Day-to-Day Departure Time Adjustment Behaviors in a Single Bottleneck Model

Dynamic traffic equilibrium implies that each commuter can correctly utilize complete, perfect and real-time information on travel conditions released by advanced traveler information systems (ATIS). In reality, traffic information provided by ATIS is rarely perfect, and drivers' capability of processing information is restricted within certain limits. This paper investigated commuters' day-to-day adjustment to departure time through comparing the differences of the private total costs between time intervals, based on their own experience and aforetime information, and explored their learning behaviors on the traffic dynamics of the bottleneck model. Finally, a bottleneck network example showed that how commuters' adjustment behavior affects the evolution of traffic flow over days. Commuters' departing early or later leads to two morning peaks in the departure rates. The departure rates under different ω-values are distinguishable from, but approximate to the departure rate at equilibrium state.