摘要
针对传统人口空间化模型普适性较差,且数据处理较依赖人工操作等缺陷,该文提出了一种基于多智能体模型从微观个体的居住选择入手,在以往学者研究的基础上同时考虑居民居住建筑的内、外部环境因素模拟人口居住空间分布。创新性地以人口智能体为中心,结合GIS利用地理国情数据、POI兴趣点数据、夜间灯光数据、遥感影像数据来模拟居住环境,将心理学和经济学观点引入到智能体对居住地选择的决策行为中,结果表明,基于建筑物内、外部环境的多智能体人口空间分布模型,可以获取高空间分辨率的人口分布情况,且更接近于实际人口分布。
Aiming at the flaws existing in the traditional spatial model of population,such as poor universality and the dependence of data processing on manual operation,this paper put forward a model based on multi-agent for the analogue simulation of human living environment,started from the residential choice of macro individuals to simulate the living spatial distribution of people through considering the internal and external environment of their residential buildings while basing on the previous researches of scholars.Centered on population agent,psychology and economics perspectives were introduced into the decision-making behavior of agents in their choice.The research result showed that the multi-agent spatial distribution of population based on the internal and external environment of buildings could obtain the population distribution with high spatial resolution,and it was closer to the actual population distribution.
作者
董春
尹诗画
张玉
DONG Chun;YIN Shihua;ZHANG Yu(Chinese Academy of Surveying and Mapping,Beijing 100036,China;School of Geomatics,Liaoning Technical University,Fuxin,Liaoning 123000,China)
出处
《测绘科学》
CSCD
北大核心
2019年第8期113-119,共7页
Science of Surveying and Mapping
基金
国家自然科学基金面上项目(71773117)
国家社会科学基金项目(17BJL004)
国家社会科学基金重大项目(18ZDA066)
中国测绘科学研究院基本科研业务费项目(7771718)
关键词
人口
多智能体
空间分布
城市
population
multi-agent
spatial distribution
city
