There exist several approaches for deriving a common set of weights in data envelopment analysis (DEA) literature. However, most of these approaches are based on complicated models. In this paper, a new practical approach is proposed to provide a common set of weights. The results of the new approach are compared with some of the existing models through several numerical examples. Manuscript profile

This paper proposes improvements and revisions to a recent approach in a voting system, and provides an effective approach with a stronger discriminate power. For this purpose, the advantage of a linear transformation is utilized to redene a previously used concept of virtual worst candidate, by incorporating the existing weight restrictions. Then, the best score of this virtual candidate is used as a lower bound for the scores of the candidates in the models. It is shown that the proposed approach overcomes some drawbacks in the existing models. The approach determines interval eciencies which can be used in ranking the candidates. Numerical examples illustrate some advantages of the proposed approach. Manuscript profile