Airport ground service dispatching optimization method considering time window and airport condition

Abstract

The application discloses an airport ground service dispatching optimization method considering a time window and an airport condition, and comprises the following steps: establishing an airport ground service mixed-skill dispatching model based on a team-oriented problem with a time window, and describing an airport ground service scheduling problem through the model; establishing a target function of multi-objective optimization of the airport ground service mixed-skill dispatching model, and determining the number of completed ground service tasks and the balance of employee working hours as optimization targets; setting constraint conditions of the airport ground service mixed-skill dispatching model according to the actual condition of the airport, and setting parameters of the airport ground service mixed-skill dispatching model; and based on Python, a program is written, a Gurobi solver is called after a data file is read, and an airport ground service dispatching optimization scheme is obtained by solving. The application is helpful to improve the time utilization rate of each ground service employee, improve the comfort of the employee, and save the cost of the airport. Meanwhile, the model algorithm of the application has good solving characteristics, and has a strong practical application prospect.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of airport ground service dispatching, in particular, to an airport ground service mixed skill dispatching optimization method considering time windows and random airport conditions. BACKGROUND

[0002] Airport ground service is an industry that provides ground services for airlines at airports, including baggage check-in, crew pick-up and drop-off, airport transportation, cleaning services, cargo handling, maintenance, and other services. The quality of airport ground service is crucial to ensuring the normal operation of flights and improving passenger experience. In recent years, with the rapid development of the civil aviation industry and the continuous optimization of airport operation modes, the importance of airport ground service has become increasingly prominent. Airport ground service, as an important part of airport operation, is an important guarantee for the normal operation of flights.

[0003] However, the current flight arrangements of major airlines mainly rely on the human resources of various departments or ground service companies. Traditional dispatching methods cannot observe the overall situation of the entire transportation system and it is difficult to make correct decisions based on actual conditions, resulting in low efficiency or even irrationality of the transportation system. The quality of airport ground service is directly related to the normal operation of flights and the travel experience of passengers, therefore, how to optimize airport ground service has become an important problem for airport managers and ground service providers to continuously explore and research.

[0004] Currently, the algorithms used for individual operation scheduling or overall scheduling of various ground service vehicles mainly include NSGA-II algorithm-based scheduling, multi-agent algorithm-based scheduling for aircraft deicing, and scheduling algorithm based on greedy strategy. However, existing models do not fully consider the changes in flight information during airport operation, and can only provide static reference, and cannot well reflect the dynamic changes in the actual operation process. Therefore, there is an urgent need for a transportation scheduling method based on flight data to adapt to complex and variable real-world environments. SUMMARY

[0005] The purpose of the present application is to overcome the shortcomings and deficiencies of the prior art, and to provide an airport ground service dispatching optimization method considering time windows and airport conditions, which helps to improve the time utilization rate of each employee, improve the comfort of employees, and save airport costs.

[0006] To achieve the above purpose, the technical solution adopted by the present application is as follows:

[0007] An airport ground service dispatching optimization method considering time windows and airport conditions, comprising the following steps:

[0008] S1. Based on the team-oriented problem with time window, an airport ground service mixed skill dispatching model is established, and the airport ground service scheduling problem is described through the model;

[0009] S2. The objective function of the multi-objective optimization of the airport ground service mixed skill dispatching model is established, and the number of completed ground service tasks and the balance of employee working hours are determined as the optimization objectives;

[0010] S3. According to the actual situation of the airport, the constraint conditions of the airport ground service mixed skill dispatching model are set, and the parameters of the airport ground service mixed skill dispatching model are set;

[0011] S4. Based on Python, a program is written to read data files and call the Gurobi solver to obtain the dispatching optimization scheme of the airport ground service.

