A bi-objective data envelopment analysis model using a common set of weights

Conventional data envelopment analysis (DEA) stems from benefit/cost theory to evaluate the technical efficiency of units as the ratio of weighted outputs to weighted inputs. DEA in this regard allows each unit to maximize its own efficiency under total input and output weights flexibility. However, the freedom of weights in DEA leads to distinct weights for each factor that is in question in some situations, in particular performance evaluation of centralized systems. In addition, the lack of power to discriminate between the efficient units is another problematic issue in DEA models. In this paper, we develop a bi-objective common-weights DEA model involving two evaluation objectives at the simultaneous time; that is, minimizing the maximum and sum of inefficiencies to generate a common set of weights as well as to improve the discriminating power. We also highlight the role of lower limit on weights in the proposed model. A case study of banking industry is finally presented to illustrate the efficacy of the proposed approach.