期刊
ANNALS OF OPERATIONS RESEARCH
卷 302, 期 1, 页码 173-210出版社
SPRINGER
DOI: 10.1007/s10479-021-04027-x
关键词
Data envelopment analysis (DEA); Cross efficiency; Reciprocal behaviors
资金
- National Natural Science Foundation of China [71901178, 71904084, 71910107002, 71725001]
- Sichuan Provincial Social Science Foundation [SC20C052]
- Natural Science Foundation for Jiangsu Province [BK20190427]
- Social Science Foundation of Jiangsu Province [19GLC017]
This study considers reciprocal behaviors among DMUs, determines positive or negative behaviors using a novel threshold, and develops a game-like iteration process to adjust evaluation strategies. The optimal ultimate cross efficiency score with reciprocal behaviors is then calculated.
Data envelopment analysis (DEA) has proven to be a powerful technique for performance evaluation since its inception. Since the traditional DEA approaches lack discrimination power among efficient decision-making units (DMUs), the cross efficiency method has been proposed for peer appraisal in the literature. However, the previous cross efficiency approaches imposed a single and identical evaluation strategy across all DMUs simultaneously. In addition, all the related studies have considered a static issue without the dynamic alternation of evaluation strategies. In this paper, the reciprocal behaviors among DMUs are considered to address the cross efficiency evaluation, and a novel threshold value is used to determine positive or negative reciprocal behaviors by comparing the peer-evaluated efficiency with the threshold value based efficiency. This study assumes that a DMU would show positive behavior and apply a benevolent strategy toward other DMUs that evaluate it friendly, while it also shows negative behavior and apply an aggressive strategy toward DMUs that evaluate it hostilely. Furthermore, a game-like iteration process is developed for each DMU to determine and further adjust its evaluation strategy toward other DMUs in the evaluation process. Afterward, we calculate the optimal ultimate cross efficiency score with reciprocal behaviors. Finally, the proposed approach is applied to both a numerical example and an empirical study of 31 Chinese manufacturing industries to demonstrate its usefulness and efficacy.
作者
我是这篇论文的作者
点击您的名字以认领此论文并将其添加到您的个人资料中。
推荐
暂无数据