A review of data envelopment analysis models for handling data variations

Abstract

Conventional data envelopment analysis (DEA) models require that the inputs and outputs to be measured deterministically. However, in real world applications, the measurements are subjected to random noise and errors. Ignoring the randomness in the measurement would render an evaluation using DEA unreliable. In response to this particular weakness of DEA, a number of DEA models have been proposed in the literature. This paper's aim is to review the major DEA models for handling data variations. The models include Stochastic DEA (SDEA), Fuzzy DEA (FDEA), and Imprecise DEA (IDEA). Some future research directions in this area will be highlighted as well.