基于GIS和多智能体的城市人口分布模拟  

Simulation for Urban Population Distribution Based on GIS and Multi-agent

作　　者：康停军[1] 张新长[1] 赵元[2] 王海鹰 张维[1] 

机构地区：[1]中山大学地理科学与规划学院,广东广州510275 [2]华南农业大学信息学院,广东广州510642

出　　处：《中山大学学报（自然科学版）》2012年第3期135-142,共8页Acta Scientiarum Naturalium Universitatis Sunyatseni

基　　金：国家自然科学基金资助项目(40971216;41071246)

摘　　要：传统的城市人口空间分布方法仅能在数量上达到较高的模拟精度,但是不能对城市人口分布表现的居住分异特性进行描述。针对常用的城市人口空间分布模拟方法的不足,构建了基于多智能体的城市人口分布模拟模型,模型由影响要素、智能体、决策规则等组成。居住用地划分为均匀的居住元胞,初始状态下,每个居住元胞分布相同数量的智能体(家庭为单位),智能体在各自居住压力作用下做出是否迁居的决策,如果迁居则选择合适的居住元胞,通过智能体的不断决策、协商、迁居模拟城市人口分布的变化。实验结果表明,相对于传统的重力模型,本文模型所模拟的结果不仅具有较高的精度,而且模拟的城市居住分异现象与现实居住格局相符合,为城市人口分布模拟提供了新的思路。The importance of spatial distribution of urban population data is analyzed. Aiming at the disadvantages of the usual methods of urban population distribution, a simulation model for urban population distribution is developed with integration of multi-agent and population migration theory （ Simulation Mod- el for Urban Population Distribution-SMUPD） , which consists of external environment, multi-agent and rules. Residential land has been divided into regular residential cells, based on the initial condition; each residential cell has the same amount of agent. Agents decide whether to migrate or not, and select the appropriate residential cell to make migration. The population distribution is formed by means of the constant decision-making, consult and migration of agents. Experiments indicate that, the simulation result of SMUPD not only has higher accuracy, but also is close to the actual residential segregation pattern. In conclusion, the SMUPD provides a new way of modeling urban population distribution.

关 键 词：人口分布 居住分异 多智能体 模拟 广州 

分 类 号：P208[天文地球—地图制图学与地理信息工程]