Improved DEA models in the presence of undesirable outputs and imprecise data: an application to banking industry in India

Abstract

Data envelopment analysis (DEA) is a widely used non-parametric technique for measuring the performance of a homogeneous set of decision making units (DMUs). The basic DEA models are typically based on the conjecture of input minimization and output maximization, and are limited to crisp data. Therefore, the present paper focuses on the DEA models that can handle undesirable outputs and imprecise input–output data forms like intervals or ordinal relations or fuzzy numbers. These models measure the final efficiency of each DMU as an interval for interval and ordinal data, and a fuzzy number for fuzzy data. Moreover, comparison with the existing models show that the proposed models are more theoretically accurate, numerically efficient and measure less number of DMUs as efficient. In addition, some numerical examples with different data sets and an application to the banking industry in India are presented to validate effectiveness of the proposed models. © 2017, The Society for Reliability Engineering, Quality and Operations Management (SREQOM), India and The Division of Operation and Maintenance, Lulea University of Technology, Sweden.