Abstract
Contemporary activity-travel models use a discrete-continuous structure to model activity type and duration. Analysis of raw-activity-traveldiary data shows frequent rounding of start and end times by respondents to multiples of 5 or 10 min. A continuous representation of activity duration in the model could lead to a significant error in estimation that might not always be random. A Poisson process model that uses discrete time to jointly model activity type and duration is presented. It is proposed that a quantum representation of time would possibly provide a more rigorous behavioral interpretation to the individual’s activity choice process. A theoretical description of the time-quantum hurdle model is provided, followed by an estimation of the model for nonworkers in the 1990 Bay Area Travel Survey data set. Also included are an analysis of computational efficiency of the model and possible extensions.
Get full access to this article
View all access options for this article.
