Sensitivity of Four-Step Versus Activity Based Models to Transportation System Changes

(PI: Chandra Bhat; Co-PIs: David Schmitt, Mark Bradley, John Bowman, Ram Pendyala)

Dr. Bhat is the Lead PI on an Ohio Department of Transportation funded project to compare the practical performance of activity-based and trip-based modeling systems for the Columbus region. This project is being undertaken in collaboration with David Schmitt (AECOM Consult, Inc.), Mark Bradley (Mark Bradley Research & Consulting), John Bowman (Bowman Research and Consulting), and Dr. Ram Pendyala (Arizona State University). The primary objective of this study is to assess and quantify the differences between the activity-based travel demand models and the traditional trip-based models for transportation project evaluation, policy analysis, and travel demand forecasting. To this end, the specific objective is to conduct before-and-after studies and compare the accuracy of predicted changes in travel patterns from the activity-based and trip-based models of MORPC. A second objective of this project is to identify how the two model systems differ in their predictions for actual projects and for hypothetical policies, and determine the components of the model systems that are responsible for major differences in the forecasts. For example, in the analysis, it may be found that key differences between the models are due to the fact that the activity-based model uses a destination choice model instead of a gravity model, or explicitly models time-of-day choices. This detailed-level analysis approach will help identify particular weakness and strengths of alternative model systems, and will provide guidance regarding the most practical and efficient way to address different kinds of policy questions. Importantly, the analysis will shed light on the limitations of trip-based models and the current MORPC activity-based model, identify key areas for improvement of both types of models, and make recommendations regarding the use of alternative model systems and components within the context of policy analysis.