A comparative study of robust efficiency analysis and Data Envelopment Analysis with imprecise data. (15th September 2017)

• Record Type: Journal Article
• Authors: Wei, Guiwu; Wang, Jiamin
• Is Part Of: Expert systems with applications. Volume 81(2017)
• Page Start: 28
• Page End: 38
• Publication Date: 2017-09-15
• Subjects: Data Envelopment Analysis (DEA) -- Performance measurement -- Robust efficiency analysis -- Perfect efficiency -- Potential efficiency -- Imprecise data; Expert systems (Computer science) -- Periodicals; Systèmes experts (Informatique) -- Périodiques; Electronic journals; 006.33
• DOI: 10.1016/j.eswa.2017.03.043
• Languages: English
• ISSNs: 0957-4174

Abstract

Highlights: Integrating the results of CCR efficiency analysis and robust efficiency analysis is critical. Models are developed to determine the bounds of an entity's robust efficiency score. It is the first experimental study to assess efficiency analysis procedures with imprecise data. Computational results suggest that a perfect efficient DMU performs well on average. Abstract: Data Envelopment Analysis gauges the performance of operating entities in the best scenario for input and output multipliers. Robust efficiency analysis is a conservative approach that is concerned with an assured level of performance for an entity across all possible multiplier scenarios. In this study, we extend the robust efficiency analysis procedure to the situation where precise information on some input and output data is unavailable. Perfect efficiency analysis and potential efficiency analysis methods are developed to determine, respectively, the lower and upper bounds of an entity's robust efficiency rating. The concepts of robust efficiency are expanded to classify entities in consideration into three groups: perfectly robust efficient, potentially robust efficient and robust inefficient. Two approaches are presented to convert robust efficiency analysis models into linear programs. It is claimed that Data Envelopment Analysis and robust efficiency analysis together provide a comprehensive picture of an entity's relative efficiency. A computational experiment is conducted to compare the traditional efficiency analysis method with robust efficiency analysis in the presence of imprecise data. The results illustrate that perfect efficiency analysis exhibits a superior power of discrimination than potential efficiency analysis and that an entity recommended by perfect efficiency analysis has a satisfactory average performance.