Study On The Optimization Of Evacuation Routing Measures In View Of Traffic Management And Control

With the frequent occurrence of various disasters, emergency management has caught great attention from various countries in the world. China is confronted with even more arduous task of emergency management, because it is densely populated and disasters can cause huge lost of lives and properties. Evacuation, serving as a critical measure to protect people and to reduce casualties, is an important task of emergency management. The regional evacuation process indeed involves movement of people by vehicles, which may generate large plenty of traffic trips during limited time. To guarantee safety and efficiency, it is important to establish routing measures in evacuation planning, including evacuation routes and departure schedules etc. Feasible routing plans or measures should be established in view of the conditions of evacuation network, especially the Traffic Management and Control (TMC) measures existing in the network.Previous approaches mainly use network flow models to optimize evacuation routing measures. Such models, which give straightforward description of the evacuation problem and are convenient to solve, can reflect the existence and influence of TMC measures on links such as one-way road and do-not-enter road, but can not reflect the influence of TMC measures at intersections and hold some unrealistic assumptions about the evacuation problem. For example, static flow models used for route optimization can not reflect the existence of delay and capacity of intersections; dynamic flow models used for route and departure time optimization can not reflect queues caused by TMC measures at intersections and give too simplified description of flow propagation.This dissertation gives an extensive study on the optimization of evacuation routing plans or measures in view of TMC measures, aiming at making up the deficiency of previous methods. The optimization of TMC measures during evacuation is also studied based on the optimization of evacuation routing measures.The details are listed below:(1) Optimization of evacuation route in view of TMCIn most previous models for route optimization, delay and capacity of intersections are ignored. Some approaches consider such factors, but they managed to do it through expanding network, which makes the problem much difficult to solve in large scale network. In the dissertation, the problem of route optimizing was described as a minimum cost flow problem in a special network with directional weights at nodes representing capacity and delay of intersection movements, evacuation routes and flows assigned to each route can be got by searching the minimum cost flow in such network. A modified successive shortest path algorithm that can deal with directional weights at nodes has been presented. A numerical example has been given, and route plans got when considering/ignoring intersection movements were compared. The results showed that congestion was quite likely to happen when executing the route plan that ignores movements at intersections.(2) Optimization of evacuation route and departure schedule in view of TMCTo model the problem of optimizing evacuation route and departure schedule simultaneously, the key is how to describe traffic flow propagation in view of TMC measures reasonably without losing analytical properties of models. Dynamic flow models used in previous approaches have good analytical properties, but give too simplified description of flow propagation and can not reflect the existence of TMC measures at intersections. A simulation-based model has been developed for evacuation route and departure schedule optimization, in which the critical constraints that describe traffic flow propagation were addressed through a simulation loading module. A simulation-based heuristic algorithm has been proposed according to the necessary condition for the optimal solution of the model. In light of the need of the optimizing process, the implementation of the loading simulation module has been given, including a flow propagation model that can describe queues before intersections as well as flow propagation along links, and the algorithm to implement the loading simulation based on such model. A numerical example has been given, and the evacuation plans (route and departure schedule) got when considering/ignoring the influence of TMC measures at intersections were compared. The results showed that ignoring the influence of TMC measures at intersections can lead to illogical plans that may lengthen evacuation time.(3) Simultaneous optimization of TMC measures and evacuation routing plans based on bi-level programmingAccording to the bi-level structure of the process of establishing TMC measures for evacuation, a bi-level programming model has been developed for optimizing traffic control parameters (green ratio) and evacuation routing plans(route and departure schedule) simultaneously. A genetic-based algorithm has been presented. A numerical example has been given.