Evaluation method based on ranking in data envelopment analysis

► We propose Rank-Based Measure (RBM) model to evaluate DMU by using the ranking. ► We construct an algorithm to calculate the optimal solution of RBM model. ► We suggest a new cross efficiency evaluation by using RBM model. Data envelopment analysis (DEA) has been developed as a method to evaluate...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:Expert systems with applications 2013-01, Vol.40 (1), p.257-262
Hauptverfasser: Washio, Satoshi, Yamada, Syuuji
Format: Artikel
Sprache:eng
Schlagworte:
Online-Zugang:Volltext
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Beschreibung
Zusammenfassung:► We propose Rank-Based Measure (RBM) model to evaluate DMU by using the ranking. ► We construct an algorithm to calculate the optimal solution of RBM model. ► We suggest a new cross efficiency evaluation by using RBM model. Data envelopment analysis (DEA) has been developed as a method to evaluate efficiency of Decision Making Unit (DMU). In order to analyze DMU in detail, each DEA model is formulated as a mathematical programming problem utilizing the values of inputs and outputs of all DMUs as coefficients. Each DMU is evaluated by a different weight. Then, the efficiency score of each DMU is determined by using an advantageous weight for itself. In general, the efficiency score is obtained by selecting the most advantage weight. In some real cases, seeking the best ranking is sometimes more important than maximizing the efficiency score. In this paper, we propose a model called rank-based measure (RBM) to evaluate DMU from a different standpoint. We suggest a method to obtain a weight which gives the best ranking, and calculates a weight between maximizing the efficiency score and keeping the best ranking. In order to calculate an efficiency score and the best ranking, we repeatedly solve linear programming problems. Moreover, we apply RBM model to the cross efficiency evaluation. Furthermore, a numerical experiment is shown to compare the rankings and scores with traditional evaluations.
ISSN:0957-4174
1873-6793
DOI:10.1016/j.eswa.2012.07.015