Dynamic network models and driver information systems

Abstract

Dynamic network models are needed to analyze traffic congestion patterns for new real-time motorist information systems. In previous research, a dynamic network modeling framework incorporating behavioral models of drivers' route and departure time choices and their day-to-day adjustment processes was developed. Network performance in this framework is represented by time dependent arrival and departure rates, link occupancies, and queuing delays. The purpose of this paper is to extend this framework to include explicit models of drivers' information acquisition and integration. The need for these models is motivated by considering the possible beneficial and counter-productive effects that may be caused by enhanced motorist information. Information on network conditions influences the set of routes considered by a driver and also affects the perceived values of the level of service attributes. The paper presents the structure of a dynamic model in which newly acquired information may affect pretrip and en-route travel decisions. To assess the potential magnitudes of the effects that were identified further theoretical and empirical research is needed.