Large Scale Evacuation Transportation Systems: Robust Models and Real Time Operations

This report documents the development of several transportation optimization models under uncertainty. First, a robust evacuation transportation planning model for freeway traffic systems was developed, in which demand is uncertain. The study results provide preliminary evidence that routing traffic using such a robust, optimization-based model can improve the average objective performance. Next, we extend the robust optimization approach for dynamic traffic flow problems by allowing dynamic adjustments as data uncertainty is realized. The study demonstrates that the method is computationally tractable using an example evacuation network of Cape May, New Jersey. Insights are presented with regard to the potential for use of robust optimization for humanitarian relief supply chains specifically and transportation modeling in general. Finally, a new model for robust dynamic network design was developed to reduce network traffic congestion through capacity expansion policies. The study is the first to apply robust optimization methodology to a network design problem with dynamic flows and uncertain demand.