A new approach to overall performance evaluation based on multiple contexts: An application to the logistics of China

Data envelopment analysis (DEA) is a useful method for evaluating the performance of decision making units (DMUs). In this paper, we propose a new approach to overall performance evaluation of DMUs based on multiple contexts in the framework of DEA. For a given set of DMUs, an algorithm is performed to identify frontiers of different efficiency levels as evaluation context. An ideal case is supposed in which all evaluation contexts lead to a consistent report on the performance of DMUs. Shannon entropy is employed to measure the entropy deviations of evaluation results from the real case to the ideal case. A constrained optimization model is constructed to integrate the results against multiple evaluation contexts into an overall performance score for each DMU. The proposed approach is applied to evaluate the logistics performance of China. Its comparisons to some previous methods are also illustrated using the empirical application. It is shown that the proposed approach is robust and provides a more comprehensive evaluation for logistics performance.