Modeling advanced traveler information services: static versus dynamic paradigms

This paper develops a cell-based variational inequality formulation of the dynamic traffic assignment (DTA) problem to evaluate the impact of advanced traveler information systems (ATIS) services. It considers two classes of drivers: those with ATIS and those without. Both classes are modeled to follow the stochastic dynamic user optimal conditions, with the equipped drivers having a lower perception variation of the network travel time due to the availability of better information. The model represents traffic dynamics according to the cell-transmission model, including such physical effects as queue spillback, shockwaves, etc. One objective of this study is to compare and contrast the static versus the dynamic modeling paradigms for this problem. The numerical study indicates that some aspects of the results from these two paradigms could be diametrical. The discrepancy is mainly attributed to the fundamental characteristic of a model--how traffic is represented. It appears that simplifications from the physical queue representation are inadequate in producing correct results, especially for already congested networks wherein junction blockage is common.