From Academia to Application

Abstract

This paper presents an overview of a new architecture for travel demand forecasting models. The new hybrid accessibility-based models are neither traditional trip-based models nor standard activity-based simulation models. Instead, they are a blend of aggregate and disaggregate component models offering much of the sensitivity typical of activity-based models but with more modest development and application costs. This paper presents results from the first full application of the new model design. It documents the experience of the Knoxville (Tennessee) Regional Transportation Planning Organization from the data collection and visioning stage to the stage of investing time and resources to develop the new model. Ultimately, some basic comparisons are made between the new hybrid model and the previous trip-based model. Some key innovations in the new model design are highlighted, making special note of added sensitivity to planning and policy variables. The paper also reports findings related to the effects of residence location on travelers' willingness to travel and the effect of psychological barriers such as river crossings and county lines on stop location choice.