摘要
由于从最优与最劣两个维度建立的CCR模型、第二目标模型和交叉效率模型权重的不唯一性,往往会造成排序结果的不同甚至是较大的差异,降低了排序的可接受性。本文提出的基于可辨识矩阵的二级DEA效率评价模型提高了决策单元排序的可接受性。本文首先以六种模型的效率分数构造连续值信息系统并建立优势关系辨识矩阵,从而进行属性约简并基于此进行对象的优势度比较;进一步通过建立序数型有向距离指数获得全序化结果。最后通过一个算例验证了该方法排序的有效性。
Because of no uniqueness of weight,the CCR,the second goal model and the cross-efficiency model based on the pessimism and optimal will lead to different ranking results,with even large discrepances,which will reduce the acceptability of ranking.Therefore,this paper developes two-grade composite efficiency model of data envelopment analysis based on discernable matrix and ordinal directional distance index to improve the ranking result acceptability.First it sets up the continuous value information system and discernable matrix by using efficiency score of six models,thus comparing the prior degree of different objects based on the attribute set by attribute reduction.Moreover,by using developed ordinal directional distance index on paralleling objects for linearization.At last,a numerical example is provided to test the validity of the model.
出处
《系统工程》
CSSCI
CSCD
北大核心
2015年第5期147-151,共5页
Systems Engineering
关键词
DEA
粗集
序数型有向距离指数
DEA
Rough Sets
Ordinal Directional Distance Index
