Research On Emergency Vehicle Path Planning And Real-time Dispatching Of Urban Road Network Under Multiple Accidents

A severe threat to the public’s personal safety and property security exists as a result of the rapid social and economic development,which has led to an increase in the number of people congregating in inner cities.In addition,various emergencies occasionally occur.Emergency response is a complex system project,and after an accident,rapid rescue can reduce the loss of the accident,of which emergency vehicle path planning and dispatching is the core part of the rescue work carried out after the disaster.In this thesis,we first look at the emergency vehicle path planning problem and then,using it as a foundation,we look at the emergency vehicle dispatching problem.In light of the dynamic features of the urban road network and the dynamic events in the dispatching process,we build a real-time dispatching model for emergency vehicles and design appropriate algorithms to solve it.This thesis conducts research on the path planning and real-time dispatching of emergency vehicles on urban road networks under multiple incidents,with the following main aspects:(1)On the taxi Origin-Destination(OD)data,map matching and spatiotemporal feature analysis are done.The reasonable path set of each OD pair is screened,and the Path Size Logit(PSL)model is introduced to find the probability of taxi drivers choosing different paths using the trip distance and travel time in the reasonable path set to represent the path cost,and the trip expectation time is obtained by calculating the travel time of each possible path multiplied by the probability of that path being chosen.The path traveled by the taxi driver is used as an implicit variable to transform the generated link travel time into the error squared between the minimized trip expectation travel time and the trip observation time,and the time-varying link travel time is obtained using the Levenberg-Marquardt(LM)algorithm.In addition,Considering the dynamic characteristics of urban road networks,a link travel time calculation method that combines the real-time speed of the link with the time-varying speed of the link is proposed,where the time-varying speed is converted from the time-varying link travel time.Due to the priority traffic characteristics of emergency vehicles,the speed of social vehicles and the speed of emergency vehicles obey a linear relationship within a certain range,so as to calculate the link travel time of emergency vehicles.(2)Considering the impact of the estimation error of the link travel time of emergency vehicles on the vehicle path planning,the reliability of the path travel time is introduced to minimize the influence of the estimation error on the vehicle path travel time.An emergency vehicle path planning model with the shortest path travel time and the maximum reliability of the path travel time as the optimization objective is constructed,and an Improved Shuffled Frog Leaping Algorithm-K Shortest Paths(ISFLA-KSP)algorithm is designed to solve the model.The algorithm designs a random coding scheme that meets the path connectivity,and based on the property that the subpath of each shortest time path in the First In First Out(FIFO)network is also the shortest time path,an internal optimization strategy is designed.(3)An emergency vehicle pre-optimization dispatching model is established with the severity of the accident and the accident time window as the key factors and the shortest total response time as the optimization goal.For the acquisition of link real-time speed,a periodic optimization strategy is used to adjust the dispatching scheme.Aiming at the change of the number of incident points in the rescue process,the continuous optimization strategy is used to adjust the dispatching scheme,and the real-time dispatching model of emergency vehicles is established.The improved shuffled frog leaping algorithm is designed to solve the above model.Because shuffled frog leaping algorithm is easy to fall into local extremum,and the frog position update does not take its own update experience into account,the inertia weight and cauchy distribution operator are introduced into the local search operation of the shuffled frog leaping algorithm.In the multi-core environment,the parallel computing toolbox of Matlab is used for parallel optimization,and a parallel shuffled frog leaping algorithm based on inertia weight-cauchy distribution is proposed.(4)A part of the road network in Yubei District,Chongqing,is selected for case studies on the estimation of the link travel time of emergency vehicle,the path planning of emergency vehicles and the real-time dispatching of emergency vehicles to verify the applicability of the model and algorithms in this thesis,which can ensure the rapid response to emergencies to a certain extent.