Research On Emergency Rescue Vehicle Dispatching And Route Optimization Under Airport Demand Response Connection Service

In the past ten years,benefiting from the vigorous development of China’s economy,the consumption attribute of aviation demand has gradually increased.The airport connection as a derivative issue has gradually become a research hotspot.The key to punctual,efficient,convenient,and safe airport connections is to handle the connection between the dispatch system and data center which integrates passenger information and vehicle information.Based on the study and combing of the existing literature,this thesis extends the research scenario from general demand response airport connection to emergency rescue airport connection in the case of disturbance.The research content is as follows:Clarify the connection between the general demand response airport connection and the emergency rescue airport connection in the case of disturbance.Emergency rescue refers to the damage to the connecting vehicle that cannot be recovered in a short time,and the implementation of the rescue of the passengers that have been carried by the disturbed vehicle and the passengers that have not been carried.It is the key that querying the current state of the vehicle for the emergency rescue vehicle scheduling,including passenger information and road network information.The demand-response vehicle dispatching plan is not only the vehicle pre-scheduling plan,but also the query basis for the vehicle status when an emergency occurs.Four emergency rescue strategies are given and two operating states of the disturbed vehicle are refined.The path optimization is realized by solving the multi-source shortest path in the road network,and the Floyd-Warshall algorithm is adopted to reasonably solve the problem of path impedance input between demand points.The pre-scheduling model established takes minimizing the vehicle running time as the objective function,and considers four sets of constraints including general vehicle path constraints,capacity constraints,time window constraints,and detour constraints.The model inputs include locations related to connection points,the number of people,the expected time window of arrival at the airport,and the vehicle capacity.The connection and difference between the emergency rescue model and the pre-scheduling model mainly lie in the update and division of the set,as well as the restriction conditions for the refinement of different sets.The objective function of the emergency rescue model is set to minimize the sum of vehicle running time and rescue waiting time.Through specific parameters,the pre-scheduling plan is used as the input emergency rescue model.Additional constraints on waiting for rescue and departure time on virtual parking lots have been set up to ensure the effectiveness of emergency rescue.The column generation algorithm is used to decompose the original model into a main problem and multiple sub-problems.The main problem is the generalized assignment of unconnected passengers,and the sub-problems are vehicle routing problems that depend on vehicles.In the process of continuously passing "columns" between the main and sub-problems,the dynamic addition of solutions in the model is realized.A case with the Sunan International Shuofang Airport in Wuxi City as the background airport,with 81 road nodes,20 demand points,and 272 road sections as a large-scale case road network,with random disturbance moments and disturbed vehicles as emergencies has been taken.The AMPL environment is used to call Gurobi to solve the pre-scheduling model,and the pre-scheduling vehicle scheduling plan is obtained within one minute,including the number of vehicles,connection sequence,and connection time.In the same environment,the column generation algorithm is used to solve the emergency rescue model and obtain the emergency rescue scheduling plan,which not only provides a high-quality solution for the scheduling system,but also improves the efficiency of the scheduling.In addition,it is proved that the column generation algorithm for emergency response model construction,including a generalized assignment main problem and multiple vehicle routing sub-problems,can effectively be used to solve the emergency rescue model.