Managing Day-to-Day Network Traffic Evolution via an Altering Ex-Post Information Release Strategy

Abstract

Providing travelers with full ex-post travel time information from day to day can only make a disequilibrium network evolve to the user equilibrium, rather than the system-optimal or other better-off states. To make best use of the ex-post information, this paper suggests an altering information release strategy to optimize the day-to-day disequilibrium traffic evolution. According to present strategy, the ex-post travel time information is not released directly but is properly altered in advance, with altering volume falling into travelers’ memory or perception error scopes so as not to incur distrust. The problem is formulated as dynamic programming in which the classical proportional-switch adjustment process is applied to describe travelers’ day-to-day rerouting behavior against the released information. The properties of the model are analyzed sufficiently. Due to poor properties, the problem is treated as a bound constrained optimization problem, and a constrained compass search method is suggested to solve it. Numerical examples based on two networks are given to perform and characterize the current methodology.