A DEA model under correlated polyhedral uncertainty sets

Precise, well-defined and deterministic data are the underpinning of conventional data envelopment analysis (DEA) models. However, a plethora of real-world applications includes uncertainty to some extent. Robust optimisation (RO) approaches have been used far and wide to cope with uncertain data in practical and engineering optimisation problems along with capturing more robust solutions against data fluctuations. In the existing literature, uncertain data are assumed to be independent. This paper adopts aRO approach under polyhedral uncertainty sets to develop a new robust DEA model which takes the correlation between the input and output data. As a consequence, it is envisaged that the price of robustness of DEA models decreases when there are significant correlations between the uncertain data.

Keywords:

Data envelopment analysis , Robust optimisation , Correlated polyhedral uncertainty set