Traffic operation and control during evacuation conditions: An integrated management system with multi-level coordination.

摘要

During the past decade, large storms, mudflows, hurricanes, floods, and many other extreme conditions have caused massive economic and human losses in cities and regions around the world. In the United States, Hurricane Katrina and the September 11th, Terrorists Attacks are relevant examples of such conditions where explicit cost-benefit analyses showed that by presenting a systematic and efficient evacuation scheme in response to such hazards, people lives can be preserved at a much lower cost.;Evacuation study is a complex subject that mainly includes (1) a demand side, (2) a network supply side, (3) a control side and (4) a behavioral side. Focusing on the control side of the problem, the objective of this thesis is to offer a dynamic integrated control strategy that can respond in real-time to any change in demand and supply during extreme conditions. Given the complexity of the problem, a simulation based model is required to get a better understanding and analysis of the process at hand. Through the use of the dominant path concept and transforming the evacuation problem into multiple corridor-based evacuation problems, the corridors where lanes can be reversed and where "intelligent" control tools can be deployed are better assessed. Evaluating evacuation in such framework allows the integration of real-time control strategies into the problem; the demand aspect (specification of the safe versus unsafe zones) and the complexities related to the dynamic assignment of evacuees to different routes are better captured: the model can integrate multiple data sources and network-level traffic routing while allowing improved coordinated control strategies (optimal signal timing, use of variable message signs and ramp-metering).;The logic adopted to realize the objective above is a reactive integrated control. A control "optimal" module mainly includes VMS, path-based coordinated signals and ramp metering control strategies; this module is integrated into a dynamic simulation-assignment framework. A base case simulation using the "every-day" origin-destination demand pattern allows determining the network-wide performance measures including the experienced delays and travel-times. Once determined and with the knowledge of the regions to be evacuated, the highest number of impacted vehicles (traveling to, from or through the impacted areas) and the corresponding dominant paths are identified. Accordingly, coordinating between the controls strategies along these paths (corridors) allows a faster and smoother evacuation. After identifying the components of the above control "optimization" logic, the formulated method is tested on a portion of the Maryland CHART network, USA. The portion considered consists of the I-95 corridor network between Washington, D.C. and Baltimore. The impacted area to be evacuated is inside the capital beltway (i.e. Washington DC) and the safe area to be reached is along the path towards Baltimore. The network is bounded by I-695 in the north, I-495 in the south, US 29 to the west and I-295 to the east. The network includes four main freeways (I-95, I-295, I-495 and I-695), as well as two main arterials (US29 and Route 1). The Maryland CHART network reduces to 2182 nodes, 3387 links and 111 zones.