Enhancements of ATMIS Using Artificial Intelligence

Dynamic traffic assignment (DTA) has been a topic of substantial research during thepast decade. While DTA is gradually maturing, many aspects of DTA still needimprovements, especially regarding its formulation and solution capabilities under thetransportation environment impacted by the Advanced Transportation Management andInformation Systems (ATMIS). It is necessary to develop a set of DTA models toacknowledge the fact that the traffic network itself is probabilistic and uncertain, anddifferent classes of travelers respond differently under uncertain environment, givendifferent levels of traffic information. This paper aims to advance the state-of-the-art inDTA modeling in the sense that the proposed model captures the travelers’ decisionmaking among discrete choices in a probabilistic and uncertain environment, in whichboth probabilistic travel times and random perception errors that are specific to individualtravelers, are considered. Travelers’ route choices are assumed to be made with theobjective of minimizing perceived disutilities at each time. These perceived disutilitiesdepend on the distribution of the variable route travel times, the distribution of individualperception errors and the individual traveler’s risk taking nature at each time instant. Weformulate the integrated DTA model through a variational inequality (VI) approach.Subsequently, we discuss the solution algorithm for the formulation. Experimental resultsare also given to verify the correctness of solutions obtained.