Fictitious play for finding system optimal routings in dynamic traffic networks

We introduce a novel procedure to compute system optimal routings in a dynamic traffic network. Fictitious play is utilized within a game of identical interests wherein vehicles are treated as players with the common payoff of average trip time experienced in the network. This decentralized approach via repeated play of the fictitious game is proven to converge to a local system optimal routing. Results from a large-scale computational test on a real network are presented.