Simulating Household Travel Survey Data in Metropolitan Areas.

Abstract

This dissertation proposes and tests a new approach for developing the travel survey data required for use in regional travel-demand modeling applications. Using local sociodemographic information in conjunction with a national travel survey, a simulation procedure is developed to create a synthetic household travel survey data set. The simulation procedure is tested in a region that has completed a recent travel survey---this provides a direct source of comparison for the merit of the approach. Comparisons of salient trip characteristics (trip frequencies, mode shares, departure times, and reported trip lengths) and travel-demand models estimated with these synthetic data suggest the approach has substantial potential. However, discrepancies do remain which are attributed to contextual differences between regions. Procedures are proposed and tested to capture these unexplained differences, creating a travel survey data set that is more sensitive to local conditions. It is anticipated that this approach will enable a region to develop their own travel data set and estimate travel-demand models at a fraction of the cost of conducting a traditional household travel survey.