Abstract: Classic Multi-Criteria Decision Making（MCDM） method is unable to deal with the dynamicity in real world. Also, it is difficult to reflect the feedback action that Alternative Cluster（AC） dominates Criteria Cluster（CC）. To overcome the above mentioned drawbacks, in this paper a multi-criteria group decision making approach with variable weights in a dynamic environment is suggested by synthesizing both data envelopment analysis and analytic network process. Compared with classic MCDM method, the outstanding advantages for the approach lie in following three points. A dynamic evolvement mechanism for MCDM model is proposed. A novel group information distilling way is given by analyzing these statements that each alternative lies in criteria to express expert preference, which ensures no loss of every information. It may realize the decision with variable weights for alternatives, therefore, nonlinear interaction relation between AC and CC could be well reflected. The approach is validated to be feasible and scientific and can be well applied to solve the real world dynamic decision issues, such as selecting dynamic disaster relief alternatives, supplier dynamic evaluation, etc.

Key words: dynamic environment, group experts, Multi-Criteria Decision Making（MCDM）, variable weights

摘要： 为克服经典多准则决策（MCDM）方法不适应动态的决策环境、难以反映方案集对准则集的非线性反馈效应等方面缺陷，通过运用网络分析和数据包络分析技术，提出一种动态环境下的群组专家多准则变权决策方法。较之于经典MCDM方法，新方法主要创新之处在于：给出了MCDM模型的动态演化机理；通过专家对方案所处准则状态予以有偏好（无偏好）判断，提出一种保证信息无损的群组专家信息提取方式；实现了对方案的变权评价，有效反映出蕴含在系统内部的准则集与方案集的非线性交互作用关系。实例验证结果表明，所提方法是科学可行的，能够有效解决救灾方案动态优选、供应商动态评价等实践问题。

关键词: 动态环境, 群组决策, 多准则决策, 变权