interpreting aid for determining the relative importance of criteria and impact categories, enabling the ranking and scoring of products, processes, and organizations in regard to sustainability. MCDM methods have been applied by researchers to balance trade-offs and select sustainable textile suppliers e.g., in [53], evaluate corporate sustainability e.g., in [54] and identify sustainable textile products e.g., in [55]. Furthermore, we analyse the environmental, social, and economic criteria used and highlighted in MCDM systems for sustainable decision-making in the textile supply chain and identify the scope of studies in which MCDM sustainability assessments are conducted.

3. Research Method

This study serves to establish connections between the attributes of the textile supply chain (e.g., environmental, social, and economic), the decision-making processes supported by MCDM, and their relevance to achieving sustainable outcomes in this domain. This study employs a systematic literature review (SLR) methodology, following six key stages of (1) defining research questions, (2) determining study characteristics, (3) retrieving a baseline sample of relevant literature, (4) filtering and selecting final set of synthetize studies, (5) synthesizing the findings, and (6) reporting the results [56]. We follow the structured approach outlined by Durach et al. (2017) which is tailored to the distinct characteristics of SLRs within the field of supply chain management [16], see Table 2.

Primary study characteristics are defined by establishing inclusion/exclusion criteria that align with the research purpose and questions, considering both the content and quality of the primary studies. Key concepts including “sustainability”, “assessment”, “MCDM”, “textile”, and “supply chain” are derived from the theoretical framework and expanded with relevant free text terms in line with the research questions. The keywords encompass various supply chain stakeholders, including production, manufacturing, retail, and corporations. Study contexts include dyadic relationships, evaluations within a single stage, and system-wide sustainability assessments. Since this review focuses on MCDM methods, the research method in the papers is primarily "model development," with often also involving "practical validation" through case studies. Various types of data (e.g., crisp and fuzzy) and MCDM models are included, along with papers that use single or hybrid models for criteria weighting and/or alternative ranking. In order to focus to upstream and production-oriented supply-chain decisions, studies with pivotal focus on consumer and post- consumer phases are excluded. Similarly, studies with highly technical scopes focusing on specific technologies or chemical developments are excluded, as they do not align with the research questions of this study.

A baseline sample of potentially relevant literature was retrieved through searches on Scopus and Web of Science on Aug. 07, 2025 with temporal constraint set to limit publish date after 2010. This period reflects recent technological, regulatory, and conceptual developments in the field and reduces the inclusion of studies based on substantially different institutional and market conditions. A combination of keywords and synonyms is used to ensure broad coverage and capture all aspects of the initial theoretical framework, including its limitations. Search terms are applied to titles, abstracts, and keywords based on database specifications. The final list of search terms reflects the diverse definitions and terminology related to sustainability assessment in the textile supply chain using MCDM methods. Inclusion and exclusion criteria are applied to filter relevant studies and refine the sample into a "synthesis sample." Following Durach et al. method [16], the screening process extends beyond titles and abstracts to assess each study’s relevance in refining the theoretical framework. As a result, 625 articles were identified. After title-and-abstract screening 97 papers were selected as the synthesis sample for comprehensive analysis. One study included in initial synthesis analysis was later retracted from database. Therefore, the analyses and results were re-run and updated after eliminating data from that specific study.

As suggested by Durach et al. [16], papers are categorized by scope of studies to offer a clearer contextual perspective and enable a structured and coherent analysis within each scope of study (see Table 3). Article titles, abstracts, and keywords were scanned to identify recurring themes, resulting in five categories which are presented in Table 4 of the results section. The primary analysis was conducted by one researcher using a standardized extraction and classification framework. This approach enabled the consistent application of the same coding and harmonization rules across all included studies. The resulting analytical framework, category definitions, and aggregated findings were subsequently reviewed by both authors.

Furthermore, to address terminological inconsistencies across studies and ensure consistency in reporting and analysis, within each scope of study, sustainability criteria were harmonized and subsequently categorized under a framework consisting of economic, environmental, social, technological, and governance dimensions [17,18,19,20]. For example, in the supplier selection scope of study, terms such as “price,” “material cost,” “purchase cost”, “discount opportunity,” and “payment terms” that were merged under the unified criterion of “Costs and financial flexibility”. Similarly, “repair service,” “recycle rate,” and “reverse logistics for recycling and reusing” were labelled with unified criterion of “circular supply chain practices”. To support transparency, the complete study-level evidence matrix and criteria-harmonization dictionary are provided in the Supplementary Materials.

4. Results

This section presents the results of an in-depth analysis of the synthesis sample. As illustrated in Figure 1, the number of publications addressing the application of MCDM methods for sustainability assessment in the textile supply chain has increased notably since 2019, peaking at 20 studies in 2024. In terms of publication venues, the Journal of Cleaner Production features the highest number of contributions (n=8), followed by Sustainability and the International Journal of Production Economics, each with 6 studies.

MCDM applications in the textile supply chain are highly context dependent. In the reviewed literature, MCDM methods are applied to diverse sustainability problems, including supplier selection, corporate sustainability assessment, prioritization of influential factors in sustainable and circular economy implementation, product evaluation, and strategy development. These applications involve different decision environments, stakeholders, and sustainability priorities, which directly influence the selection of criteria and MCDM techniques. Accordingly, Table 4 categorizes studies based on their thematic scope. This structure enables identification of methodological patterns and dominant sustainability priorities across application contexts. It also facilitates comparison of how different MCDM methods address sustainability trade-offs in textile supply chains, while revealing underrepresented dimensions and research gaps.

