Dealing with generalised inverse DEA models for input-output estimation

Abstract

This article addresses the inverse DEA models to deal with such a question: if a DMU changes some of its input (output) levels among a group of DMUs, to what extent should the unit change outputs (inputs) such that its efficiency score remains unchanged? The article unfolds some problematic aspect of inverse DEA model with saying that inverse DEA models do not work in some situations which should be considered on process of input or output estimation. This is done by proposing a numerical example and mathematical ratiocination. Additional arguments and theorems are provided to overcome to the current drawbacks. This article revises and improves the existing literature. Proposed models are illustrated using numerical examples.