Study On Simulation Method Of Traffic Evacuation For Emergency Decision Support

Evacuation is a critical task in emergency management,while traffic evacuation plays a leading role in large-scale evacuation.The evucation process in macroscopic view is with complexity and uncertainty,which is mainly caused by the heterogeneity of microscopic driving behaviors.This dissertation searches into the simulation method of large-scale traffic evacuation for emergency decision support,focusing on the heterogeneous behavior of evaucees and its impact on the efficiency of the entire evacuation process.The study aims at developing a traffic evacuation simulation method as well as a software,enabling the simulation system to reflect uncertain and diverse evacuation scenarios,and improve the predictive capability of complex evacuation environment.A variety of traffic simulation models are integrated to obtain a method for heterogeneous behaviors of evacuees.A large-scale traffic evacuation system is designed based on the concept.Several simulation experiments are conducted to demonstrate the capability of the system on decision support.The dissertation presented the development and implementation of a simulation method of traffic evacuation by designing an agent-based model of vehicles,roadmap and decision-makings.Considering the existing traffic simulation models and software system of traffic evcaution,a variety of microscopic evacuation models were reviewed and analyzed,and a microscopic simulation method was proposed to describe the heterogeneity of evacuation behaviors by using the concept of “dynamic parameters + integrated models”.A corresponding numerical model was also developed for calculating this method,and the computational error was discussed.The dissertation thus solved the problem of dynamicly controlling the computational error of the simulation method.Based on these theoretical research,the dissertation designed a simulation system of traffic evacuation for emergency decision support,and proposed a multi-level driving decision process named Strategic,Cognitive,Tactical,and Operational(SCTO)decision process,to integrate a variety of evacuation models and improve the capablility of describing diverse evacuation scenarios.The simulation system,focusing on the overall decision-making process in evcuations,was able to support different kinds of deicisions,from evacuation zoning,demang modeling,additional trip considerations(including background traffic,intermediate trips and shadow evacuaiton),to traffic assignment,route selection,and evacuation strategies.The dissertation compared the proposed simulation system with MATSim,a widelyused simulation system for large-sacle evacuation,to validate its accuracy and effectiveness.The results showed the capability of the proposed system on generating uncertain and diverse evacuation seceraios.An experiment on a real road was conducted to collect the heterogeneity of driving behaviors in real situations,and calibrate the corresponding parameters for setting up a more reasonable simulation environment.As an example,the dissertation conduct a simulation experiment under the case of largescale evacuation,and provided a series of decisions for evaluation as well as suggestions,indicating that the proposed method and simulation has important theoretical value and application prospect in decision support.Achievements in this dissertation provide not only theoretical guidance on the development of traffic evacuation simulation models and systems,but also scientific methods and technical support on the practice of emergency decision-making.