Travel demand and traffic forecasting models, and specifically the four-step process (trip generation, trip distribution, mode choice, and route choice), have been used for more than 50 years. While they have given reasonable forecasts given the limits of the profession’s understanding of travel behavior and the limits of computational tools, in the last two decades the weaknesses of this modeling approach have become obvious. Currently, the profession is shifting from the traditional four-step process to models that start at the individual traveler level and look at travel generation as an outgrowth of activity involvement (shopping, recreation, and work). Also critical in such an approach is the concept of tours, which are trips that sequentially link multiple activities (from home to exercise class, to shopping, and back home). While more complex models of traveler behavior have been appearing in the academic literature for decades, models more sophisticated than the simple four-step process have only recently begun to appear in practice. To be sure, the transition from traditional four-step models to tour-based and activity-based models presents many challenges. Included among these challenges are significant increases in required
data (to be able to predict individual activity patterns) and limitations of current computer technology. However, more detailed and accurate models for travel and traffic forecasting will allow analysts to determine the impacts of many new transportation policies relating to parking controls, toll roads, congestion pricing, vehicle occupancy restrictions, reductions in energy consumption and emissions, and other emerging transportation policies.