基于Agent技术的交通流仿真研究

Research on Traffic Flow Simulation Based on Agent Technology

基于我国当前的交通问题以及传统交通管理系统的特点,在前人的理论基础上,文中将agent技术应用于智能交通系统,运用系统中各个元素进行agent建模。介绍了KQML语言规范并详细探讨了KQML消息机制在agent通信中的应用,在此基础上选择Java编写agent消息类,构建仿真系统。系统在周期边界条件下进行单车道交通流仿真,真实地再现了流量-平均速度、密度-流量之间的关系。仿真结果表明,基于agent的智能交通仿真能够细致地刻画交通流特性和交通实体的微观行为。

In view of traffic problems of our country and features of traditional traffic management systems,applies agent technology to intelligent traffic system and builds models of traffic dements. It discusses the KQML message mechanism in the communication of agents and achieves the message class and simulation system with java, and then it carries out a lane simulation, which can reveal the relationship of flow- average velocity and density- flow. The result suggests that the microscopic behavior of traffic entities and the features of traffic flow will be depicted clearly by traffic simulation based on agent technology.