Some Evaluations Based on DEA with Interval Data

Abstract

We propose a new approach for obtaining two interval efficiency values with interval data as an extension of DEA. We deal with interval data that can reflect uncertainty in real situations. The two interval efficiency values are obtained from the optimistic and pessimistic viewpoints. Their upper and lower bounds are obtained by two different extreme values in the given interval data respectively. Thus, we formulate four types of efficiency values from two viewpoints with two extreme values in the given interval data. Our emphasis is to obtain two interval efficiency values reflecting uncertainty of the given data. Thus our approach can be described as a kind of interval data analysis. A numerical example is shown to illustrate our proposed approach.