[0012] Further, the airport ground service scheduling problem is described through the model, specifically:

[0013] E represents the set of employees, e represents the e-th employee, I represents the set of tasks, i represents the i-th task, and j represents the j-th task;

[0014] In the process of ground service, it is necessary to judge whether the employee e has the qualification of auditing task i Whether the employee e has the double-machine guarantee qualification γ e Whether the auditing task i and the auditing task j belong to the same short-time transit flight b ij Whether the auditing task i and the auditing task j meet the double-machine guarantee condition c ij According to the optimization objective, the penalty score ε of the employee e executing a double-machine guarantee task once is set e ;

[0015] In the airport ground service task, the completion of the task and the arrangement of the employee are included, and during the airport ground service dispatching, it is necessary to determine the start time O i of the auditing task i, judge whether the auditing task i is located at the remote machine position ξ i , and determine the duration v i of the auditing task i;

[0016] For employee scheduling, the time τ ij that the employee moves from the machine position of the auditing task i to the machine position of the auditing task j needs to be considered, the set of executable task periods T e of the employee e, the start working time of the employee e at the period t The end working time of the employee e at the period t The duration of the h-th meal in a day Γ h The set of meal times H e of the employee e, the previous available period t corresponding to the h-th meal hthe earliest start time of the hth meal Θ h the latest start time of the hth meal Φ h the total working hours d of the employee e in a day e .

[0017] Further, the airport ground service mixed-skill scheduling model makes decisions on employee task arrangement by combining task completion and scheduling optimization, and considers whether the employee e performs the audit task j after performing the audit task i in the time period t, i.e. whether the employee e performs the audit task i in the time period t whether the double-machine guarantee task ij is successfully performed the hth start meal time of the employee e

[0018] Further, the objective function of the multi-objective optimization of the airport ground service mixed-skill scheduling model is:

[0019]

[0020]

[0021] In the formula, |E| represents the number of employees, formula (1) maximizes the number of completed tasks, i.e., the total number of tasks completed by employees minus the total score of the double-machine guarantee penalty; formula (2) minimizes the difference between the working hours of the employee and the average value of the working hours of all employees, satisfying the working hour balance among employees.

[0022] Further, the constraint conditions of the airport ground service mixed-skill scheduling model are:

[0023]

[0024]

[0025]

[0026]

[0027]

[0028]

[0029]

[0030]

[0031]

[0032]

[0033]

[0034]

[0035]

[0036]

[0037]

[0038]

[0039]

[0040]

[0041]

[0042]

[0043]

[0044]

[0045]

[0046] where k represents the kth task; g represents the gth task; t represents the tth available time period; represents whether employee e has no audit task to perform at time period t after performing audit task i; represents whether employee e has no audit task to perform at time period t after performing audit task i; represents whether employee e has no audit task to perform at time period t after performing audit task i; represents whether employee e has no audit task to perform at time period t after performing audit task i; represents whether employee e has no audit task to perform at time period t after performing audit task i; represents whether employee e has no audit task to perform at time period t after performing audit task i; j represents the start time of audit task j; represents whether employee e has no audit task to perform at time period t after performing audit task i; represents the start time of audit task j; represents whether employee e has no audit task to perform at time period t after performing audit task i; k represents the start time of audit task j; j represents whether employee e has no audit task to perform at time period t after performing audit task i;

[0047] Compared with the prior art, the application can not only solve the scheduling problem of employees, but also well solve the problem that employees eat in the uncomfortable interval and deviate from the comfortable interval too far, help to improve the time utilization rate of each employee, improve the comfort of employees and save the cost of the airport. Meanwhile, the model algorithm of the application has good solving characteristics and has strong practical application prospect. BRIEF DESCRIPTION OF DRAWINGS

[0048] Figure 1 The flow chart of the airport ground service dispatching optimization method considering time window and airport condition of the application.

[0049] Figure 2 The satellite map of Xiamen Gaoqi International Airport in the embodiment.

[0050] Figure 3 The parking position map of Xiamen Gaoqi International Airport in the embodiment.

[0051] Figure 4 The Gantt chart of part of flight duration in the embodiment.

[0052] Figure 5 The working position and sequence of employees in the embodiment.

[0053] Figure 6 The tornado chart of part of optimization projects in the embodiment.

[0054] Figure 7 The working position and sequence transfer of employee 27 in the embodiment.

[0055] Figure 8 The tornado chart of part of optimization projects in the embodiment. DETAILED DESCRIPTION

[0056] The airport ground service dispatching optimization method considering time window and airport condition of the application will be further described below in combination with the drawings and specific embodiments.

