An integrated decision-making approach for identifying sustainable tier-1 suppliers: Fuzzy Delphi-Borda-SWARA

Abstract

Outsourcing has emerged as a key strategic approach for many organizations and manufacturers, leading to the creation of complex multi-tiered supplier networks. Efficient Supplier Network Management (SNM) is necessary, as every element within the network can significantly influence the final product’s quality and delivery timeline. Among the strategies employed in SNM, the tiering of suppliers holds particular prominence, with tier-1 suppliers assuming a pivotal role in ensuring high-quality products and timely delivery. However, despite the criticality of tier-1 supplier selection, existing literature lacks comprehensive frameworks, specifically addressing the specification and prioritization of influential criteria in this context. This study addresses this gap by identifying and integrating key attributes of sustainable tier-1 suppliers through a meticulous review of the literature and expert panel consultations, followed by developing and categorizing a comprehensive list of criteria. A hybrid model is built up to determine criteria weights and relative importance values, employing Multi-Attribution Decision-Making (MADM) techniques; i.e., fuzzy Delphi, Borda, Step-wise Weight Assessment Ratio Analysis (SWARA). The model is then applied to an Iranian spare parts supplier, revealing key criteria for sustainable tier1 supplier selection wherein “proper manufacturing and supply capabilities” is the most critical one. It is worth noting that this model is applicable across various industries beyond automotive. These criteria form the foundation of an efficient assessment framework designed specifically for tier-1 suppliers. Using the fuzzy Delphi, we identify 35 critical parameters influencing tier-1 supplier selection, from which the 10 most critical ones are specified and prioritized. These 10 criteria are also weighted using the Borda and SWARA methods. A comparative analysis is then conducted between the proposed model and the well-known MADM model in order to appraise the robustness and reliability. Additionally, Principal Component Analysis (PCA) and K-Means clustering methods are employed to uncover hidden patterns and groupings among the criteria based on expert evaluations. These insights further enhance the interpretability and decision-support capacity of the framework.