| This dissertation is concerned with the development of methodologies for the control and optimization of transportation systems under real-time, dynamic, and uncertain conditions. Due to the uncertain nature of transportation demand and supply, these systems can be viewed as multi-stage stochastic optimization problems, where the strategic and tactical decisions need to account for the uncertainty of future operational and real-time network states. For example, decisions might include infrastructure improvement, control systems location and strategy, real-time control optimization, and driver routing through information dissemination. Current practices typically do not sufficiently account for the complex behavior of these stochastic dynamic conditions within the system and the role these conditions play both in online management, and strategic decision making of earlier stages. Often, this is due to the computationally prohibitive costs associated with examining the system under either one of these conditions: dynamic or uncertain. Therefore, new approaches for dynamic network design, user optimal dynamic traffic assignment, and the online shortest path problem are developed to facilitate the exploration in the impact of these conditions.; This thesis employs traditional and new approaches to analyze properties of transportation systems under uncertainty and their effect on the various stages to transportation management; i.e., strategic, tactical, operational and real-time. Numerous pertinent problems are presented or extended to account for dynamics and uncertainty such as the network design problem, dynamic traffic assignment, centralized guidance, and the online shortest path problem. Analytical formulations are developed to describe system properties, and solution algorithms are presented to model transport systems under the stated conditions and, ultimately, to improve the performance of the transportation network. |
