Research On Ambulance Scheduling Problem Of Rescue Based On Evolutionary Algorithm

In recent years,with a variety of emergencies occur frequently,ambulances are being noticed as the public medical service infrastructures.In addition to daily medical services,ambulances are the basic medical support forces in major disaster incidents.A large number of injured people require medical aids at the same time during a large scale emergence,and the efficient scheduling of ambulances to some extent affect the survival rate of wounded and economic losses as a kind of limited resource.However,the current ambulance scheduling system is still based on the dispatcher’s work experience and has not enough support about computational intelligence,which leads to greater risks.The thesis focuses on the ambulance routing problems during the response phase.In this thesis,victims are distinguished into two types: the slightly injured patients who can be aided directly in the field,and the seriously injured patients who have to be transported to hospitals.According to the type and position of victims,ambulances scheduling problem is modeled as a multivehicle routing problem with mission assignments.The main contributions of this thesis are as follows: 1.Ambulance scheduling problem of disaster rescue based on meta-heuristics algorithm: Because of the limited number of ambulances,not all requests can be rescued immediately in disaster relief.In order to reduce the longest waiting time and increase the survival rates of patients,meta-heuristic is used to solve this problem and the coding and corresponding operators is designed.And the goal is minimizing the latest service completion time.In the simulation case,the convergence of the algorithm was verified by small-scale experiments.2.Ambulance scheduling problem of disaster rescue based on multi-agent genetic algorithm: On the basis of the previous work,multi-agent genetic algorithm(MAGA)is also used to solve this problem,and the MAGA is modified with the consideration of intrinsic properties.The local search method is employed to enhance the capability of searching globally without early convergence.The effectiveness of this algorithm was verified by experiments on instances with different scales.3.Ambulance scheduling problem of disaster rescue based on multi-objective evolutionary algorithm: In most cases,the emergency managers can’t ignore the economic costs in the actual conditions.How to achieve the scheduling goal of “the best possible benefit for most people” in the case of the cost is as small as possible,that is,how to achieve the goal of a better chance to survive as the same time minimizing economic budgets in ambulance scheduling,and this is a multi-objective optimization issue.Modeling the problem between survival probability and economic cost,and the multi-objective evolutionary algorithm: NSGA-II algorithm and MOEA/D are used to solve this problem.At the same time,the effectiveness of the algorithm was verified by experiments on instances.