[0057] Please refer to Figure 1 The application discloses an airport ground service dispatching optimization method considering time window and airport condition, which comprises the following steps:

[0058] S1. An airport ground service mixed-skill dispatching model is established based on a team-oriented problem with time window, and an airport ground service scheduling problem is described through the model.

[0059] S2. A target function of multi-objective optimization of the airport ground service mixed-skill dispatching model is established, and the number of completed ground service tasks and the balance of employee working hours are determined as optimization targets.

[0060] S3. According to the actual situation of the airport, set the constraint condition of the airport ground service mixed skill dispatching model, and set the parameters of the airport ground service mixed skill dispatching model;

[0061] S4. Based on Python, a program is written to read data files and then call the Gurobi solver to obtain the dispatching optimization scheme of the airport ground service.

[0062] In order to facilitate the description of the mathematical model, Table 1 lists the symbolic variables involved in the present application.

[0063] Table 1 Set, parameter and variable

[0064]

[0065]

[0066] The mixed skill dispatching of airport ground support service is a condensation and induction of all business of the airport in normal times. Due to the diversity of the actual situation, if the whole situation is considered, the complexity of the model is greatly improved, and even loses practical significance. Therefore, in order to simplify the process and details of dispatching, make the model have higher practicality, be applicable to more extensive situations, and be able to obtain the optimal solution, the following assumptions are made on the mixed skill dispatching process and related details of the airport ground support service:

[0067] Vehicle resources are sufficient and consistent: the inventory of various service vehicles in the airport is sufficient to serve all tasks in the model and examples, and the basic indicators such as mass and load of each vehicle are consistent, and there is no special quality vehicle to interfere with the completion of the task.

[0068] Vehicle route is determined: the vehicle commuting route between two task points is determined, and the commuting time is fixed.

[0069] Time difference belonging to "the same short time": for variable b ij When judging whether the audit task i and the audit task j belong to the same short time overflight, since the constraint condition states that "if two tasks belong to the same short time overflight, they must be completed by the same employee", in order to ensure the effectiveness of adding such a judgment, flights with no deviation value (completely consistent time) are all identified as "the same short time overflight".

[0070] No commuting time is needed within the same area: there is mutual commuting time between each apron area, such as 10 minutes of commuting time for employees in area 1 and area 2. It is assumed that no commuting time is needed within the same area (such as area 3 to area 3), that is, the commuting time is 0.

[0071] The present application describes the airport ground service scheduling problem through the model, specifically:

[0072] E represents a set of employees, e represents the e-th employee, I represents a set of tasks, i represents the i-th task, and j represents the j-th task.

[0073] During the ground service process, it is necessary to determine whether the employee e has the qualification to audit the task i Whether the employee e has the dual-machine guarantee qualification γ e Whether the audit task i and the audit task j belong to the same short-time transit flight b ij Whether the audit task i and the audit task j meet the condition c of dual-machine guarantee ij According to the optimization objective, set the penalty score ε of the employee e performing a dual-machine guarantee task once e .

[0074] In the airport ground service task, the completion of the task and the arrangement of the employee are included. During the airport ground service task arrangement, it is necessary to determine the start time O of the audit task i i Determine whether the audit task i is located at a remote machine position ξ i Determine the duration v of the audit task i i .

[0075] For employee scheduling, the time τ of the employee moving from the machine position of the audit task i to the machine position of the audit task j needs to be considered ij The set of task execution time periods T of the employee e e The start working time of the employee e at time period t The end working time of the employee e at time period t The duration Γ of the h-th meal in a day h The set of meal times H of the employee e e The previous available time period t corresponding to the h-th meal h The earliest start time Θ of the h-th meal h The latest start time Φ of the h-th meal h The total working hours d of the employee e in a day e .