While all papers primarily focus on the development and application of MCDM methods to address sustainability-related decision-making challenges within the textile supply chain, they often validate their applicability through case studies, where the models are applied to real-world scenarios such as supplier selection, choice of alternative management strategies, evaluation of sustainability performance within a specific company, etc. Our scope of study classification and identification of units of analysis relies mainly on the case studies used for their practical validation. As illustrated by the number of studies in each study scope within our synthesis sample, MCDM applications in textile sustainability research are concentrated primarily on two scopes of studies: the “Identification of influential factors for sustainability and circularity implementation (drivers, barriers, metrics)” and “sustainable supplier selection”. This distribution highlights the strong research focus on strategic sustainability implementation challenges within textile industries and upstream supply chain decisions. The classification in Table 4 also shows that most studies adopt a system-wide perspective rather than limiting the analysis to a single organizational stage, reflecting the interconnected and global nature of textile supply chains. Moreover, the variation in units of analysis across scopes of studies indicates that the choice of MCDM method and sustainability criteria is strongly influenced by the specific decision-making problem being addressed.

Furthermore, grouping criteria into sustainability dimensions provides a clearer understanding of the relative emphasis placed on different sustainability aspects within textile-related decision-making problems. Table 5 provides study-level prevalence of sustainability dimensions and Table 6 presents a summary of the scope of the studies and the number of raw occurrences of sustainability criteria, categorized under economic, environmental, social, technological, and governance dimensions. A detailed overview is presented as A1 in the Appendix as well as in supplementary documents.

4.1. MCDM for Sustainable Supplier Selection in Textile Industries

Figure 2 presents the MCDM methods applied in our synthesis sample that focused on sustainable supplier selection in the textile industry. Regardless of the data type in the decision matrix (crisp, fuzzy, etc.), AHP is the most common method for criteria weighting, while TOPSIS is the most frequently applied for alternative (suppliers) ranking. The combination of these two methods is the dominant MCDM hybrid approach for supplier selection over the years. Appendix A1 presents detailed information including on data environment in which the problem of sustainable supplier selection has been solved.

However, in order to address the second research question, a structured analysis of the selected and prioritized criteria for sustainable supplier selection for the textile industry among the synthesis sample is required. Different criteria classification frameworks are visible within the studied papers. Besides Triple Bottom Line (TBL) sustainability framework, the Supply Chain Operation Management (SCOR) framework has been repeatedly embedded in criteria selection step. The SCOR model offers a standardized framework for describing supply chain processes, performance metrics, best practices, and enabling technologies. Its core structure encompasses five key processes: Plan, Source, Make, Deliver, and Return [150]. Table 7 presents the distribution of criteria categorized under economic, environmental, social, technological, and governance categories.

To enable consistent and efficient reporting, factors with similar definitions and concepts were grouped under unified criteria. Figure 3 present the raw criteria occurrences in our synthesis sample under the sustainable supplier selection scope of study. Furthermore, as presented in Figure 4, a decision importance criteria graph was created by compiling the top two ranked criteria from each study that included a criterion weighting MCDM system. As illustrated, economic dimension is prevailing both in raw occurrences decision importance criteria with costs and financial flexibility (n _{raw occurrences}=56, n _{decision importance}=14) followed by supply chain agility, capacity and delivery performance criteria in raw occurrences (n _{raw occurrences}= 50) and product quality and design specification (n _{decision importance}=10) in decision importance criteria. On the environmental front, green operational practices (n _{raw occurrences}= 23) and environmental management systems and competencies (n _{raw occurrences}= 22, n _{decision making}= 7), and circular supply chain practices (n _{decision importance}= 2) are prevailing. In social dimension, employees’ rights and interests is repeatedly used criteria (n _{raw occurrences}= 13). However, it ranks second in decision importance (n _{decision importance=} 1) after community, stakeholders and social impact (n _{decision making}= 2). IT, information systems and smart manufacturing appears both in raw occurrences and decision importance diagrams (n _{raw occurrences}= 14, n _{decision making}= 1). Similarly, corporate governance, audits and compliance emerges in the government dimension (n _{raw occurrences}= 3, n _{decision making}= 1). Details on harmonized categorization and the process of compiling up the prioritized criteria by MCDM modelling (decision making importance) are provided as S1 and S2 in supplementary documents.

4.2. MCDM for Corporate Sustainability Assessment in Textile Industries

Several studies have highlighted the application of MCDM methods in the scope of corporate sustainability e.g., in [149]. Within our synthesis sample, eight studies have focused on the context of corporate sustainability performance evaluation. Table 8 provides an overview of these papers, including the applied MCDM methods, case study details, sustainability dimensions, and sources of data. Results shows that majority of these studies (6) do not incorporate a criterion weighting system but only focus on ranking alternatives. In our synthesis sample studies, alternatives are represented either by different years [54,89,93]; or different companies (e.g., in [88,90]). Additionally, TOPSIS is found to be the most frequently used MCDM method for corporate sustainability assessment in the textile industry.

In order to elaborate on the second research question about the selected criteria in MCDM methods, criteria were categorized in harmonized terms based on their definition similarities and context with details as provided in S3 in the supplementary documents.

