传统数据包络分析(DEA)方法以静态``黑箱"的视角分析系统效率, 无法体现出决策单元(DMU)生产的实际运作情况, 从而造成结果与实际效率存在较大偏差. 文章引入合作博弈思想, 考虑系统的内部结构和不同时期的动态联系, 构建动态网络系统的松弛变量方法(SBM) 模型. 同时, 探讨系统存在非期望要素情况下的效率评价问题. 此外, 基于多目标规划理论, 构建效率分解模型. 分析不同时期和不同过程对整体系统效率的影响, 并解决了效率分解方案不唯一的问题. 最后, 给出一个应用实例来说明模型的可行性与实用性.

Conventional data envelopment analysis (DEA) evaluates efficiency of system from the static and the ``black box" perspectives. It can not effectively reflect the operational performances of decision making units (DMUs) so that some deviations are produced. By taking the internal structure of system and dynamic relationship in different periods into consideration, this paper constructs a slacks-based measures (SBM) model of dynamic network system based on the cooperative game. Then, it discusses efficiency evaluation of system, which considers undesirable factors. Moreover, the model of efficiency decomposition is constructed based on the theory of multi-objective programming to analysis effects of different periods and different processes, and make the decomposition method uniqueness. Finally, the empirical case is introduced to verify feasibility and practicability of the models.