一种具有独立子系统的决策单元DEA排序新方法

A new method of DEA efficiency ranking for decision making units with independent subsystems

数据包络分析是(data envelopment analysis,DEA)评价具有可比性决策单元相对有效性的一种工具。在排序过程中,如果同时比较决策单元的最优效率与最劣效率,排序结果会更合理。将基于完全包络面的超效率DEA模型、虚拟决策单元以及非径向C2GS2模型引入具有独立子系统的决策单元DEA排序问题中,提出了改进的能体现决策者偏好或指标重要程度的两阶段模型。最后算例证明该方法的排序结果具有公平性、合理性、完全性。

Data Envelopment Analysis （DEA） is a tool to assess relative efficiencies with a group of decision making u- nits （DMUs）. As in ranking process, it the best and worst efficiency scores of a DMU are compared at the same time, the results will be more reasonable. The CEF （complete envelopment frontier） based on super efficiency DEA model, virtual DMU and non-radial C2GS2 model were used in DMUs with independent subsystems, then a new two-stages model which considering the preference of decision-maker or the importance of different index were proposed. Final, the data example proves that the ranking results of the improved model are fair, reasonable and full.