Figure 5 shows the criteria raw occurrences results of our synthesis sample in the scope of corporate sustainability assessment. The assessment over Environmental dimension is dominant with criteria including energy and water use and waste types (n _{raw occurrences}= 16), followed by environmental impact and emissions (n _{raw occurrence}= 12) and waste types (n _{raw occurrences} = 13). Social criteria including employees’ rights and welfare (n _{raw occurrences}= 14) and organizational culture and leadership (n _{raw occurrence}= 13) are also repeatedly used for corporate sustainability assessments. As mentioned above, majority of studies under this scope of study (6 out of 8) do not incorporate a criterion ranking/weighting method. Therefore, an overview of decision importance criteria is could not be generated.

4.3. MCDM for Identifying the Influential Factors for Sustainability and Circularity Implementation in Broader Contexts (Barriers, Drivers, and Metrics)

The largest number of studies (n=37) within our synthesis sample sought to identify the influential factors for sustainability and circularity implementation in the textile industry at a broader level (e.g., country-wide). As illustrated in Figure 6, DEMATEL is the prominent MCDM method (single or hybrid models) in this scope of study.

Table 9 shows the distribution of the influential factors (drivers, barriers, metrics) that were examined for sustainability and circularity implementation in broader system-wide sustainability scope of study. Overall, the economic dimension shows the highest number of the used criteria (n _{raw occurrences}= 195) followed by environmental dimension (n _{raw occurrences}= 166).

To provide a clearer overview, the influential factors for sustainability and circularity implementation were categorized based on similarities in their definitions and contextual interpretations across this scope of study. The results are presented in Figure 7 as an overview of raw occurrences of influential factors for broader system-wide sustainability studies in textile industries. Governmental policies and legislation factor (n _{raw occurrences}= 62) is most frequently used in this scope of study, followed by costs and capital expenditure (n _{raw occurrences}= 44), environmental impacts and emissions (n _{raw occurrences}= 43), organizational culture and leadership (n _{raw occurrences}= 35), and CE infrastructure: complexities and capabilities (n raw occurrences= 33). Details of this categorization is provided as S4 in supplementary documents for transparency.

Aggregation of findings within this scope of study in the form of a decision importance diagram from our synthesis samples is potentially problematic. As shown in Figure 6 (and also detailed in the Appendix A1) despite ranking/weighting MCDM methods like ANP, AHP, and BWM being applied, the application of DEMATEL is prominent in this scope of study (23 out of 37) alongside other structural analysis methods like ISM. DEMATEL has multiple functions primarily focused on the identification of cause-effect relations and also calculation of the relative dominance factor. Different studies have operationized and interpreted these functions differently to determine the importance level of the sustainability influential factors. This interpretation complexity is more evident when using hybrid models. Therefore, rather than aggregating all results, we zoomed into studies that used DEMATEL to identify relationships the factors. Based on the positive dominance factors, the top three factors in each study were identified and compiled up as the strongest causal influential factors and reported in Figure 8. Government policies and legislation (n _{decision importance}= 9) and knowledge, skill and information gapes (n _{decision importance}= 7) show the strongest causal influence. Details of these strongest casual factors is provided as S5 in supplementary documents.

4.4. MCDM for Sustainable Management Strategies

Several studies (n=14) within the synthesis samples, apply MCDM methods for decision making about adoption of sustainable management strategies and initiatives within the textile industries. These include supply chain managements strategies (e.g., [133,144]) or, sustainable production management strategies (e.g., [140]). An overview of the applied MCDM in this scope of study along with the alternative management strategies is provided as S6 in supplementary documents.

Table 10 presents of the number of relevant sustainability criteria across different dimensions for analysing management strategies. Number of raw occurrences in economic dimension is leading (n=88) followed by environmental dimension (n=64). The criteria were further processed and grouped together in harmonized categories based on the similarities in definitions and contexts. Figure 9 shows the distribution of raw occurrences of criteria in the sustainable management strategies scope of study over the harmonized categorization. Economic performance and costs (n _{raw occurrences}= 37) appear to be the most frequently used criteria in this scope of study, followed by supply chain agility, capacity and delivery criteria (n _{raw occurrence}= 24) and community, stakeholders and social impacts (n _{raw occurrences}= 17). Details of the harmonized criteria categorisation is provided as S7 in supplementary documents.

Although ranking MCDM methods like AHP (n=4), WSM (n=2) and BWM (n=1) have been used in this scope of study, these methods have been primarily used to directly rank the alternative management strategies against the used criteria rather than initially giving weights to criteria. Therefore, providing an overview of decision importance criteria was not possible.

4.5. MCDM for Textile Product Sustainability Assessment

Lastly, few studies (n=5) within our synthesis sample use MCDM methods for product- specific sustainability assessment. As shown in Table 11, these studies primarily rely on quantitative data collected from on-site investigations, databases, and relevant literature, with the environmental dimension being the most prominent in the synthesis sample. Two studies integrate LCA with MCDM approaches to provide a more balanced evaluation of different environmental impact categories [25,146]. The remaining studies employ MCDM methods to incorporate additional sustainability dimensions, such as economic and social factors alongside environmental considerations.

