摘要
城市空间的扩展是一个非常复杂的现象 ,传统的研究方法是在相同的层次上建立起因变量和自变量的某种函数关系 ,但很难表现出宏观与微观之间的相互反馈的作用关系。本文提出了一种外部模型 (描述宏观现象 )与内部模型 (微观决策 )相结合的Agent模型 ,试图从人相互作用的关系来理解城市的扩展现象的机理。主要讨论了构造外部模型和内部模型及其相结合的方法、Agent观察和探测外部信息的方式以及构建模拟系统等问题。在此基础上基于Java和Swarm平台开发了模拟系统UrbanSwarm。实例对北京市土地利用变化动态模拟 ,结果表明 ,利用Agent模型动态模拟土地利用变化是一种可行的方案。
The phenomenon of urban land expanding is too complex that the traditional model, which based on function between independent variables and outputs at the same level in system, can not describe the interrelation between macro phenomena and the selection of micro agent in different levels. This paper presentes an agent modeling for the simulation of urban land using & covering by combing exterior model with interior model method. The authors attach importance to such questions as how agent learning 、observing or detecting the information on exterior space and reacting to and interacting with it, which drives the system evolution continuously. A simulating system named UrbanSwarm is been constructed and developed based on Java & Swarm platform. The result of simulation on Beijing shows that this methodology is promising.
出处
《东华理工学院学报》
2004年第1期80-83,共4页
Journal of East China Institute of Technology
基金
北京市自然科学基金 (40 2 2 0 0 6)