A COMBINED TRAVELER BEHAVIOR AND SYSTEM PERFORMANCE-MODEL WITH ADVANCED TRAVELER INFORMATION-SYSTEMS

The goal of this paper is to develop a framework for evaluating the ef fect of Advanced Traveler Information Systems. The framework uses a co mposite traffic assignment model which combines a probabilistic travel er behavior model of route diversion and a queuing model to evaluate A dvanced Traveler Information Systems impacts under incident conditions . The composite assignment model considers three types of travelers: t hose who are unequipped with electronic devices, i.e. they do not have Advanced Traveler Information Systems or radio in their vehicles; tho se who receive delay information from radio only; and those who access Advanced Traveler Information Systems only. The unequipped travelers are able to observe incident-induced congestion, if the congestion rea ches or exceeds their decision point. The composite model assigns trav elers with Advanced Traveler Information Systems to the shortest trave l time route. Travelers with radio information and those who can obser ve the congestion are assigned according to a behavioral model calibra ted on revealed preference data. Travelers who are completely unaware of the incident-induced congestion are assigned to their usual route. The unique feature of the composite model is the integration of realis tic traveler behavior with system performance while accounting for the effect of real-time travel information. To demonstrate the applicatio n of the composite model, we consider the evolution of queues on a two link network with an incident bottleneck. The findings indicate that the overall system performance, measured by average travel time, impro ves marginally with increased market penetration of Advanced Traveler Information Systems. However, the benefits of Advanced Traveler Inform ation Systems under incident conditions are expected to be marginal wh en there is more 'information' available to travelers through their ow n observation or radio. Specifically, delay information received throu gh radio and from observation of incident-induced congestion induces p eople to divert earlier causing the network to operate closer to syste m optimal than user equilibrium. This limits the potential benefits of Advanced Traveler Information Systems. (C) 1998 Elsevier Science Ltd. All rights reserved.