Household Travel Data Simulation

This paper presents the process of developing models that can facilitate disaggregate household travel data transferability. Household records from the 2001 National Household Travel Survey (NHTS) are clustered into several homogeneous groups representing various household lifestyles. Using an artificial neural network model, households from add-on areas of the NHTS were assigned to the same cluster schema developed for the national dataset. Travel estimates from national data are transferred to the add-on areas based on household cluster membership. Using a small local sample, transferred travel data are updates considering the observed distributions and utilizing Bayesian updating. Furthermore, the household level travel data transferability model is combined with a population synthesizing model to create synthetic household travel data that can reduce or eliminate the need for a large data collection in the application context.