Real-Time Diversion Model for Traffic on Evacuation Corridors

Caused by events such as terrorist attacks and natural disasters, evacuation planning and management has attracted growing interest from researchers, engineers, and governments for protecting people from disasters and complications. Transportation engineers are challenged by moving the extremely large groups of people that are usually associated with an evacuation. Pre-developed plans help evacuation management, but do not guarantee its success. Even carefully laid-out evacuation plans can have a diminished capacity or become ineffective because of a crash or other incident blocking a vital evacuation route. An incident blocking an evacuation corridor results in serious traffic congestion. As queue builds up along the corridor, both travel time and delay significantly increase upstream from the incident location, which leads to extra evacuation time or even a failed evacuation. Diversion, using adjacent arterials to divert stuck traffic, can be a feasible approach to evacuation incident management due to its capability to reduce traffic volume on the freeway and distribute exceeded demand to adjacent arterials. Widely deployed Intelligent Transportation Systems have made real-time traffic data collection much more efficient than before, but it is still difficult to re-route evacuation traffic in real time when precipitating incidents occur. Diversion routing during evacuation is a computationally challenging task because the number of evacuees often far exceeds the capacity, and the transportation networks involved are very large. Based on a review of the literature, two problems regarding diversion routing for traffic on evacuation corridors were raised: 1) lack of a diversion routing model considering resource cost for diversion control and characters of diversion traffic in evacuation; 2) lack of a method for real-time diversion routing in evacuation. For the first problem, a new diversion routing model was developed upon the basic minimum cost network flow model. It considers characters of diversion traffic in evacuation, while the prominent feature of the new model is to take intersection control cost for diversion operation into the route optimization process. The model is more appropriate for identifying practical diversion routes in evacuation than conventional ones. To address the second problem, a method for determining subareas for diversion routing was designed. This method decreases the road network size involved in routing algorithms and significantly reduces computational time, which makes real-time diversion routing feasible during evacuation. By using the two methods together, practicable diversion routes could be obtained in real time to alleviate congestion on evacuation corridors caused by precipitating incidents.