基于Pareto多目标遗传算法的公共服务设施优化选址研究——以深圳市医院选址为例

A Pareto Genetic Algorithm for Multi-objective Site Search Problem: A Case Study on Hospital Location in Shenzhen City

公共服务设施选址是一类复杂的多目标优化问题。传统遗传算法选址模型多将此问题转化为单目标优化问题,采用二进制或实数编码方式,在小规模数据下进行优化实验,其模型的空间搜索能力不足以满足当前设施优化选址的实际需求。文中基于Pareto多目标遗传算法,设计了行列号组合编码方式及多种重组方法相结合的遗传操作算子,构建了Pareto多目标遗传算法选址模型。实验表明,模型可较好地逼近Fonseca(2)测试函数的凹状解空间前沿,将模型应用于大规模数据环境下的深圳市公共设施优化选址中,取得了较好的实验结果。

Site selection of public service facilities is a complicated multi-objective spatial decision problem that can hardly be solved with traditional methods available from GIS. To reach such location-related decisions, genetic algorithm （GA） is an essential tool. However, as traditional location models based on GA generally use the weighting method and simple binary or real-code encoding strategy, they can hardly be used in settling large-scale site-search problems. The purpose of this paper is to propose an approach based on the modification of genetic algorithm and then to address multi-objective facilities site-search problems in the context of large-scale data. A new encoding strategy based on cells’ index and corresponding genetic operators are designed to construct the location model. The validity of this model is examined by using Fonseca（2） function. Experiment result indicates that the proposed modified GA method using the cells’ index coding strategy and multiple crossover methods can generate approximate Pareto-front. Finally, the proposed model is applied to multi-objective site selection of hospitals in Shenzhen City.