Table 11 represents the challenge of assessment due to the diversity of materials, processes, logistics policies, and product usage. Therefore, optimizing supply chain scenarios for environmental impact is not a straightforward decision but a multiple criteria problem. One study applied PROMETHEE and ELECTRE to interpret LCA environmental impact data for cotton, viscose, and polyester t-shirts from cradle to grave [25]. Assessing 10 impact categories (e.g., global warming potential, human toxicity potential), this study concluded that cotton is the worst scenario while the preference relationship between polyester and viscose is sensitive to the applied MCDM method [25]. Two studies [145,148] integrated life cycle assessment (LCA) with economic data from denim fabric production sites to evaluate the sustainability performance of fabrics made from different fibre combinations. Using TODIM method, one of these studies extended the fibre combination analysis with varying energy supply systems (combined heat and power vs. grid energy) incorporating both environmental indicators such as global warming, acidification, eutrophication, water use, and cumulative energy demand with economic criteria of cost and quality [145]. The second LCA study on denim fabric production further explored alternative allocation approaches (cut-off and waste value) alongside fibre variation using EDAS method [148]. Results showed that recycled PET fibres demonstrated lower freshwater ecotoxicity and reduced eutrophication and acidification impacts, while cotton performed better in terms of human toxicity. Additionally, the cut-off allocation method yielded lower global warming potential for recycled PET, highlighting the importance of allocation choices in recycled material LCAs [148].

The circularity potential of fashion, home textiles, and technical textiles were evaluated by applying TOPSIS using 15 criteria across environmental (e.g., fabric origin, landfill waste), economic (e.g., market demand, production volume), technical (e.g., energy efficiency), and social (e.g., employment rate) categories. Sensitivity analysis showed that landfill waste, fabric origin, and production energy efficiency had minimal sensitivity to weight changes and were assigned the lowest weight. In contrast, environmental impact and number of washes criteria were most sensitive and received the highest weight. TOPSIS rankings indicated that fashion textiles have the highest circular economy potential due to their large market share and cost-effectiveness, followed by technical textiles and home textiles [147].

Table 12 presents the number criteria used for product sustainability assessments. The highest number belongs to the environmental dimension (n=41), followed by economic (n=10).

These criteria were further processed in harmonized categories based on the similarity of their definitions and contexts. Figure 10 shows the raw criteria occurrences in the product sustainability scope of study. Environmental impacts and emissions (n _{raw occurrences}= 36) show the highest number, followed by energy and water use (n _{raw occurrences}= 5) and product quality and specifications (n _{raw occurrences}= 3) are the categories with highest number of used criteria for product sustainability assessment. Details of harmonized criteria categorization is presented as S8 in supplementary documents.

5. Key Findings and Discussion

This systematic literature review provides a context-sensitive synthesis of how multi-criteria decision-making methods have been employed to support sustainability assessment and management in textile and apparel supply chains. Whereas previous reviews have examined MCDM applications in supply chain management, sustainability engineering, sustainable development, or supplier selection more generally [12,14,15,52,152,153,154], the present review focuses specifically on the relationship between the decision context, the methodological function of the MCDM approach, and the sustainability criteria incorporated into textile-sector applications. The findings indicate that MCDM use in this field is not methodologically uniform. Instead, particular methods have become associated with particular types of decision problems, data sources, units of analysis, and sustainability interpretations. Three overarching conclusions emerge. First, the choice of MCDM method is strongly related to the decision context: AHP–TOPSIS combinations dominate supplier-selection applications, DEMATEL is particularly prevalent in studies identifying drivers and barriers, and distance-based approaches are commonly used in product-level assessments. Second, despite the nominal inclusion of environmental and social considerations, economic and conventional operational criteria remain highly influential in many applications. Third, the methodological sophistication of the reviewed models, including the use of hybrid and fuzzy approaches, does not necessarily imply that sustainability has been assessed comprehensively, transparently, or robustly. These findings have implications for both the methodological development of MCDM-based sustainability assessment and its practical use in textile supply chains.

5.1. Context-Dependent Patterns in the Application of MCDM Methods

The results confirm the broader observation that supply chain management constitutes an important field of application for MCDM methods [12,14]. Supply chains involve decisions at strategic, tactical, and operational levels and require the simultaneous consideration of multiple actors, objectives, and performance dimensions [13]. However, the present review demonstrates that the role performed by MCDM differs substantially across the five identified application contexts. Sustainable supplier selection constitutes the second-largest field of application in the reviewed sample. This prominence is consistent with previous reviews showing that supplier selection is one of the most established areas of MCDM application [52,152,153,154]. Supplier-selection decisions involve a clearly defined set of alternatives and criteria and are therefore well suited to methods that combine criteria weighting with alternative ranking. In the textile-sector literature, AHP is most frequently used to derive criteria weights, while TOPSIS is commonly used to rank suppliers according to their distance from positive and negative ideal solutions. The prevalence of AHP–TOPSIS combinations can therefore be explained by their functional complementarity rather than by the intrinsic superiority of either method.

AHP provides a relatively accessible procedure for structuring a decision hierarchy and eliciting decision-maker preferences through pairwise comparisons. TOPSIS then produces an intuitive ranking based on the relative proximity of each supplier to an ideal solution. This combination may be particularly attractive for company-level applications because the results are comparatively easy to communicate to managers. The dominance of AHP and TOPSIS is consistent with earlier reviews of green and sustainable supplier selection [152,153,154]. Nevertheless, their widespread adoption may also reflect methodological familiarity and ease of implementation. The popularity of a method should therefore not be interpreted as evidence that it is necessarily the most appropriate method for every supplier-selection problem. The compensatory logic underlying many AHP–TOPSIS applications deserves particular attention. In a fully compensatory model, poor performance regarding one criterion can be offset by good performance regarding another. From a conventional supplier-performance perspective, such compensation may be acceptable. From a sustainability perspective, however, it may become problematic when low costs or high product quality compensate for serious deficiencies in occupational health, labour rights, pollution control, or legal compliance. Future supplier-selection models should therefore consider whether particular social and environmental criteria should be treated as minimum requirements or exclusion thresholds rather than as fully compensable performance attributes. Outranking approaches or two-stage procedures combining eligibility thresholds with subsequent ranking could be particularly relevant in this regard.

