Title: Increasing the discrimination power of data envelopment analysis
Authors: Alireza Amirteimoori; Sohrab Kordrostami; Atefeh Masoumzadeh; Mahnaz Maghbouli
Department of Applied Mathematics, Islamic Azad University, Rasht, Iran
Department of Applied Mathematics, Islamic Azad University, Center-Tehran Branch, Lahidjan, Iran
Department of Applied Mathematics, Islamic Azad University, Lahidjan, Iran
Department of Applied Mathematics, Islamic Azad University, Sciences and Research Branch, Guilan, Rasht, Iran
Abstract: In data envelopment analysis (DEA), to discriminate between efficient units, several ranking methods have been proposed by different authors from various points of views. However, despite of the fact that each technique has some advantages and disadvantages, all of them will face the relatively high computational complexity level. Moreover, different ranking procedures yield to different results. Although there are many ranking approaches in DEA literature, there is a need to provide a complete ranking on efficient units with a lower complexity. The current paper proposes a complete ranking method for fully ranking of all DMUs. Furthermore, this approach makes use of a common set of weights for all DMUs. In the proposed ranking approach, the infeasibility and instability problems of the existing methods have been removed. Moreover, the computational effort of the approach is relatively low in comparison to the existing methods. Sample and real cases will be presented for more illustration.
Keywords: data envelopment analysis; DEA; efficiency ranking; super efficiency; common weights; discrimination; infeasibility; instability.
Int. J. of Operational Research, 2014 Vol.19, No.2, pp.198 - 210
Available online: 25 Jan 2014