一种基于虚拟决策单元的排序方法的完善和扩展

本文对文1]中提出的基于虚拟决策单元的排序方法进行了完善和扩展。首先,根据CCR模型,给出了两类特殊的DEA模型,分别是仅有投入数据的DEA模型和仅有产出数据的DEA模型;其次,基于这两个模型,应用上述方法实现了对仅有投入(或产出)数据的决策单元的排序;第三,给出了排序方法中参数a的计算方法;最后,通过修正排序模型,有效提高了排序方法的计算精度。改进后的排序方法避免了两个决策单元因为相对效率值过小而不能排序的情形,其应用范围也进一步扩大。

Perfection and Extension of a Ranking Method Based on the Virtual Decision-making Unit

In the paper, a ranking method based on the virtual decision-making unit is perfected and extended, which is developed in1]. Firstly, according to CCR model, two special models are presented, which are DEA model only with input data and DEA model only with output data. Secondly, based on these two models, decision-making units only with input data or output data can be ranked by applying the mentioned-above method. Thirdly, the computational method of parameter a is given. Finally, the calculation accuracy of ranking method is improved efficiently by modifying ranking models. As a result, the improved method can avoid the situation, in which two decision-making units can’t be ranked as their relative efficiency values are very small. Furthermore, its application range is enlarged.