System and user performance in multi-class networks under multi-criteria route guidance systems

This paper investigates the influences of multi-criteria route guidance systems on transportation systems and user performance under various market penetrations. The research utilizes a logit-based multi-class stochastic network equilibrium modeling approach, where users can be differentiated by their own value of time (VOT), which follows some probability distribution. Users are assumed to select routes based on the tradeoffs between travel time and (monetary) travel cost. The travel demand of each user class is governed by the discretized VOT distribution. We carry out numerical experiments in an example network through changing various model parameters and input data, including advanced traveler information systems (ATIS) market penetration, origin-destination (O-D) demand, link free-flow cost, and information quality. The findings suggest that the larger the number of user classes, the better the mixed multi-class equilibrium with discrete VOT distribution approximates that of continuous VOT distribution. In addition, the model under multi-class multi-criteria route guidance can provide more insights about the impacts of ATIS on systems and user performance than the conditional single-class single-criterion model.