Existing common route choice models are based on random utility theory, which follows the maximum utility assumption. Recent intelligent transportation system applications have highlighted the need for better models of the behavioral processes involved in route choice decisions. Therefore, prediction of travelers' responses to uncertainty was analyzed. Route choice experiments were conducted to evaluate the effect of the feedback mechanism on decision making under uncertainty. The experimental results were compared with those from a model based on cumulative prospect theory and models based on learning approaches. It is shown that a traveler's sensitivity to travel time differences is lower when variances in travel times are higher. This better understanding of route choice behavior predicted by learning models may improve traffic predictions, as well as the design of traffic control mechanisms.
Samenvatting