Multiple Criteria DEA-Based Ranking Approach With the Transformation of Decision-Making Units

Abstract

Though various ranking methods in the data envelopment analysis (DEA) context have emerged since the conventional DEA was introduced, none of them has not been accepted as a universal or a superior method for ranking decision-making units (DMUs). The DEA-based ranking methods show some shortcomings as the numbers of inputs and outputs for DMUs increase. To overcome such shortcomings, this paper proposes a two-step procedure of ranking DMUs more effectively and consistently. In the first step, the multi-objective programming (MOP) is applied for the multiple criteria DEA to transform the original DMUs into the new simpler DMUs with two inputs and a single output, regardless of the numbers of inputs and outputs that the original DMUs use and produce. With the transformed DMUs, some conventional DEA based-methods for ranking DMUs are applied in the second step. A numerical example demonstrates the efficient performance of the proposed method.