基于遗传多目标优化的居民地自动选取  

Automated Selection of Settlement Based on Multi-objective Optimization by Genetic Algorithm

作　　者：李国辉[1,2] 龙毅[1,2] 周侗[1,2,3] 许文帅[1,2] 陈林[1,2] 

机构地区：[1]南京师范大学地理科学学院,江苏南京210023 [2]虚拟地理环境教育部重点实验室,江苏南京210023 [3]南通大学地理科学学院,江苏南通226007

出　　处：《地理与地理信息科学》2014年第4期59-62,101,共5页Geography and Geo-Information Science

基　　金：国家自然科学基金项目(41171350;41271449;41301514);江苏省高校自然科学研究项目(13KJB170020)

摘　　要：基于遗传算法思想,针对地图综合中离散面群居民地的选取,考虑地图综合需要保持的统计、专题、拓扑和度量等4类关键信息,提出一种基于遗传多目标优化的离散面群居民地的自动选取模型。该模型考虑了面积、密度、属性意义等多种约束指标,通过遗传迭代,逐步优化得到理想结果。采用1∶10 000居民地数据进行了地图综合实验,结果表明,该算法能够有效保持地图所包含的关键信息。This paper presents an automated selection method of settlement based on multi-objective optimization by genetic algorithm which is concentrated on the discrete polygon of settlements.This selection method aims to keep four types of key information during generalization,which is statistical information,thematic information,topological information and metric information.It takes area,density,importance,etc.as the constraint indicator so that the genetic algorithm can run in a specific direction and as long as the number of genetic iteration is large enough,the system would get the global optimal solution.Then the whole examination is taken to test the feasibility of the method,and the result shows that this method can maintain four types key information to some extent.However,the distribution range factor is not kept very well.This problem can be solved by increasing the significance of the border settlements.In addition,some other knowledge of the settlements can be added to the genetic algorithm to improve the accuracy of the results.

关 键 词：地图综合 居民地选取 多目标优化 遗传算法 

分 类 号：P283.7[天文地球—地图制图学与地理信息工程]