Optimization & Operations Research
Jafar Pourmahmoud; Ahad Abbasi; Alireza Ghaffari-Hadigheh
Abstract
Data Envelopment Analysis (DEA) is a well-established methodology for assessing the efficiency of decision-making units. In complex systems comprising multiple interconnected subsections, Network DEA provides a structured framework for efficiency evaluation. However, ...
Read More
Data Envelopment Analysis (DEA) is a well-established methodology for assessing the efficiency of decision-making units. In complex systems comprising multiple interconnected subsections, Network DEA provides a structured framework for efficiency evaluation. However, traditional DEA models rely on the assumption of deterministic data, which inadequately reflects the inherent uncertainty present in real-world scenarios. Traditional uncertainty-handling methods, such as fuzzy logic, stochastic models, and interval-based techniques, often fail when there is limited historical data and when expert opinions significantly influence the dataset. To address these limitations, this study introduces an uncertain network DEA model based on Liu’s uncertainty theory, facilitating a more accurate assessment of efficiency under conditions of data imprecision. The proposed model is designed for three interconnected subsections and is further extended into a generalized multi-stage framework, allowing it to adapt to increasingly complex systems. Its effectiveness and practical applicability are demonstrated through two numerical case studies in the banking industry, highlighting its capacity to support decision-making under uncertainty. The findings emphasize the model's potential to enhance efficiency evaluation methods, particularly in environments characterized by limited and uncertain data.