The largest group of studies uses MCDM to identify and prioritize metrics, drivers, barriers, or critical success factors for the implementation of sustainability and circular economy practices. DEMATEL is the most frequent method in this application context. This finding is methodologically plausible because DEMATEL is specifically designed to represent interdependencies and perceived cause–effect relationships among factors [155]. Unlike methods that provide only an ordinal ranking, DEMATEL distinguishes between cause and effect groups and visualizes the structure of relationships within a decision problem. Such capabilities are relevant to sustainability transformations in textile supply chains, where regulatory, technological, organizational, financial, and market-related factors interact. Nevertheless, DEMATEL results should be interpreted carefully. The relationships identified by the method represent structured expert judgements and should not automatically be understood as empirically demonstrated causal relationships. The outcome depends on the composition of the expert panel, the formulation and aggregation of judgements, the threshold used to construct the relation map, and the specific institutional and geographical context. A factor identified as a “cause” in one national textile industry may operate differently in another institutional environment. DEMATEL is therefore most appropriately understood as a method for structuring perceived influence relationships and generating hypotheses about system dynamics rather than as a substitute for statistical or longitudinal causal analysis.

Hybrid DEMATEL–ANP applications partly address the limitation that DEMATEL alone does not provide a complete ranking procedure. DEMATEL can first reveal interdependencies, while ANP can subsequently derive criteria weights within a network structure. Such hybridization is useful when each method performs a clearly defined function. More generally, however, the value of hybrid MCDM models should be assessed according to whether their components solve distinct methodological problems. Combining several methods does not inherently improve the validity of the decision model. Hybridization may increase complexity without increasing explanatory power when methods are added primarily to demonstrate methodological novelty. Corporate sustainability assessment represents a smaller but distinct application field. In these studies, TOPSIS is commonly used to compare companies or to evaluate the performance of the same company across several years. This application is compatible with the structure of TOPSIS because companies or reporting years can be represented as discrete alternatives assessed against a common set of indicators. The use of TOPSIS for environmental and corporate performance evaluation has also been justified by its comparatively transparent mathematical structure and straightforward interpretation [94].

However, corporate sustainability rankings are only as reliable as the underlying data. The reviewed studies frequently draw on corporate sustainability reports, annual reports, or public databases. Although these sources provide quantitative information on emissions, energy, water, materials, and waste, their use raises questions regarding comparability. Companies may apply different organizational boundaries, reporting periods, consolidation rules, estimation methods, and levels of external assurance. Changes in a company’s reported performance may consequently reflect changes in reporting practices rather than changes in its actual sustainability performance. MCDM can aggregate available indicators, but it cannot resolve inconsistencies in the underlying data. Future corporate applications should therefore document data provenance, reporting boundaries, missing-data treatment, and external verification more explicitly. Studies focusing on sustainable management strategies apply a wider variety of methods. This methodological diversity reflects the heterogeneity of the underlying decisions, which range from energy-efficiency measures and cleaner-production practices to circular supply-chain and reverse-logistics strategies [136,137,141,145]. In these contexts, no single method can be expected to dominate because the alternatives, time horizons, stakeholders, and data types differ considerably. The appropriate method should therefore be selected according to the structure of the decision problem, the form of uncertainty, and the intended decision output, as emphasized in general MCDM literature [29].

Product-level sustainability assessment remains the least developed application field. Most product-related studies combine an environmental assessment, such as LCA or chemical footprint analysis, with a distance-based or outranking MCDM method [25,146,147,148,149]. This combination addresses a recurring problem in life-cycle assessment: environmental impact categories are expressed in different units and may point in conflicting directions. MCDM can support the interpretation of such multidimensional results by making preferences and trade-offs explicit. At the same time, the integration of LCA and MCDM must be undertaken cautiously. Normalization, weighting, and aggregation can strongly influence the resulting product ranking. Moreover, the use of economic, technical, and social criteria alongside environmental LCA indicators creates potential differences in scale, data quality, and system boundaries. Product cost may be measured directly, whereas social impacts may be represented through qualitative proxies and environmental impacts through modelled life-cycle data. Combining these different forms of evidence in a single index can create an appearance of precision that is not necessarily matched by the underlying information. Product-level studies should therefore report alternative weighting scenarios, rank stability, and the consequences of different normalization and aggregation procedures.

5.2. Handling Uncertainty: The Prevalence and Limitations of Fuzzy Approaches

The frequent use of fuzzy MCDM, particularly in supplier-selection studies, reflects the high degree of uncertainty and imprecision characterizing sustainability decisions. Decision-makers often lack precise numerical information regarding a supplier’s environmental competence, organizational commitment, social performance, flexibility, or innovation capacity. Linguistic assessments such as “high,” “medium,” or “low” can be represented by fuzzy numbers, thereby allowing qualitative judgements to be incorporated into a mathematical decision model. The reviewed applications of fuzzy AHP–TOPSIS are consistent with previous research demonstrating the usefulness of fuzzy representations for textile supplier selection under uncertain conditions [67]. However, fuzzy modelling should not be equated with the elimination of uncertainty. It transforms linguistic or imprecise judgements into numerical ranges but does not necessarily increase the reliability of the original judgement. The selection of linguistic scales, membership functions, defuzzification procedures, and aggregation operators introduces additional assumptions. Moreover, fuzzy models primarily address vagueness and imprecision. They do not automatically address other forms of uncertainty, such as disagreement between stakeholders, variability over time, incomplete system knowledge, measurement error, or uncertainty concerning future regulatory and market developments.

