Data Perturbations for All DMUs in DEA with Interval Uncertainty

This paper discusses the sensitivity analysis in Data Envelopment Analysis (DEA) with interval data. According to the interval efficiencies, all DMUs can be classified in three categories. Based on that, corresponding models are established to explore the sensitivity analysis and stability radius for all DMUs in the interval data setting, within which the classification keep unchanged. A very close numerical value to stability radius is selected to see the tiny change of efficiency. The approaches in this paper not merely can be adopted into the situation when the data variations are only applied to specific DMUs, but also to all DMUs and simultaneously different in various inputs and outputs subsets.