[0076] The airport ground service mixed skill task arrangement model makes decisions on employee task arrangement by combining task completion and scheduling optimization, considering whether the employee e executes the audit task j after executing the audit task i at time period t, that is Whether the employee e executes the audit task i at time period t Whether the dual-machine guarantee task ij is successfully executed The h-th start meal time of the employee e

[0077] The objective function of the multi-objective optimization of the airport ground service mixed skill task arrangement model is:

[0078]

[0079]

[0080] In the formula, |E| represents the number of employees, formula (1) maximizes the number of completed tasks, i.e., the total number of tasks completed by employees minus the total score of the penalty of double-machine guarantee; formula (2) minimizes the difference between the working hours of employees and the average of all employees' working hours, satisfying the working hour balance among employees.

[0081] The constraint conditions of the airport ground service mixed skill dispatching model are:

[0082]

[0083]

[0084]

[0085]

[0086]

[0087]

[0088]

[0089]

[0090]

[0091]

[0092]

[0093]

[0094]

[0095]

[0096]

[0097]

[0098]

[0099]

[0100]

[0101]

[0102]

[0103]

[0104]

[0105] where k denotes the kth task; g denotes the gth task; t denotes the tth available time period; denotes whether employee e has no auditing task to perform at time period t after performing auditing task i; denotes whether employee e has no auditing task to perform at time period t after performing auditing task i; denotes whether employee e has no auditing task to perform at time period t after performing auditing task i; denotes whether employee e has no auditing task to perform at time period t after performing auditing task i; denotes whether employee e has no auditing task to perform at time period t after performing auditing task i; denotes whether employee e has no auditing task to perform at time period t after performing auditing task i; j denotes the start time of auditing task j; denotes whether employee e has no auditing task to perform at time period t after performing auditing task i; denotes the start time of auditing task j; denotes whether employee e has no auditing task to perform at time period t after performing auditing task i; k denotes the start time of auditing task j; i denotes whether auditing task j is located at a remote position.

[0106] Equation (3) denotes that any employee performs a series of tasks starting from a virtual node and returns to the virtual node after performing the series of tasks; equation (4) denotes a flow balance constraint, i.e., if it is known that an employee completes three tasks in a certain time period, then it can be determined that the employee completes the tasks in the middle order in the time period; equation (5) requires that any auditing task is performed at most once; and equation (6) determines whether employee e can perform auditing task i, and if not, employee e cannot perform auditing task i.

[0107] Equation (7) relates two decision variables, i.e., the total number of all tasks completed in a time period in a certain order is the total number of tasks completed in the time period; equations (8) and (9) require that if two tasks belong to the same short-time transit flight, then the two tasks must be completed by the same employee; and equation (10) determines whether two auditing tasks can be performed consecutively, and if the two auditing tasks are double-machine guaranteed, then there is no time limit, otherwise, the start time of the second auditing task must be before the employee arrives at the position of the second auditing task after completing the first auditing task.

[0108] Formula (11) requires no task to be assigned 10 minutes before work; formula (12) and (13) require auditing tasks within a specified time range, formula (12) requires the start time of auditing task i to be no earlier than the time when employee e starts work in time period t; formula (13) requires that if an employee performs a task, the start time of the task is no later than the end time of the task after the working duration of the task; formula (14) requires no task to be assigned 10 minutes before work.

[0109] Formula (15) determines whether an employee can perform dual-machine support work; formula (16) relates two decision variables; formula (17) requires that an employee cannot perform consecutive dual-machine support work; formula (18) requires that the interval between two times of dual-machine support work performed by an employee be greater than 150 minutes; formula (19) requires that the time interval between two consecutive remote-machine work performed by an employee be greater than 150 minutes; formula (20) requires that an employee cannot perform remote-machine auditing tasks consecutively four times.

[0110] Formula (21)-(23) define meal times, formula (21) provides that the time when a certain time period t ends work is the time when the employee starts to eat the hth meal, wherein the time period t is equal in value to the available time period t h before the hth meal; formula (22) provides that the start time of work of a certain employee after a meal is equal to the end time of the previous work plus the meal time; formula (23) provides that the start time of a meal of a certain employee is between the earliest start time and the latest start time (time window); formula (24) calculates the working hours of an employee in a day, which is the total duration of performing each task and the transfer duration between tasks; formula (25) defines part of the decision variables as 0-1 variables.

