Network optimization with dynamic traffic information and tolling: a probabilistic model and economic analysis

International audienceA model is developed for traffic network subject to recurrent congestion and incidental disturbances on link travel times with the presence of dynamic traffic information, and two classes of users respectively informed and uninformed. This traffic assignment model is bilayered every informed user chooses his route for the short run in any circumstance, whereas every non-informed user chooses his route on for the long run an average basis. In addition to to dynamic traffic information service, congestion pricing with various tolling strategies like no-tolling, flat tolling or dynamic tolling is also integrated to the model. This model allows us to analyze the combined operation of two traffic demand management tools as well as the interplays between these tools with recurrent congestion, non-recurrent congestion and users' behaviors. The model is applied for a two-parallel-link network with one O-D pair and linear travel time functions and additive random disturbances. Analytical formulae and numerical investigation show that the benefits from such combined traffic management system split into three parts: benefit from information and benefit from congestion pricing and benefit from joint operation. The complementarities are highly efficient at high equipment rate with dynamic tolling