Development of dynamic traffic assignment models for planning applications

Dynamic Traffic Assignment (DTA) models have been studied extensively in the last two decades and have reached a sufficient level of maturity to be potentially applicable for planning applications. However, there are still numerous obstacles in using DTA in actual applications. This dissertation aims to address some of the issues and to develop a framework for allowing DTA to be used on large-scale networks for planning purposes.; Travelers can be grouped into two classes, those with fixed departure times and those with preferred arrival times, according to their trip purposes. This dissertation proposes a simple heuristic procedure that utilizes information contained in travel survey to generate time-dependent demand for each class. A combined model of time-dependent trip distribution and traffic assignment is also presented and discussed, and a solution algorithm is proposed by using Dantzig-Wolfe decomposition algorithm and Lagrangian relaxation technique.; Most of existing DTA models assume demand with fixed departure times and system performance under arrival time based demand has not been studied extensively. A new set of linear programming (LP) formulations is presented to study system behavior with arrival time based demand. Insights into the problem are obtained by constructing the dual of the LP and using complementary slackness theory.; Although solution algorithms are very well established for LP problems, the size of the problem that can be solved using the LP formulation is still limited by the power of mathematical programming solvers. To relax such limitations, the network structure within the LP formulations is explored to develop decomposition schemes. Such decomposition of the LP formulation can potentially allow us to solve DTA analytically on large size networks. In addition, insights into the problem can be obtained to help better understanding of the behavior of the transportation network under dynamic demand.; Development of a simulation-based heuristic DTA model that can be implemented for real world applications is also presented. The model uses a mesoscopic simulator and a time-dependent shortest path algorithm and is tested on an actual urban network with more than 16,000 links.