基于Agent的Repast仿真分析与实现

基于复杂适应系统(CAS)理论在计算机上建立模型是研究社会复杂系统的一个基本方法.选择一个合适的基于A-gent的建模与仿真平台能降低社会科学工作者的研究难度.Repast是目前使用较多的基于Agent的建模与仿真平台之一,主要应用于社会科学仿真领域.简要介绍了Repast平台的背景,从建模与仿真的两个角度分析了Repast仿真的特点,并介绍了Repast的主要类库.从仿真实现的基本步骤,Repast内含的时间表机制和显示机制,仿真程序框架几方面探讨了Repast仿真模型的设计与实现流程.从CarryDrop模型的实现结果,能看出基于Repast平台能较方便地实现基于Agent的仿真.最后指出Repast平台突出的特点是能充分利用.Net和Java类库的众多功能.

Implementation and Analysis of Agent Based Simulation on Repast Toolkit

Since computer simulation based on complex adaptive system theory is a basic approach of social complex system research, a suitable modeling toolkit should be considered. Repast is one of most popular agent - based modeling and simulation toolkits, which is specifically designed for social science applications. The background of Repast is introduced in this paper. The main class packages of Repast are also briefly introduced. The characters of Repast simulation are analyzed from modeling and simulation aspects. The implementation of Repast simulation is diseussed from several aspeets, such as simulation steps, program framework ,meehanism of sehedule and display. A CarryDrop Repast Model is given as an example, which shows that agent -based simulation on Repast toolkit is convenient. In the end, the trait of Repast Toolkit is pointed out.