Sensitivity analysis and performance security margin of decision making units with imprecise data

As the calculated efficiencies for the DMUs is relative, so each DMU attempts to improve its performance to don’t miss the position of efficiency in compete with other DMUs. Generally, the performance of DMUs can be evaluated from two perspectives-optimistic and pessimistic. A part of data envelopment analysis examines the sensitivity of the set of efficient DMUs to changes in input and output values. In real world, DEA is sometimes faced with imprecise inputs and/or outputs. We focus in this paper on the one of important subjects of sensitivity analysis and present an algorithm which use the classic interval and fuzzy DEA models, can determine the relative performance security margin of DMUs with imprecise inputs and outputs and their simultaneous perturbation. Of course, in addition to the optimistic frontier, here we also consider the pessimistic frontier for the observed DMUs and call it the inefficiency improvement margin. With this information, the managers of companies can identify their closest threat and improve their performance in order to keep their position and not to miss their efficient position in ranking. Some numerical examples for illustration are given.