The results therefore suggest a need for a more differentiated treatment of uncertainty. Fuzzy methods may be appropriate when the primary concern is linguistic vagueness, whereas stochastic approaches may be more appropriate for probabilistic variability and interval or grey approaches for incomplete information. Scenario analysis may be required when the future development of prices, technologies, regulation, or consumer demand cannot be represented by a single probability distribution. Method selection should thus be based on the nature of the uncertainty rather than on the general popularity of fuzzy extensions. The reviewed literature also provides limited evidence of systematic comparisons between methods or uncertainty representations. Different MCDM models can produce different rankings even when applied to the same decision matrix. This is particularly relevant where small changes in criteria weights or normalization procedures can alter the preferred supplier, strategy, company, or product. Sensitivity analysis and robustness testing should therefore become standard components of MCDM-based sustainability assessment. At a minimum, studies should test alternative weighting scenarios, evaluate changes in ranking, and disclose criteria that exert a disproportionate influence on the final result.

5.3. Sustainability Criteria and the Persistence of Economic Decision Logics

The synthesis of criteria provides an important insight into how sustainability is operationalized in textile-sector MCDM research. The distribution (see table 6) confirms that the traditional economic, environmental, and social dimensions remain central but also supports the inclusion of technological and governance factors, as suggested in the conceptual literature [17,18,19,20]. The prominence of economic criteria is particularly evident in sustainable supplier selection and in studies of sustainability drivers and barriers. Cost, quality, delivery, flexibility, responsiveness, productivity, and financial capacity occur frequently. Some of these criteria are indispensable to the commercial viability of supply-chain decisions. A sustainability assessment that ignores economic feasibility may recommend alternatives that cannot be implemented or maintained. In this sense, the inclusion of economic and operational criteria is consistent with the multidimensional understanding of sustainable development, which requires consideration of economic, environmental, and social objectives [17,18]. Nevertheless, the results also indicate that many MCDM applications may extend conventional supplier or operational-performance models by adding environmental and social criteria without fundamentally altering the underlying decision logic. Product cost and quality not only occur frequently but also repeatedly receive high weights in supplier-selection models. By contrast, social criteria rarely appear among the most highly prioritized criteria. Occupational health and safety, employee rights, information disclosure, and community impacts are included in some models but seldom become decisive for the final ranking.

This finding has two possible interpretations. First, it may reflect the actual priorities of the decision-makers whose judgements determine the weights. Second, it may result from asymmetries in data availability and measurability. Cost, quality, lead time, and energy use are comparatively easy to quantify, while labor rights, gender equality, worker participation, community effects, and distributive justice are more difficult to represent through standardized indicators. MCDM models may therefore reproduce existing informational and organizational priorities rather than provide a neutral representation of sustainability. The distinction between frequently used criteria and highly prioritized criteria is consequently important. A criterion may occur in many models but receive little decision weight. Conversely, a less frequently used criterion may be decisive within a particular application. Frequency indicates the visibility of a topic in the literature, whereas weighting indicates its importance within a specific decision model. Neither measure independently establishes the substantive importance of a criterion for sustainability. Future reviews and empirical studies should therefore distinguish between criterion prevalence, decision weight, rank influence, and actual sustainability impact.

The results further show that the balance between dimensions varies according to the application context. Environmental criteria dominate product-level and many corporate assessments, whereas economic criteria dominate supplier-selection and strategy studies. Governance criteria occur most frequently in broader analyses of drivers and barriers, particularly through legislation, standards, certification, governmental support, and institutional coordination. This contextual variation indicates that sustainability dimensions should not be interpreted as independent and interchangeable categories. Governance and technological conditions often determine whether environmental and social improvements can be implemented, while economic and organizational conditions influence their persistence. Governance is particularly important in global textile supply chains because environmental and social impacts frequently occur outside the organizational and geographical boundaries of the focal company. Patterson et al. [19] emphasize that sustainability transformations are shaped by governance arrangements and political processes. The comparatively low overall occurrence of governance criteria may therefore represent an important limitation of current MCDM applications. Supplier-level assessments focusing predominantly on price, quality, and environmental management systems may overlook enforcement capacity, worker representation, access to grievance mechanisms, transparency, and power asymmetries between buyers and suppliers. The same applies to technological criteria. Technology can enable cleaner production, traceability, recycling, resource efficiency, and improved process control, but it should not automatically be treated as evidence of sustainability. The contribution of a technology depends on how it interacts with environmental, social, economic, and governance conditions [20]. Future MCDM models should therefore assess not only the availability or sophistication of a technology but also its actual sustainability outcomes, implementation requirements, potential trade-offs, and distributional effects.

