A new model for studying the SO-based pre-trip information release strategy and route choice behaviour

In reality, the predictive traffic information is rarely perfect. It is thus a rational assumption that travellers would not be completely in compliance with the guidance by advanced traveller information systems (ATIS). This article presents a Markovian decision programming (MDP) model for investigating the dayto- day receiving and adjusting process to the pre-trip information released by ATIS at route level. The goal is to seek an optimal information release strategy (IRS) for minimising the overall disutility under the assumption that travellers' route choice is governed by a logit model on the base of adopted travel times. The properties of the model solution are analysed. It is found that there exists an optimal IRS that can drive the flow pattern to a system optimum (SO) if the behaviour adjustment parameter satisfies a condition as 0≤ α<1. The predictive information provided to drivers results in the oscillation of traffic flow among alternative routes when α ≥ 1. The solution of the MDP model is obtained by solving a series of sequential minimisation programs. The model and algorithm are numerically verified on a test network.