[0111] The data in the "2022 National Civil Aviation Airport Production Statistics Bulletin" released by the Civil Aviation Administration of China shows that the passenger throughput of Xiamen Gaoqi International Airport ranks 17th among all airports in China, and the cargo and mail throughput ranks 12th among all airports in China. Xiamen Gaoqi International Airport is one of the twelve major trunk airports in China and is an important regional aviation hub in the southeast coast of China, belonging to the category of large airports. The following takes Xiamen Gaoqi International Airport as a model example of the analysis carrier of the invention, which has strong sufficiency and effectiveness.

[0112] The basic layout of Xiamen Gaoqi International Airport is shown in Figure 2 and Figure 3 , wherein the area of the near-machine position (bridge position) mainly has three areas, including the 1st stop area, the 20A stop area, and the 20B stop area, and the remaining stop areas are remote-machine positions. Xiamen Gaoqi International Airport currently has 97 parking spaces, distributed in 13 stop areas. For each stop area of the airport, there are a number of parking spaces in each area.

[0113] If the two consecutive audit tasks of an employee are located in different aprons, there is a commuting time between the two areas, and the specific commuting time is τ ij As shown in Table 2 below.

[0114] Table 2 Commuting time between each apron area

[0115]

[0116] Table 3 below shows the relevant variables and audit data of 3 employees about work period, meal times, meal period, meal time, etc.

[0117] Table 3 Employee work and meal schedule

[0118]

[0119]

[0120] At the same time, Table 4 and Table 5 list part of the flight schedule of the arrival (A) at Xiamen Gaoqi International Airport and part of the flight schedule of the departure (D) at Xiamen Gaoqi International Airport. Among them, ID is the identification of each corresponding flight task (which can come from the same aircraft, such as 1 and 2 in the arrival list). In order to facilitate program identification, each time is preprocessed from the combination of "hours + minutes" to "minutes (from 0 o'clock of the day)". O i For the start time of the audit task i, there is a difference between the arrival and departure requirements - the arrival audit task is 15 minutes before the scheduled landing time, and the departure audit task is 50 minutes before the scheduled take-off time, so the O i in Table 4 and Table 5 are located in different columns.b ij List indicates the type of flight task (origin, transit, post-flight); the cargo list indicates the type of flight cargo (bulk, freighter, container). As for the "gate" data selected by the present application, all data (1-232) have been assigned to the gate; ζ i indicates whether the audit task i is located in the remote gate, i.e. whether the apron of the flight task is adjacent to the terminal; the last list indicates the apron area where the flight task is located.

[0121] Table 4 Flight task table (part) of arrival (A) at Xiamen Gaoqi International Airport (ZSAM)

[0122]

[0123] Table 5 Flight task table (part) of departure (D) at Xiamen Gaoqi International Airport (ZSAM)

[0124]

[0125]

[0126] The duration Gantt chart of the eight flight tasks in Table 4 and Table 5 is shown in Figure 1, which is within 1560 minutes (25 hours and 40 minutes), because the scheduled arrival time of the fourth flight task is 1560 minutes (2:00 am the next day). Figure 4

[0127] In addition, the qualifications of each employee are shown in Table 6 and Table 7. In Table 6, it is shown that the employees have different specialities for different tasks, i.e. "mixed skills", some employees need to perform more types of tasks, and some employees are good at a certain type of task. In Table 7, it is shown whether the employees can perform dual aircraft protection, i.e. if they have the ability to perform dual aircraft protection, the value is 1, otherwise it is 0.

[0128] Table 6 Employee qualifications for tasks Corresponding table

[0129]

[0130] Table 7 Employee qualifications for dual aircraft protection γ e Corresponding table

[0131]

[0132] The following parameters involved in the solving process are set, and the specific settings are shown in Table 8. The penalty score of dual aircraft protection refers to the number of "completed tasks" deducted from an employee after performing a dual aircraft protection task. This parameter appears in the first optimization objective, aiming to allow the employee to complete more tasks while minimizing the number of dual aircraft protection tasks.

[0133] Table 8 Model parameter table

[0134]

[0135] The present application is based on Python programming, reading data files and writing related codes for calling Gurobi solver, translating variables, objectives and constraints. The following shows part of the program code.