5.4. Trade-Offs and the Limits of Aggregated Sustainability Scores

The textile supply chain is characterized by trade-offs between environmental impacts, economic feasibility, technical performance, and social conditions [24]. Comparisons between cotton, polyester, recycled fibers, and other materials illustrate that no alternative necessarily performs best across all relevant impact categories. Morrison-Saunders and Pope [23] argue that identifying and managing trade-offs is a central function of sustainability assessment. MCDM methods are potentially valuable because they make the aggregation of conflicting criteria more systematic and transparent. However, the reviewed applications often focus on producing a final ranking or identifying a single preferred alternative. Such an output may conceal rather than clarify the underlying trade-offs. A supplier can achieve a high aggregate score despite weak performance on a particular social criterion, while a product can be ranked as preferable even though it performs poorly in an individual environmental impact category. The final ranking therefore needs to be accompanied by a transparent account of which improvements are obtained, which disadvantages remain, and which forms of compensation were accepted. A more informative sustainability assessment would not only answer which alternative ranks first but also explain why it ranks first, under which weighting assumptions it remains preferable, which stakeholders’ benefit, which stakeholders bear the disadvantages, and whether minimum sustainability conditions are satisfied. Non-compensatory criteria, veto thresholds, dominance analyses, and scenario-based rankings may help preserve information that would otherwise disappear during aggregation. The review also suggests that the term “trade-off” is sometimes used broadly to describe the existence of multiple criteria without examining the substantive consequences of balancing them. The presence of an MCDM method does not demonstrate that trade-offs have been resolved legitimately. The weighting procedure determines whose preferences are represented, while the aggregation method determines which forms of compensation are permitted. Both steps therefore involve normative choices that should be disclosed and justified.

5.5. Data Sources, Stakeholder Representation, and Geographical Concentration

The validity of an MCDM model depends not only on the selected method but also on the actors and information used to construct it [150]: Many studies, particularly those concerned with drivers, barriers, and management strategies, rely heavily on expert panels. Expert knowledge is valuable where quantitative data are unavailable or where organizational and institutional relationships must be assessed. Nevertheless, the term “expert” can encompass very different stakeholder groups, including managers, academics, consultants, policymakers, suppliers, and technical specialists. The composition of expert panels affects the criteria selected, the relationships assumed, and the weights assigned [151]. A panel dominated by managers may place greater emphasis on cost, productivity, technological feasibility, and market performance, whereas workers, local communities, environmental organizations, or consumer representatives may prioritize different issues. The limited prominence of social criteria in final rankings may therefore be related not only to the characteristics of the methods but also to the distribution of decision-making power within the assessment process. Future studies should document panel composition, expertise, institutional affiliation, stakeholder role, selection procedure, and the treatment of disagreement. Including affected stakeholder groups would improve both the substantive validity and legitimacy of MCDM-based sustainability assessments. Where full participation is not feasible, separate stakeholder weighting scenarios could make differences in priorities visible rather than averaging them into a single consensus value.

The geographical distribution of studies also requires careful interpretation. Supplier-selection studies are concentrated particularly in Turkey, India, Vietnam, Pakistan, and Bangladesh, while broader studies of drivers and barriers frequently examine the Indian and Bangladeshi textile industries. This concentration is understandable because these countries play major roles in textile and apparel manufacturing and face significant sustainability challenges. However, the findings cannot automatically be transferred to other institutional and supply-chain contexts. Regulatory enforcement, industrial structure, labor-market conditions, technological capabilities, access to finance, infrastructure, and buyer–supplier relationships differ across countries. A barrier identified as central in one national industry may be less relevant in another. Likewise, criteria weights derived from one group of experts cannot be treated as universal sustainability preferences. The current evidence base is therefore more representative of export-oriented manufacturing contexts than of the textile supply chain as a whole. Other stages remain comparatively underrepresented, including fiber cultivation, logistics, retail, product use, repair, collection, sorting, reuse, and recycling. This is particularly relevant because circular economy strategies require coordination before, during, and after product use. Restricting analysis primarily to manufacturing and supplier selection risks overlooking decisions that determine product longevity, consumer behavior, reverse logistics, material recovery, and secondary-market development.

5.6. Theoretical and Practical Implications

The findings contribute to the MCDM and textile-sustainability literature by demonstrating that the relationship between method and application context can be represented as a sequence of interconnected design choices [150]. These choices include the decision problem, unit of analysis, stakeholder group, data source, type of uncertainty, methodological function, aggregation logic, and sustainability dimensions considered. The suitability of an MCDM method cannot therefore be assessed independently of these contextual characteristics. Based on the synthesis, a context–method–criteria framework can be proposed. Supplier-selection problems generally require criteria weighting and alternative ranking; analyses of drivers and barriers require the structuring of interdependencies; corporate assessments require comparable and verified performance data; management-strategy decisions require methods capable of accommodating heterogeneous alternatives and future uncertainty; and product assessments require transparent integration with life-cycle information. Across all contexts, stakeholder representation, data quality, trade-off assumptions, and robustness analysis constitute cross-cutting requirements.

For practitioners, the findings suggest that MCDM model design should begin with the decision problem rather than with a preferred method. Companies should identify:

• decision to be supported
• actors affected by the decision
• relevant sustainability boundaries
• available data and their quality
• form of uncertainty
• criteria that may or may not be compensated
• required type of decision output.

