基于动态信任网络的突发事件模型库系统的研究  

Research on emergency model base systen based on dynamic belief network

作　　者：邵荃[1] 翁文国[1] 袁宏永[1] 

机构地区：[1]清华大学工程物理系公共安全研究中心,北京100084

出　　处：《中国安全生产科学技术》2009年第3期19-24,共6页Journal of Safety Science and Technology

摘　　要：根据突发事件应急决策的需求,本文结合模型网络法和基于信任筛选器的模型组合方法,提出了动态信任网络的模型组合方法,实现了根据不同的突发事件和不同环境选择适合的模型进行模型组合。在运用动态信任网络模型组合方法的基础上,本文应用Agent技术提出了一种智能化较高的突发事件模型库,不仅增强了模型库的适应性和模型库的自学习功能,而且还可以对耦合灾害和复杂事件链的应急处置提供良好的模型决策支持功能。According to the needs of decision-making for emergency events, a dynamic belief network model integration method which combined the model network and filter spaces method was presented in this paper. This method can select suitable models to achieve model integration for different emergency decision-making objectives. Based on this model integration method and agent technology, architecture of emergency model base system was presented, which not only improves the applicability and self-learn function of the emergency model base system, but also provides well decision-making for coupling disasters and event chain.

关 键 词：突发事件 应急模型库 模型组合 智能决策系统 

分 类 号：X924.2[环境科学与工程—安全科学]