Traffic-Assignment Techniques for Smaller Cities

Abstract

The primary purpose of the research was to determine which of the heuristic assignment techniques currently available to transportation planners in the United States provides the most accurate results for small and medium-sized cities and to assess the potential for future applications of equilibrium assignment techniques. A number of all-or-nothing, multipath, and capacity-restrained assignment techniques were applied to highway networks for Fond du Lac and Madison, Wisc., using common trip tables. Using the percentage of root mean squared error as the primary accuracy measure/percentage of the accuracy of the assignments in the order of increasing accuracy was all-or-nothing, multipath, and capacity-restrained. The accuracy of the capacity-restrained assignments appeared to be more sensitive to the assumptions made in computing the peak-hour assigned volumes and capacities than to differences in the capacity-restraint techniques. An equilibrium assignment for Winnipeg, Manitoba, Canada, was slightly less accurate than the most accurate assignment for Madison. A technique for assessing the behavioral validity of equilibrium assignments is proposed.