粒子群算法在地震灾害应急避难所选址中的应用:以云南省鲁甸县文屏镇为例

Analysis of earthquake emergency shelter location selection based on particle swarm optimization algorithm:Wenping of Ludian in Yunnan Province as a case

以云南省鲁甸县文屏镇为例,在政府指定的地震灾害应急避难所候选场址的基础上,结合基础地理数据和人口分布数据,建立以最短化避难总距离和最小化避难所总容量为目标、不超出最大服务距离和不超过避难所最大容量为约束的多目标模型,采用改进的粒子群算法对其进行解算,并结合地震灾害避难所要求建立综合评价指标,对避难所分配结果进行分析评估.结果表明:共得到12种优化的方案;从安全性角度方案6为最优方案,从收容性角度方案12为最优方案,从成本投入方面方案1为最优方案;综合3个指标考虑方案12为最优方案.各种不同评估方案能为该区域避难所的责任区划分提供一定的理论基础.

Locations and allocations of earthquake shelters were examined for Wenping,Ludian,Yunnan Province.Geographic data,population data,candidate shelter data from local government were used to model location and allocation with two objectives（minimizing total evacuation distance and total shelter area）and with two constraints（maximal service radius and maximal capacity）.The model was then used to obtain schemes of location and allocation with MPSO algorithm.A comprehensive evaluation index was used to analyze simulation results.Schemes 6,12,1,12 were found to be best in safety,capacity,cost and comprehensiveness respectively.These data would provide some theoretical basis for earthquake shelter location and allocation.