针对突发事件应急决策的复杂性特征,本文提出一种专家基于领域知识对应急方案集合进行局部评价的群决策模型。以区间数互反判断矩阵对属性进行偏好评价,采用线性规划求解权重区间,并利用集值统计序列得到属性权重向量的平均估计;在专家局部评价基础上,基于目标优化方法对方案群权重进行分配,并同时得到专家权重;通过集结个体局部偏好得到应急决策群组成员对备选方案偏好的一致性排序。最后,以地铁火灾事故为例模拟仿真应急疏散方案的群决策,对模型的实际运用说明了模型的可行性与有效性。
For the complexity of emergency decision making, a local group evaluation model for emergency project set based on the domain knowledge of expert is proposed. First, the expert gives his preference for each attribute, and gets weight ordinal interval by linear programming, and then the comprehensive evaluation value of attribute weight is achieved by set valued sequence. On the base of the expert's local evaluation, the optimization is used to distribute the project group weight, and the expert's weights are obtained too. By aggregating the individual preference, the best project for emergency group decision is available. At last, the paper illustrates metr fire accident to simulate the model application for emergency evacuation project, and the model is proved to be feasible and valid.