摘要
将空间信息与Pareto多目标遗传算法相结合可解决具有多个相互制约目标准则的GIS选址问题。以NSGA-Ⅱ算法为基本算法,以空间选址涉及的服务人口密度、交通成本、道路可达性等因素定义多个目标函数,通过表达空间相互作用的权重矩阵将空间信息融合在NSGA-Ⅱ中,形成GIS空间对象的多目标优化选址算法流程,并以山东省10个流行病监控点最优位置的选址为案例进行对比分析。实际应用表明,与普通遗传算法、普通NSGA-Ⅱ算法相比,融合空间信息的多目标遗传算法可有效解决复杂的空间优化选址问题,不仅可以收敛到Pareto最优集,而且解集的分布性更好,算法也具有很好的稳定性。
Combining spatial information with Pareto multi-objective genetic algorithm can solve GIS location problem, which has mutual restriction of multiple objective criteria. Based on NSGA-Ⅱ, this paper used the service population density, transportation cost and road accessibility to define multiobjective functions at first. Secondly, the paper fused spatial information into NSGA-Ⅱ algorithm through the form of weight matrix, which was used to express spatial interaction, and put out the process of the multi-objective optimal location algorithm of GIS spatial objects. Lastly, this paper illustrated and analyzed the algorithm by searching optimal location of 10 monitoring places for epidemic disease in Shandong Province. The application results show that the improved multi-objective genetic algorithm, compared with the normal genetic algorithm and normal NSGA-Ⅱ algorithm, can solve complex spatial location problem effectively. Not only can this algorithm converge to Pareto-optimal set, but the optimal set has better distribution and the algorithm has better stability.
出处
《地理空间信息》
2018年第3期26-29,共4页
Geospatial Information
基金
国家自然科学基金资助项目(41471322)
山东省自然科学基金资助项目(ZR2012DM010)
