台风灾害下无车群体疏散行为分析与服务设施选址设计

The frequent typhoon and hurricane evacuation activities around the whole world have highlighted the fact that carless evacuation is the key link for conducting successful emergency evacuation practice. In the complex context of disasters, how to make effective carless evacuation plan to improve the efficiency of carless evacuation is a challenging problem in the field of emergency management. This project will start from systematically analyzing the characteristics and regularity of carless evacuees’ typical evacuation behaviors. Based on the consensus that carless evacuees will follow such a trip chain during emergency evacuation that is departing from home, gathering to pick-up facilities, and transported to shelters by evacuation vehicles, the project will propose a series of location models for the selection of important evacuation service facilities, e.g., carless evacuee pick-up facilities and shelters, by using system optimization modeling methods. Especially, the trade-off between the evacuees' risks and the evacuation agency's operation costs occurring in the pick-up facilities, evacuees’ en-route traffic congestion and in-shelter waiting delay occurring in the shelters, and the probabilistic disruption of both these two service facilities against catastrophic typhoon disasters will be integrated in the modeling framework respectively. This research aims to overcome the shortcomings of classic facility location models on worse robustness when facing varied catastrophic events and underestimates of practical evacuation problems. This research can help understand the generation and evolution mechanism of carless emergency evacuation under typhoon disasters and consequently provide support to solve similar carless evacuation problems under other devastating events.

世界范围内频繁发生的台风（飓风）灾害疏散实践表明，无车群体疏散已经成为制约应急疏散活动成功与否的关键。如何在复杂多变的灾害背景下，为无车群体制定有效的疏散规划方案，提高无车群体疏散效率是应急管理领域的前沿课题。从系统分析台风灾害下无车群体关键疏散行为的特征和规律入手，以无车群体“由家出发、集合到集散设施、再由疏散车辆运送至避难所”的疏散活动出行链为主线，以其中发挥关键节点作用的集散设施和避难所选址问题为核心，采用系统优化建模方法，将疏散实践中可能发生在集散设施的无车群体等待、疏散车辆资源有限，避难所相连路网交通拥堵、避难所内排队拥挤，以及两类服务设施都会面临的随机服务失效等问题整合在各自的选址模型中，弥补传统选址方法应对灾害环境鲁棒性较差、对疏散实践问题考虑不足等方面的缺陷。本研究有助于深入理解台风灾害下无车群体应急疏散发生和演化机理，也可以为解决其它灾害下的无车群体疏散问题提供参考。

本课题以台风灾害为应急背景，从前期规划、应急准备、以及为应急管理者提供决策支持的角度，围绕无车群体应急区域疏散规划问题展开研究。.以台风灾害下无车群体应急疏散行为意向调查为基础，系统分析了无车群体的疏散行为特征与规律，弥补以往研究中过多关注有车群体，而对无车群体考察较少的不足之处；通过统计分析和回归建模等方法，刻画了台风灾害下无车群体疏散决策行为的特征与规律，达到预测无车群体疏散需求的目的；.在可能发生集散设施服务失效的情况下，采用EOQ规则度量被疏散群体在集散设施中的等待风险与有限的疏散车辆资源二者之间的均衡，并将其整合到无车群体应急集散设施选址模型中，建立了无车群体集散设施选址模型，不仅能够降低被疏散群体暴露在台风灾害环境中的风险，优化疏散车辆资源的分配，而且使得集散设施选址结果更能充分考虑无车群体的利益，也具有更好的可靠性；.以随机优化建模思想为核心，建立了无车群体避难所选址模型，能够整合台风灾害下避难所发生随机服务失效的各种场景，有助于提高模型在应对复杂的灾害环境方面的鲁棒性；在避难所选址模型中综合考虑与避难所相连路网的交通拥堵以及避难所内的服务排队拥挤，能够有效避免或减少实际疏散活动中可能出现的问题，弥补以往研究中对这些因素考虑的不足之处。.结合GAMS建模系统建立集成优化结果展示与数据统计分析的可视化软件平台，通过提取GAMS计算结果，将抽象繁杂的数值结果以直观可视的方式展现出来，并记录在后台数据统计分析平台中，因此在应对多目标、重复性的结果展示、方案对比等方面具有显著的优势。主要服务于应急区域疏散路网中主要节点设施（如避难所、集合点等）的选址设计以及疏散路径规划等，便于为政府部门和应急管理者提供决策支持。