【摘要】 为提升非常规突发事件应急决策的敏捷性与科学性，本文提出基于“情景—应对”模式的动态决策框架。通过界定承灾体、致灾因子等情景要素，构建耦合关系模型，并结合动态贝叶斯网络，实现情景演化的时序概率推理；创新性地融合结构相似度、逻辑属性与语义属性，计算全局情景相似度，实现目标情景与历史案例的精准匹配。以地铁水灾事件为案例，通过虚拟仿真验证了模型有效性：基于情景演化推演生成的分级响应方案与实际处置需求高度契合。研究表明，该框架可突破传统“预测—应对”模式局限，为非常规突发事件应急决策提供科学支撑。更多还原

【Abstract】 To improve the agility and scientific rigor of emergency decision-making for unconventional emergencies, this paper proposes a dynamic decision-making framework based on the “scenario–response” paradigm. By defining scenario elements such as hazard-affected bodies and disaster-inducing factors, it constructs a coupling relationship model integrated with a dynamic Bayesian network(DBN) to achieve temporal probabilistic reasoning of scenario evolution. The study innovatively combines structural similarity, logical attributes, and semantic attributes to calculate global scenario similarity, enabling precise matching between target scenarios and historical cases. Using a subway flooding event as a case study, it verifies the effectiveness of the model through virtual simulation: the graded response plan generated through scenario evolution reasoning aligns closely with actual response requirements. The findings demonstrate that the proposed framework overcomes the limitations of the traditional “prediction–response” model and provides scientific support for emergency decision-making in unconventional emergencies.更多还原