Abstract
In this paper, a multi-agent simulation model of urban traffic behavior using mathematical optimization methods is constructed. Traveler agents decide their routes and means of mobility by optimizing their utility functions. There are too many selection candidates in their decisions to be calculated in practical time, thus modeling techniques which reduces calculation time as well as keeps accuracy is proposed. The decision making process of agents is divided into the common part and the individual part. Results of the former part are shared by multiple agents and the later part is constructed as parallelizable to calculate multiple agents simultaneouslly. Suitable value of the parameter which decide balance of accuracy and calculation load is estimated by computer experiments. Through some simulations, behavior and potential of the proposed model is demonstrated.
