综合网络DEA模型有效性理论分析  

Efficiency Theory of Generalized Network DEA Model

作　　者：李凌[1] 刘建永[1] 付成群[1] 伍中军[1,2] 

机构地区：[1]解放军理工大学工程兵工程学院,江苏南京210007 [2]防化指挥工程学院,北京102205

出　　处：《系统工程》2010年第1期53-57,共5页Systems Engineering

摘　　要：针对综合DEA模型无法反映系统内部子过程的相互作用对绩效的影响,提出一种综合网络DEA模型,并给出该混合模型的相关定义及定理,得到综合网络DEA模型由参数取值不同,各(弱)网络DEA有效性间的关系;指出了综合网络DEA有效的决策单元,也为评价决策单元总体有效性的综合DEA有效,同时该单元内部子过程也均为综合DEA有效;证明决策单元为(弱)综合网络DEA有效时,评价决策单元总体有效性的综合DEA最优解与相应综合多目标规划的(弱)Pareto解之间的对应关系。这些结果进一步丰富了DEA有效性理论。Considering that generalized data envelopment analysis （DEA） can not reflect the reciprocity of subsystems and its influence on system efficiency, we propose a generalized network DEA model. The correlative definition and theorem of the hybrid model are given. According to different parameters of generalized network DEA, the （weak） efficiency relationship of each kind of network DEA is obtained. This study indicates that the decision making unit （DMU） being generalized network DEA efficient is also generalized DEA efficient when the model doesn＇t consider the subsystems, and the subsystems of the DMU are all generalized DEA efficient synchronously. At the same time, the relationship between DEA optimal solution which doesn＇t involve the subsystems and Pareto solution of generalized multi-objective programming was proved. These results further enrich DEA efficient theory.

关 键 词：综合网络DEA 综合DEA 网络DEA 有效性Pareto解 

分 类 号：N945[自然科学总论—系统科学]