Research on airport trailer emergency scheduling model based on genetic simulation annealing algorithm

With the increase of air transport business year by year, the airport comprehensive service capabilities are increasingly facing severe challenges. As a key part of airport service, ground scheduling is crucial to the departure of flights on time. by analyzing the traditional rules of airport trailer and combining with the service processes and the characteristics of trailers, the paper develops a multi-target trailer emergency scheduling model with rolling windows. When the model is solved, the simulation annealing algorithm is introduced into the genetic algorithm to prevent the optimal result falling into the local optimal solution. The simulation results show that the model can more effectively solves the problem of airport trailer emergency scheduling in the case of flight delays than the traditional scheme,and to relieve the pressure of flight delay for busy airports.


Introduction
With the increase of air transport business year by year, the airport comprehensive service capabilities are increasingly facing severe challenges. As a key part of airport service, ground scheduling is crucial to the departure of flights on time.
At present, most of the airport flight scheduling service is still in the manual scheduling phase [1], adopting the First in first out scheduling strategy.Although the operation is simple,efficiency is low. It is obvious no value for solving the flight delay when the weather suddenly changes,flight delays, Special vehicle breakdown and other emergencies,Especially.the ground scheduling of trailers during the flight access to station operation is studied by Jiayan Du [2][3],developing a VRPTW model according to the operation flow of the towing vehicle and Minimizing the total scheduling time cost of the entire process .simultaneously,Xia Feng [4]has established the coordinated scheduling model of refuel tanker and shuttle bus with taking the refueling service and pick-up service time of flights as the constraint,and aimed at the objective of minimizing the number of safeguards and making the total starting time of service earliest, and given model solution based on multi-objective genetic algorithm.
The paper develops a trailer emergency scheduling model, by analyzing the traditional rules of the trailer scheduling based on the existent researches and related papers,and combining with the service flow and time characteristics of the trailer,then gives the optimized scheduling scheme.

Traditional trailer scheduling strategy
Now, most airports adopt the First in first out strategy in the airport trailer scheduling [5].The strategy,which is to select a grouping scheme from the scheduling system,and the grouping which can be completed in the shortest time i t in the GPS system corresponding to the scheduling system is selected for transmission. Among all the grouping schemes waiting to be scheduled (see figure 1).It is 2 1234567890''"" confirmed that the time which takes for the packet completion service lags behind the time it takes to transmit the longest packet on the process ,comparing with the corresponding GPS system [6]. When a trailer group arrives,it will be sorted into the corresponding queue by classification.the scheduling system will calculate the completion time of the process,at the same time [7].Then select the packet with the shortest completion time among the packets as the next packet transmitted from the output port, and the packet period is T.
Among them, the scheduling time is calculated as follows: In the above equation ( is used as the constraint of scheduling algorithm, and the node trailer scheduling delay as follows equation(2) [8]: Among them, C is the service rate of the server,and max L is the maximum of the packet length of vehicles.

Trailer emergency scheduling model
The paper develops a trailer emergency scheduling model,and gives the optimized scheduling scheme.

putting forward of the problem
According to the conventional ground specific vehicle scheduling strategy, the nearest trailer is assigned at the current time, but the actual distance between the current failed trailer and the next trailer in the group is ignored. And the trailer's work flow is chained with strict scheduling rules and time constraints. A change of a link or a group of trailers may cause subsequent trailers to fail to schedule normally and affect the normal launch of subsequent flights. In order to efficiently schedule flights under the condition, it is necessary to optimize the scheduling strategy and improve the level of ground scheduling and security.
So, the paper develops a multi-target trailer emergency scheduling model with rolling windows.and gives a certain number of delayed flights and serviceable trailers, we construct a coordinated emergency scheduling model for all kinds of trailers with flight delays and assign them to service

Model description and symbolic description
In this paper, the rolling window strategy is used to divide the flight delay time into multiple equal time periods. After that, the flights in each time period are optimized according to their planned take time-off time and delay time.In this way, not only can the scheduling plans of delayed flights be developed quickly, but also the real-time processing of flight information can be ensured.
In the actual operation of the airport scheduling, the trailers are divided into three categories called small, medium and large.The specific information as the table1 show.
For each scheduling path k, kj s represents the time when the trailer j starts to schedule a delayed flight in the flight schedule path k, that is the trailer j arrives at the scheduled downtime. If there is a trailer j in the scheduling path k, then

Model assumptions
In the following study in this paper, for the delay of departure flights, the establishment is based on the following assumptions: • Each flight happens to be towed once.
• The flight has been completed refueling, boarding and other work, the trailer towed to the taxiway can be released takeoff. • During the process of receiving the towing service, the towing vehicle will not malfunction and other events that affect scheduling will occur. • Once the trailer has started its work, it will not be interrupted until the end of the service.

trailer emergency scheduling model constraints
Trailer emergency scheduling model constraints is extremely limited and in order to operate safely, all kinds of special vehicles must travel at a fixed path and speed.in the flight scheduling process, the use of the trailer must be in accordance with the "flight-trailer" corresponding mode to draw the flight.
Time constraints in the scheduling process are as follows equations (5)- (7):

Trailer emergency scheduling optimization goals
According to the actual operation of the airport and the emergency scheduling model of the above trailer, the various types of trailers under the condition of flight delay are optimized to achieve the goal of minimizing the total service time of all delayed flights,as follows equation (8).

Model solution
According to the above model, the key to solve the delay flight scheduling path problem is to reasonably determine the relationship between the flight and each trailer, to minimize the total delay time when the corresponding types of flights and trailers and the departure time of flight are satisfied, and to avoid genetic algorithm (GA) is easy to fall into the condition of local optimization and slow convergence of simulation annealing algorithm (SAA), and a model based on simulation annealing genetic algorithm to solve the emergency scheduling of trailers is constructed [9][10][11].

Algorithm steps
The specific idea of simulation annealing genetic algorithm is as follows: Step 1: Initializing the control parameters. The population size is N , the mutation probability is m P , the initial annealing temperature is 0 T , the temperature cooling parameter is .
Step 2: Construct chromosomes to generate initial population.For example, chromosomes {6, 2, 5, 4, 1, 3} indicate that flight number 1 is the 6th towing service, flight number 2 is the second towing service, and so on.
Step  (  ,  1  min{ , b is a random number between 0 and 1, then the current individual is received; Step 5: The cross after the above-mentioned individual mutation operation, combined with the fourth step of the method to determine whether to receive the variant solution.
Step 6: to determine whether to meet the convergence conditions. Otherwise, ,turn to the step 3,and continue to find the optimal solution.

Application examples
Based on the above trailer emergency scheduling model, this section will use the delayed flight data as an example that to verify the reliability of the scheming model,and comparing with the conventional trailer scheduling method. The scheduling time for each flight is assumed to be 10 minutes.Specific flight information as follows table 2.     table 2 and table 4,can see that the above trailer emergency scheduling method which is only a total delay of 15 minutes on the basis of the current flight schedule.

summary
By analyzing the traditional rules of airport trailer and combining with the service processes and the characteristics of trailers,the paper develops a multi-target trailer emergency scheduling model with rolling windows.And the simulated annealing algorithm is introduced into the genetic algorithm to prevent the optimal result falling into the local optimal solution when the model is solved. The simulation results show that the model can more effectively solve the problem of airport trailer emergency scheduling in the case of flight delays.In the future research work,we can refine the relevant factors and go further quantitative analysis to improve the efficiency of airport scheduling.