Pareto–Koopmans efficiency in two‐stage network data envelopment analysis in the presence of undesirable intermediate products and nondiscretionary factors

• Published in: Expert systems 2019-06, Vol.36 (3), p.n/a
• Main Authors: Maleki, Samira, Ebrahimnejad, Ali, Kazemi Matin, Reza
• Format: Article
• Language: English
• Subjects: Algorithms; Computational efficiency; Computing time; Data analysis; Data envelopment analysis; data envelopment analysis (DEA); Efficiency; nondiscretionary factors; Operations research; Pareto–Koopmans efficiency; two‐stage network; undesirable intermediate products
• ISSN: 0266-4720; 1468-0394
• DOI: 10.1111/exsy.12393

Many studies have recently been conducted on the evaluation of system performance with a two‐stage network structure in data envelopment analysis (DEA) literature. One of the topics of interest to researchers has been the mitigation of undesirable products or nondiscretionary factors into their corresponding possible production set (PPS) and their impact on overall efficiency calculations. Determination of decision‐making units (DMUs) with Pareto–Koopmans efficiency status is decisive in identifying benchmark units. The calculated overall efficiency status is compromised when both undesirable products and nondiscretionary factors are present. This work utilizes an axiomatic approach. A novel PPS for a two‐stage network in presence of undesirable intermediate products and nondiscretionary exogenous inputs is introduced. Based on this PPS and by focusing on the principle of mathematical dominance, new models for evaluating overall and divisional efficiencies are presented. In addition, by proposing a two‐step network DEA approach, a necessary and sufficient condition for detection of DMUs with Pareto–Koopmans efficiency status is provided. And by introducing a two‐step algorithm, a novel technique for determining overall efficiency conditions is produced. Finally, the proposed technology is applied to a practical example, and outcomes are discussed.