Evolutionary game-theoretic model for dynamic congestion pricing in multi-class traffic networks

This paper describes an evolutionary game-theoretic learning model for dynamic congestion pricing in urban road networks, taking into account route choice stochasticity and reliability considerations, and the heterogeneity of users, in terms of their value of travel time and real-time information acquisition. The learning model represents the dynamic adjustments of users to travel cost changes which may take place in the day-to-day as well as the within-day timescales. The implementation into a simplified and a real urban road network signifies the important implications of modeling the dynamic and stochastic learning components of users' behavior for accommodating the efficient deployment of congestion pricing schemes.