Convex cone-based ranking of decision-making units in DEA

Akram Dehnokhalaji, Behjat Hallaji, Narges Soltani, Jafar Sadeghi

Research output: Contribution to journalArticle

Abstract

One of the major research streams in data envelopment analysis (DEA) is ranking decision-making units (DMUs). Utilizing a multicriteria decision-making technique, we develop a novel approach to fully rank all units. Motivated by the convex cone-based total order for multiple criteria alternatives proposed by Dehnokhalaji et al. (Nav Res Logist 61(2):155–163, 2014), we consider DMUs in DEA as multiple criteria alternatives and obtain their total ordering. Initially, some pairwise preference information is provided by the decision maker for units and the concepts of convex cones and polyhedral sets are defined in a DEA framework, correspondingly. We apply a modification of Dehnokhalaji et al. method to extract additional preference information for each pair of units and consequently obtain a full ranking (strict total ordering) of DMUs. The benefit of our approach to their method is that we apply non-radial models to overcome the instability drawback of radial models and their infeasibility occurring in DEA applications. The proposed approach is implemented for two numerical examples, and the accuracy of it is investigated through a computational test.
Original languageEnglish
Pages (from-to)861-880
JournalOR Spectrum
Volume39
Issue number3
Early online date31 Mar 2017
DOIs
Publication statusPublished - 1 Jul 2017

Fingerprint Dive into the research topics of 'Convex cone-based ranking of decision-making units in DEA'. Together they form a unique fingerprint.

Cite this