An urban traffic simulation system based on multi-agent modeling
Han, Zhi,Zhang, Kun,Yin, Hongpeng,Zhu, Yong
DOI: https://doi.org/10.1109/CCDC.2015.7161966
2015-01-01
Abstract:The compute simulation technology plays a key role in developing Intelligent Transportation System (ITS). The agent-oriented technology provides the advantage in modeling the system and reflecting the ITS components' interactive and self-managing behavior than the conventional simulation technologies. In this paper, a multi-agent traffic simulation system is presented and implemented in NetLogo platform. In this system, the urban traffic components including the vehicles, road sections, intersections are abstracted to agent models. Each agent has the basic abilities of obtaining knowledge, autonomy, interaction and communication. In the road agent model, the traffic flow forecasting ability is integrated to induce vehicle agents' action and help intersection agent control the traffic signal. Each intersection agent is the abstracted model of the signal controller and the traffic situation of intersection. In each intersection agent model, the signal control function is implemented by analyzing the real-time and predicted traffic flow information from the interaction with related road agents. Simulation result shows that the proposed multi-agent traffic system performs well in reflecting the evolution of dynamic traffic system's behavior. Besides, it shows the advantage in reducing the vehicles' delay time of the road network by the interaction-based signal control function of the intersection agent.