Sensitivity of efficiency classifications in the inverse DEA models

摘要

The previous methods of sensitivity analysis have restricted attention to technical aspects of efficiency. Evaluations arising from prices, costs and preference on decision making units (DMUs) are not addressed in these methods. In this paper, we propose a modified version of inverse data envelopment analysis (DEA) models for sensitivity of efficiency classifications of efficient and inefficient DMUs in which important policies over inputs, outputs and DMUs are represented by preference cones. We use the non-dominated solutions of the modified version of inverse DEA to obtain the upper and lower bounds of inputs and outputs variations range. The method is illustrated through an example in which data set is taken from previous research on DEA.