TRAVEL TIME PREDICTION AND DEPARTURE TIME ADJUSTMENT BEHAVIOR DYNAMICS IN A CONGESTED TRAFFIC SYSTEM

This paper examines two heuristic rules proposed for describing urban commuters's predictions of travel time as well as the adjustments of departure time in response to unacceptable arrivals in time daily commute under limited information. It is based on the notion that the magnitude of the predicted travel time depends on each commuter's own experience, including recallable travel time, schedule delay, and difficulties in searching for a satisfactory departure time. An exploratory analysis is first performed to compare these two rules, based on the information provided by a set of commuters interacting over 24 days through a simulated traffic system. A more elaborate model specification which captures the dynamic interrelation between the commuter's cumulative and recent experience with the traffic system's performance is then proposed. The model parameters are estimates with explicit conideration of the serial correlation arising from repeated decisions by the same individuals and the contemporaneous interaction with other system users' decision through the traffic system's performance.