Only after these elements have been specified should a weighting, ranking, outranking, causal-structuring, or optimization method be selected. This problem-oriented procedure would reduce the tendency to apply familiar combinations such as AHP–TOPSIS without sufficient justification. The findings also have implications for policymakers and standard-setting organizations. Standardized indicator definitions and reporting boundaries could improve the comparability of corporate and supplier-level assessments. Policy frameworks could further encourage the use of minimum social and environmental requirements rather than relying exclusively on relative rankings. MCDM should support compliance and sustainability improvement but should not convert violations of fundamental requirements into ordinary trade-offs.

5.7. Future Research Directions

The results point to several priorities for future research. First, studies should move beyond demonstrating the technical feasibility of a newly developed hybrid model. Comparative applications using the same decision problem and dataset are needed to examine whether different methods produce materially different rankings and recommendations. Such comparisons should include sensitivity, uncertainty, and rank-stability analyses. Second, future research should evaluate the external validity of MCDM recommendations. Most reviewed studies validate their models through case application, but applying a model to a company or expert panel does not demonstrate that the recommended alternative produces better sustainability outcomes. Longitudinal studies could examine whether the selected supplier, strategy, technology, or product subsequently improves environmental and social performance. Third, stakeholder participation should be broadened [152]. Workers, suppliers, communities, consumers, civil-society organizations, and policymakers should be included alongside corporate and academic experts. Multi-actor approaches could reveal conflicts in preferences that are obscured by aggregated expert weights.

Fourth, social and governance dimensions require further conceptual and methodological development. Future models should incorporate more substantive indicators of labor rights, living wages, gender equality, worker participation, grievance mechanisms, community effects, transparency, regulatory enforcement, and power relations. Particular attention should be given to criteria that should function as thresholds rather than as compensable attributes. Fifth, product-level applications should be expanded. The limited number of studies integrating MCDM with LCA, LCC, social assessment, and circularity indicators suggests substantial scope for further research [25,146,147,148,149]. Such studies should cover the entire product life cycle and explicitly address use intensity, durability, repairability, reuse, collection, sorting, recycling quality, and end-of-life treatment. Finally, future research should develop more dynamic approaches. Most existing models provide a static assessment based on current expert opinions or available data. Textile supply chains are affected by changing regulation, technology, prices, consumer expectations, geopolitical conditions, and climate-related risks. Scenario-based, dynamic, and periodically updated MCDM models could therefore provide more relevant support for long-term sustainability decisions.

5.8. Limitations of the Review

The findings should be interpreted in light of several limitations. The review is restricted to publications indexed in Scopus and Web of Science, written in English, and published after 2010. Relevant studies available in other databases, languages, conference proceedings, industry reports, or grey literature may therefore have been omitted. This limitation may be particularly relevant for region-specific applications and practitioner-oriented decision-support models. The harmonization of heterogeneous criteria represents another limitation. Criteria reported under different names were grouped into unified categories to enable comparison. Although this procedure reduces terminological fragmentation, it may also conceal differences in definition, measurement, system boundary, and decision context [153]. Raw occurrence counts are additionally influenced by the number and granularity of criteria used in individual studies. They should therefore be interpreted as indicators of visibility within the literature rather than as direct measures of sustainability importance.

Each study was assigned to its most relevant application scope, although some studies could plausibly belong to more than one category. Furthermore, the current analysis distinguishes methodological functions, data environments, and sustainability criteria but does not provide a standardized quality appraisal of every included MCDM model. Future reviews could assess reporting transparency, expert-panel composition, consistency tests, sensitivity analysis, data quality, validation, and reproducibility in a more systematic manner. Despite these limitations, the review offers a structured overview of how MCDM methods have been applied across different textile-sector decision contexts and identifies recurring methodological and substantive imbalances. It therefore provides a foundation for developing more transparent, context-sensitive, stakeholder-inclusive, and robust sustainability decision-support models.

6. Conclusion

This systematic review examines how multi-criteria decision-making methods have been used to support sustainability-related decisions in textile and apparel supply chains. The findings reveal a clear relationship between decision context and methodological choice. AHP–TOPSIS combinations dominate sustainable supplier-selection studies, DEMATEL-based approaches are particularly prevalent in analyses of sustainability drivers and barriers, and product-level assessments primarily employ distance-based methods to support the interpretation of life-cycle information. At the same time, the review shows that sustainability remains unevenly operationalized. Conventional economic and operational criteria, including cost, quality, productivity, and delivery performance, are particularly prominent, whereas social and governance considerations are less frequently represented and rarely become decisive in final rankings. This indicates that many existing applications extend conventional decision models by adding sustainability criteria but do not necessarily transform the underlying decision logic. Methodological sophistication, including hybrid and fuzzy models, should therefore not be equated with comprehensive or robust sustainability assessment.

The principal contribution of this review is a context–method–criteria–framework that demonstrates how the suitability of an MCDM approach depends on the decision problem, unit of analysis, data source, stakeholder composition, uncertainty, and sustainability concept. For practitioners, the findings underline the need to justify method selection, disclose weighting assumptions, incorporate affected stakeholder perspectives, apply minimum social and environmental thresholds, and test the robustness of rankings. Future research should compare alternative MCDM methods using common datasets, examine sensitivity and rank stability, validate decision recommendations against observed sustainability outcomes, and improve the representation of workers, communities, consumers, and policymakers. Greater attention should also be given to product use, repair, reuse, collection, recycling, and other post-consumer stages. The findings must nevertheless be interpreted in light of the review’s database, language, temporal, and scope restrictions, as well as the challenges associated with harmonizing heterogeneous criteria and comparing studies of differing methodological quality.