Network Data Envelopment Analysis with Data-Driven Absolute Standards

A major shortcoming of the network data envelopment analysis (DEA) models is their inability for quantifying the absolute performance of the evaluated decision-making units (DMUs). Although these models effectively identify best practices in performance evaluation, they fail to determine the frontier at both the divisional and overall levels. In this paper, we create a set of standard DMUs at both the divisional and overall levels that are seamlessly integrated into the performance analysis using the traditional network DEA structure and activity matrices. The proposed approach is capable of incorporating predefined production standards in network or multi-division systems and accommodating complicated interdependencies and undesirable outputs, presenting a robust performance evaluation. We demonstrate the effectiveness of the proposed approach via a practical application in the food industry, and with an emphasis on the importance of providing a measure of absolute performance for evaluated DMUs in network structures.