Two-stage Network Models in DEA and DEA-R with Desirable and Undesirable Outputs

Abstract

This paper proposes two-stage network models within the frameworks of Data Envelopment Analysis (DEA) and DEA-R, designed to accommodate both desirable and undesirable outputs. These non-radial models are developed under the assumption of constant returns to scale. By employing a multi-objective linear programming approach within non-radial additive DEA and DEA-R models, this study introduces a novel method for identifying suitable benchmarks for decision-making units, even in the presence of undesirable outputs [10] to [13]. The proposed models evaluate decision-making units based on the level of inefficiency within two-stage networks, with the calculation of total inefficiency in DEA and DEA-R serving as a criterion for assessing units in two-phase networks with undesirable outputs. The paper concludes with a case study on storage centers for electric power supply equipment in Fars province, illustrating the practical application and effectiveness of the proposed models.