一种基于MAS和GIS平台的城市人口变迁模拟仿真方法  

Simulation of urban population growth with a multi-agent system and GIS platform

作　　者：危辉[1,2] 白宇[1,2] 

机构地区：[1]复旦大学计算机科学与工程系认知算法模型实验室,上海200433 [2]复旦大学波散射与遥感信息教育部重点实验室,上海200433

出　　处：《智能系统学报》2009年第1期44-58,共15页CAAI Transactions on Intelligent Systems

基　　金：国家973计划资助项目(2001CB309401);国家自然科学基金资助项目(60303007)

摘　　要：中国的城市化进程非常迅速,人口增长和迁移是重要特征之一.以地理信息系统、Multi-agent系统和随机过程为基础,建立一个研究人口增长的仿真模型.首先,利用GIS建立一个数字环境,用效能函数的概念在GIS中表示各类资源的效能.其次,详细阐述了一个Multi-agent系统设计和实现,重点是用agent代表某个地块及其属性和行为设计.由于发展的不确定性,设计了一个基于随机过程的行为模型,这也和agent宏观水平上的特性一致.最后,以上海市浦东新区为对象,进行了实验模拟和分析.结果表明这个基于Multi-agent系统和GIS的人口变迁仿真模型,具有模块性和可扩展性方面的优势,可应用到其他的城市规划研究中去.The Chinese economy grew rapidly for more than 20 years, promoting rapid urbanization. New cities emerged and great changes took place in old cities. Population changes complicated city planning, infrastructure construction, housing, and environmental preservation. A model simulating this population transformation was built merging methods from Geographical Information Systems, Multi-agent Systems and Stochastic Processes. First an efficiency function was proposed to represent the effectiveness of a resource within the GIS. Second, a multi-agent system was designed and implemented, with agents representing ground segments and profits acquired after that. Third, static attributes and dynamic behaviors of agents were discussed. To model the uncertainty of urban development better, a behavioral mode was designed based on a stochastic process, consistent with macro-level characteristics of the agent. The combined system and its implementation in a computer were then discussed. Finally experiments, using data from Pudong District of Shanghai, were run and analyzed. This urban evolution model can be adapted to more types of research and owes its good performance to modularity and scalability.

关 键 词：复杂系统仿真 Multi—agent系统 城市化模型 

分 类 号：TP391.9[自动化与计算机技术—计算机应用技术]