Prototype Design for a Predictive Model to Improve Evacuation Operations

Texas Department of Transportation, 02/17/2009-01/31/2011

Investigators: Gino Lim, May Feng, and Hamid Parsaei

Objective:

Mass evacuations of the Texas Gulf Coast remain a difficult challenge. These events are massive in scale, highly complex, and entail an intricate, ever-changing conglomeration of technical and jurisdictional issues. This project focused primarily on the specific issue of developing a new technical tool to help TxDOT and other key operating agencies/stakeholders better predict when major elements of evacuation operations should be implemented. In particular, a variety of technical analyses were employed to develop a new, prototype decision support system that provides additional insights to more effectively decide when evaculane shoulder operations versus full contraflow operations are needed to manage evacuation demand. This new tool has a predictive mechanism designed to provide lead time for implementing these two prospective operational scenarios. The work conducted during this research involved a large-scale application of the DynusT model, and integrates several different factors into the evacation operation decision-making process - namely real-time traffic conditions, hurricane characteristics (strength and size) and human behavior.