基于邻域互信息的区间DEA交叉效率评价方法

A Cross-efficiency Evaluation Method of Interval DEA Based on Neighborhood Mutual Information

传统的交叉效率评价方法多数是采用一个模型进行求解,并且往往会出现多个决策单元(DMU)均有效,因此不能对所有的决策单元进行充分排序。文章基于区间DEA,同时用进取型、仁慈型和竞合交叉效率模型进行求解,并对竞合模型中的分组方法进行改进,提出利用区间值邻域互信息进行分组,之后利用潜在信息函数对三个模型的效率分数进行集结,得到每个决策单元最终的效率分数及排名结果。最后通过一个算例验证了所提方法的合理性和可行性。

Most of the traditional cross-efficiency evaluation methods use one model to solve problems, and tend to show up multiple efficient decision-making units(DMU), thus it is unable to sufficiently order all decision units. Based on interval DEA,this paper uses aggressive approach, benevolent approach and competing cross efficiency models to solve problems, improves the grouping method in the competition model, and proposes using interval valued neighborhood mutual information to group. Then the paper utilizes the potential information function to assemble the efficiency scores of the three models to obtain the final efficiency score and ranking result of each DMU. Finally, an example is given to verify the rationality and feasibility of the proposed method.