Regret-Theory-Based Two-Stage Cross-Efficiency Models for Supply Chain Evaluation under Bounded Rationality

Data envelopment analysis (DEA) is widely used to evaluate supply chain performance in companies, but traditional DEA models have three key limitations: (1) limited differentiation in efficiency scores, where multiple supply chains may be classified as efficient with identical scores; (2) lack of transparency in internal system structures, which fails to account for interactions between different stages; and (3) the assumption of perfect rationality in decision-makers (DMs), neglecting real-world factors such as bounded rationality and subjective preferences. To address these issues, this paper introduces innovative two-stage cross-efficiency models — neutral, benevolent and aggressive — by integrating two-stage cross-efficiency evaluation with regret theory. The main innovations include: (1) models developed within the framework of bounded rationality; (2) a method for constructing a reference supply chain, which allows DMs to assess the perceived utility of regret from collaborating with evaluated supply chains; and (3) the establishment of maximum target efficiency values for each supply chain, ensuring that targets are achievable. A case study demonstrates the effectiveness of the proposed models, which outperform existing DEA approaches by accounting for DMs’ bounded rationality and providing more reasonable evaluation results.