面向突发事件应急决策的情景建模方法

Scenario modeling method for unconventional emergency decision-making

情景是非常规突发事件应急决策的基础和依据.针对情景建模中跨学科、跨领域的知识支撑和对事件经验知识依赖的不足,本文从事件本身、事件属性以及属性间关系三个角度对非常规突发事件的情景进行刻画,提出了一种基于知识元的情景表示模型;在此基础上,针对非常规突发事件的不确定性特点,运用贝叶斯网络扩展知识元模型的不确定性表示和推理,构建了适用于情景推演的贝叶斯网络模型,解决了非常规突发事件情景的不确定推演问题;最后,通过舟曲泥石流事件验证本文提出的模型与实际情况相符,能够为非常规突发事件应急决策提供科学有效的辅助支持.

Scenarios are the foundation of unconventional emergency decision. To solve the problem of knowledge support of cross-disciplines and the dependency on experiential knowledge when modeling scenarios, the paper analyzes the scenario structure from the perspective of event,event attribute, and attribute relationships,and proposes a scenario representation model based on knowledge unit model. In view of the uncertainty characteristic of scenarios, the paper extends the uncertainty representation and uncertainty inference of knowledge unit model using the Bayesian network, and then builds a Bayesian network model for scenarios inference.The model solves the problem of uncertainty inference of unconventional emergency scenarios. Finally, the experiment of Zhouqu debris flow proves that the proposed model is consistent with the reality and that it can support the unconventional emergency decision scientifically and effectively.