基于多目标规划和智能体仿真的社区疏散方案

Community evacuation scheme based on multi-objective programming and agent simulations

快速疏散灾民是地震发生后短时间内的首要任务之一,提前规划避震场所、设计社区人员疏散方案能为灾后应急提供有效保障。该文提出了一种基于多目标规划和智能体仿真的社区人员疏散方案设计方法,综合考虑救援机构可达性、交通便利性、避震场所容纳能力等因素,基于多目标规划设计不同避震场所个数下的最优疏散方案,结合Pathfinder软件进行人员疏散行为仿真实验验证。结果表明:粒子群算法对求解避震场所选址和分配的多目标规划问题收敛性好,适用于快速建模计算;指定避震场所的疏散方案具有引导性,能大幅缓解自由疏散出现的"超载"现象,同时保障全员尽快抵达避震场所。

Residents need to be quickly evacuated from buildings after earthquakes.However,shelters need to be built in advance and community evacuation schemes need to be developed to provide effective emergency response after an earthquake.This paper presents a community evacuation scheme design method using multi-objective programming and agent simulations.The optimal evacuation schemes are developed based on the accessibility of rescue agencies,traffic conditions,shelter capacity and other factors based on various numbers of earthquake shelters using multi-objective programming with the evacuation behavior simulated using the Pathfinder software.The results show that the particle swarm optimization algorithm has good convergence for the multi-objective programming problem of location selection and allocation of earthquake shelters,and is suitable for quick modeling and calculation.The evacuation scheme of designated earthquake shelters is instructive,which can greatly alleviate the"overload"phenomenon of free evacuation scheme and ensure that all personnel can arrive at the shelters as soon as possible.