TY - EJOU AU - Fujii, Hideki AU - Yoshimura, Shinobu AU - Seki, Kazuya TI - Multi-agent Based Traffic Simulation at Merging Section Using Coordinative Behavior Model T2 - Computer Modeling in Engineering \& Sciences PY - 2010 VL - 63 IS - 3 SN - 1526-1506 AB - Traffic system is a typical complex system emerged through interactions among a large number of traffic actors such as car drivers, pedestrians and others. We model each actor as an intelligent agent that can judge and act by itself autonomously, and develop a traffic simulator named MATES, which is based on an intelligent multi-agent model. In this simulator, microscopic behaviors of the traffic actors cause macroscopic traffic phenomena through their interaction. In this research, we newly develop a coordinative behavior model among multiple cars, and implement it to MATES to enhance the reproducibility of traffic behaviors at merging sections. In order to verify the developed simulation model, we apply MATES to the simulation of a merging section near Kandabashi Entrance of Tokyo Metropolitan Expressway. Through the detailed comparison between the results of the simulation and the observation, we confirm practical effectiveness of the developed model. KW - Traffic simulation KW - intelligent multi-agent model KW - coordinative behavior KW - merging section DO - 10.3970/cmes.2010.063.265