Abstract: This study aims to explore the regional evacuation of mass vehicles before the predictable natural disasters’ landfall including typhoon, hurricane, and flooding. A two-level framework based on optimization and micro-simulation is presented to model the traffic management of regional evacuation. The upper level model based on the revised cell transmission theory is capable of identifying the candidate set of optimal evacuation plans in a timely manner even for large networks. Then, the responsible system users can finalize the control plans with a pre-calibrated simulator that can realistically replicate some critical operational features and driver response difficult to fully capture with analytical formulations. The results of case study show that the proposed optimization-simulation framework has a great potential to improve the efficiency of evacuation planning and operations.

Key words: regional evacuation, traffic management, optimization, simulation

摘要： 针对台风、飓风等可预报自然灾害，在其登陆城市之前一定的安全时限内，大规模区域路网疏散的交通管理问题进行了研究。基于最优化理论与微观交通仿真相结合的方法，构造了面向区域疏散交通管理的双层建模框架。上层模型（最优化层）以修正元胞传输理论为基础，构建了适用于大规模路网疏散的交通管理方案优化模型，其输出的疏散交通管理策略备选集合，输入到下层模型（微观交通仿真模型）中，从而在微观层面考察驾驶员行为反应等上层模型难以捕捉的疏散特性，进而制定最优的疏散交通管理方案。案例分析的结果表明，该框架兼具最优化方法与微观交通仿真方法的双重优点，能够较好地服务于区域疏散交通规划与具体的区域疏散交通管理实践。

关键词: 区域疏散, 交通管理, 最优化, 仿真