Study On Identifying Critical Emergency Facilities On Traffic Network:Models And Algorithms

In recent years,disasters like fire hazards(including explosions,industrial accidents,etc.),earthquakes,geologic hazards occurred frequently in China,and emergency resource deployment optimization issues outstanding.For the conflicts between supply and demand of emergency resource deployment,this thesis provides analysing methods and tools for managing emergency and public security resources in the point of location-routing-network flow.Based on above-mentioned background,this thesis analyses emergency vehicle deployment objective with demand zone service level evaluation,public service system reliability,traffic network congestion,proposes measurements and calculating method for the emergency deployment modes of static location,redeployment,patrolling,and proposed the analytic bilevel model for identifying critical emergency vehicles considering traffic network conditions.This thesis constructs heuristics for different traits of these models.Genetic algorithms and hamming distance strategy based multiagent evolutionary algorithms are employed for single level model and solving frameworks for two-stage or bilevel problems.Based on the spatial decision and data organization for emergency vehicles traits,identifying critical vehicle system on traffic network based on ArcGIS Engine is designed and developed.A case study is proposed and discussed for emergency vehicle planning in Xinjiang Uygur Autonomous Region.The conclusion is drawn as follows.1.This thesis constructes emergency vehicle maximal preparedness covering model,reliability-oriented emergency vehicle location-routing chance constrained programming model under stochastic fixing cost and traffic cost,and critical emergency vehicle identification and traffic assignment bilevel model with user equilibrium and stochastic user equilibrium describing traffic network flow.2.Multiagent evolutionary algorithm is revised based on the models.Two-stage genetic algorithm and iterative algorithm for bilevel programming are constructed for solving the relative models.3.In the developing environment of Visual Studio,the emergency vehicle deployment decision-support system is developed for the urban emergency management center with Visual C#.NET programming language and ArcGIS Engine application programming interface.The emergency vehicle deployment decision-support system provides united emergency management service for the emergency management center such as visualization of data and optimization results,real-time information enquiries,decision support,dispatching,etc.The emergency vehicle deployment decision-support system includes emergency facility planning decision-support system,emergency vehicle dispatching system,risk monitoring and warning system,and emergency evacuation guiding system.The application of the emergency vehicle deployment decision-support system is introduced with the emergency vehicle management problems in Xinjiang Uygur Autonomous Region.