Scholarlink

<p>As a data-driven non-parametric method, data envelopment analysis (DEA) has been widely employed to evaluate the relative efficiency of homogeneous decision making units (DMUs). Cross efficiency evaluation and network DEA represent two important extensions of traditional DEA, enhancing it along distinct dimensions: the evaluation mode and the internal structure, respectively. Fairness concern, a critical determinant influencing individual behavior, has been separately introduced into cross efficiency evaluation (focusing on fairness in efficiencies among DMUs) and two-stage DEA (emphasizing fairness in efficiencies between two sub-stages within DMUs). In reality, the concern for fairness may exist both across DMUs and within individual DMUs. Thus, this paper integrates dual fairness concerns into the cross efficiency two-stage DEA framework, developing different models under both multiplicative and additive efficiency decomposition methods. The weighted Chebyshev norm is employed to solve the developed models. Moreover, an algorithm is proposed to ensure the uniqueness of the optimal indicator weights. The validity of the proposed model is verified through a numerical example. Finally, we discuss the extension of the proposed model to efficiency evaluation of DMUs with a general two-stage network structure. This study has significant practical implications for fairness-conscious organizations to diagnose operational inefficiencies and develop targeted improvement strategies.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Cross efficiency evaluation in two-stage data envelopment analysis model with dual fairness concerns

  • 期刊论文

  • 发表日期：

    2025年10月8日

查看全文
  • Chen Yingwen,
  • Zhu Zhixi,
  • Zhao Yingying,
  • Zhuang Yahan

摘要

As a data-driven non-parametric method, data envelopment analysis (DEA) has been widely employed to evaluate the relative efficiency of homogeneous decision making units (DMUs). Cross efficiency evaluation and network DEA represent two important extensions of traditional DEA, enhancing it along distinct dimensions: the evaluation mode and the internal structure, respectively. Fairness concern, a critical determinant influencing individual behavior, has been separately introduced into cross efficiency evaluation (focusing on fairness in efficiencies among DMUs) and two-stage DEA (emphasizing fairness in efficiencies between two sub-stages within DMUs). In reality, the concern for fairness may exist both across DMUs and within individual DMUs. Thus, this paper integrates dual fairness concerns into the cross efficiency two-stage DEA framework, developing different models under both multiplicative and additive efficiency decomposition methods. The weighted Chebyshev norm is employed to solve the developed models. Moreover, an algorithm is proposed to ensure the uniqueness of the optimal indicator weights. The validity of the proposed model is verified through a numerical example. Finally, we discuss the extension of the proposed model to efficiency evaluation of DMUs with a general two-stage network structure. This study has significant practical implications for fairness-conscious organizations to diagnose operational inefficiencies and develop targeted improvement strategies.