A HYBRID INTEGER DATA ENVELOPMENT ANALYSIS BASED ON AN ALTERNATIVE APPROACH OF SUPER SLACK BASED MEASURE FOR MEASURING SUPER EFFICIENCY AND INEFFICIENCY OF DECISION MAKING UNITS

Abstract

Conventional data envelopment analysis (DEA) models assume that all values of input and output variables are non-integers. However, in many situations of real life DEA application, some of the input and/or output variables are to be integers.The rounding process of the non-integer values to their nearest integers can cause inaccurate efficiency and performance measurement. Thus, it is important to develop a non-radial hybrid integer DEA model. Moreover, the projection of decision making units (DMUs) in the existing super efficiency slack based measure hybrid integer model may not be strongly Pareto-efficient.Thus, our hybrid integer DEA model is developed based on an alternative super efficiency slack based measure DEA model for ranking efficient and inefficient DMUs.The developed model can project efficient DMUs resulted from super efficiency strongly Pareto-efficient by identifying input excesses and output shortfalls of non-integer and integer slack variables. It then discriminates inefficient DMUs whose efficiency scores resulted from super efficiency slack based measure model equal to one.An application to measure the efficiency of 41 academic departments at National Cheng Kung University in Taiwan (NCKU) is used to test the applicability of developed models. The results showed that 51.20% of the departments (DMUs) were super efficiency. Moreover, the result also showed that most of the inefficient departments were caused by their input excesses of expenses and teaching space, and output shortfalls associating with publications and grants.