Evaluating No-Notice Evacuation Strategies for an Urban Area

Abstract

Although the occurrence of no-notice events is not as large as short-notice events, they do occur. The prime example of this in United States history occurred on September 11, 2001 when New York, Pennsylvania, and the Pentagon were under attack by Al-Qaeda terrorists. Nearly 2,800 fatalities occurred by the end of that day (1). In response to the no-notice events, countermeasures are needed (i.e., evacuations). In short-notice evacuations (e.g., hurricane evacuations) residents of an impacted area have time to prepare for the evacuation. That amount of time given to evacuees does not exist under no-notice evacuations. This thesis investigates the performance of no-notice evacuation strategies for Northern Virginia. No-notice evacuations can be difficult to execute especially for densely populated urban areas that have a mixed group of road types, such those that make up Northern Virginia. The objective of this thesis is to determine whether implementing a set of signal timing plans optimized for evacuation demand and a group of congestion warning Variable Message Signs are suitable strategies for no-notice evacuations. The simulation-based Dynamic Traffic Assignment model, known as DynusT, was used to execute the evacuation simulations. The results show that improvements to travel time were produced due to the evacuation strategies but they were quite marginal and insignificant for most of the evacuation scenarios.