Contemporary Urban Travel Forecasting Methods

This course provides an overview of the travel forecasting problem. The course begins with defining the objectives of the problem, and reviews the weaknesses of the traditional 4-step procedure. A general formulation of the travel forecasting problem is stated. An integrated mathematical forecasting model is derived, in which the traffic assignment, mode choice and origin destination steps are solved simultaneously. Traditional solution algorithms including Frank Wolfe, Dijkstra and Evans are presented; a more robust algorithm for the traffic assignment problem is introduced. Potential extensions to the model are pursued. Alternate model formulations are also described including activity based and combined land use - transportation models.