面向大型活动的交通应急预案快速生成与动态优化方法

Rapid Generation and Dynamic Optimization of Traffic Emergency Plans for Large-scale Events

现阶段大型活动突发事件下的交通组织应急预案多依赖人工经验,缺少针对性和定量化应急措施。针对预案的快速匹配问题,提出利用交通应急预案库中的历史案例提取突发事件特征属性,基于案例推理和朴素贝叶斯分类快速生成初始预案,并依据贝叶斯分类算法所得后验概率选取属性不完备情况下的最佳匹配预案。在此基础上,基于改进的规则推理建立交通应急预案库知识库和规则库,并采用正向推理修改预案内容。为构建评价指标体系,基于模糊层次分析法对预案应急能力水平和突发事件严重程度进行模糊评价。以北京2022年冬奥会为仿真案件,试验结果表明该方法可以快速生成最佳匹配预案并完成动态调整与完善。

The generation and dynamic optimization technology of traffic-emergency plans under large-scale emergencies rely on experience,and the manual plan lacks targeted and quantitative emergency measures.The feature attributes of emergencies are extracted to quickly generate the optimal matching plan.The initial plan is quickly generated based on case-based reasoning(CBR)and naive Bayesian classification and selected according to the posterior probability obtained by the Bayesian classification algorithm.The knowledge base and rule base of traffic emergency plans are established based on rule-based reasoning(RBR),and the content of the plan is modified by forwarding reasoning.The evaluation-index system of emergency-response capability and emergency severity is established based on the fuzzy analytic hierarchy process(FAHP).Taking Beijing2022 Winter Olympic Games as the case background,the method can quickly generate the optimal matching plan and realize dynamic adjustment and improvement.