中文摘要：

采用面向多智能体的建模与仿真方法,对先进的交通信息系统（ATIS）提供的出行前信息下通勤者出行行为进行了研究。首先,基于贝叶斯理论,建立了在出行前信息与以往出行经验作用下,通勤者出行时间感知动态更新模型;进而以信念-愿望-意图（BDI）模型为基础,将每个通勤者-车辆单元刻画为一个具有双层结构的智能体,并采用面向多智能体语言AgentSpeak（L）刻画了智能体的出行行为决策机制;最后,采用面向多智能体编程平台Jason与微观交通仿真平台Paramics相结合的方式进行了相应的仿真试验。结果表明：贝叶斯理论可以较好地解释驾驶员（尤其是通勤者）出行行为动态特性,同时验证了多智能体技术与微观交通仿真技术的结合在驾驶员行为分析中应用的有效性,为驾驶员行为分析提供了新思路。

英文摘要：

Authors studied the commuter＇s travel behavior under advanced traffic information system （ATIS） pre-trip information based on the combination of multi-agent and micro- simulation methods. Firstly, a perception dynamic updating model which commuters update their travel time perceptions by Bayesian theory in light of pre-trip information and their previous experiences. Then, based on the belief-desire-intention （BDI） model, every commuter-vehicle unit was described as an agent which possesses a two-layer architecture and decision-making mechanism of agent travel behavior was specified by extended AgentSpeak（L） language. Finally, a simple simulation test was carried out by combination of Jason and Paramics. The simulation results show that the Bayesian theory can effectively explain the commuter＇s travel dynamics behavior and the combination of multi-agent technology and micro-simulation is a new way of studying driver＇s behavior.