[0136] The objective function Python program is as follows:

[0137] model.setObjectiveN(-obj1,index=0,weight=1,priority=5,name='obj_1')

[0138] model.setObjective(obj1,GRB.MAXIMIZE) ​

[0139] model.setObjective(sum(y_ite[i,t,e]for e in np.arange(E)for iinnp.arange(1,I+1)for t in T_et[e])-sum(epsilon_e[e] / (4*E)*z_ije[i,j,e]fore innp.arange(E)for iin np.arange(1,I+1)for j in np.arange(I+1)if i!=j),GRB.MAXIMIZE)

[0140] model.setObjectiveN(sum((d_e[e]-sum(d_e[e] / E for e in np.arange(E)))for e in np.arange(E)),index=0,weight=1,priority=5,name='obj_2')

[0141] Constraint (Equation (14)) Python program is:

[0142] for e in np.arange(E):

[0143] for iin np.arange(1,I+1):

[0144] if len(T_et[e])>0:

[0145] lhs=LinExpr(0)

[0146] lhs.addTerms((P_et[e][T_et[e][-1]]-10)*(1-M),y_ite[i,T_et[e][-1],e])

[0147] model.addConstr(lhs>=O_i[i]-(P_et[e][T_et[e][-1]]-10)*M,name='constr12'+str(e)+str(i))#+str(t))

[0148] Python program (excerpt) for output results:

[0149] if model.status<3:

[0150] y_ite_all=[]#np.zeros((I+1,T,E))

[0151] y_ite_all_list = []

[0152] for i in np.arange(I + 1):

[0153] for t in np.arange(T):

[0154] for e in np.arange(E):

[0155] # y_ite_all_list[i,t,e] = y_ite[i,t,e].X

[0156] y_ite_all.append(pd.DataFrame([{'i': i, 't': t, 'e': e, 'y_ite': y_ite[i, t, e].X}]))

[0157] # print(y_ite_all)

[0158] y_ite_all = pd.concat(y_ite_all, axis=0, ignore_index=True)

[0159] y_ite_all.to_csv("y_ite.csv")

[0160] else:

[0161] print('has not solution')

[0162] For objective function 1, "Optimal solution found, Best objective 7.000060000000e+00", the optimal solution is 7. Through the output of the related.csv file, 3 employees are assigned to 8 tasks, and the double-machine guarantee times are 1, so deduct 1 double-machine guarantee penalty score ε e , the optimal solution value 7 is the total number of tasks completed on the day minus the value of the double-machine guarantee times.

[0163] For objective function 2, "Optimal solution found, Best objective 1.600000000000e+02", the optimal solution is 160, through the output of The relevant.csv file shows that 3 employees respectively perform ① 1 inbound task, 3 outbound tasks, ② 2 inbound tasks, 1 outbound task, and ③ 1 inbound task, 0 outbound task, and the individual working hours are 190, 130 and 40 respectively. The average working hours of the employees are 120, and the working hour deviations are 70, 10 and 80 respectively. The total working hour deviation is 160, and the optimal solution value 160 is the sum of the working hour deviations of all employees on that day.

[0164] The output results are converted into a more intuitive table of employees e and time periods t, tasks i, and combined with double-machine support, the same short-time overflight flights and other special situations and working hours, which can more clearly browse the dispatching situation of the model output.

[0165] Among them, each employee is matched with the corresponding task, and each employee has a task in each working period, that is, if an employee eats 2 meals a day, he has a task in 3 working periods.

[0166] Table 9 Specific situation of employee e performing task i in time period t

[0167]

[0168]

[0169] Comparing Table 3, Table 6 and Table 9, each employee has obtained a task in his working period, there is no similar situation that an employee who has 3 working periods (eats 2 meals) has 1 or 2 completely idle periods, and each employee can eat according to his relatively comfortable time range, as shown in Table 10 and Table 11.

[0170] Table 10 Comparison of employee qualification and task execution

[0171]

[0172] In Table 10, the bold italic "1" indicates that the employee has the qualification and is assigned to the task, the normal black "1" only indicates that the employee has the qualification, and the space indicates that the employee does not have the qualification and cannot be assigned to the task.

[0173] Table 11 Employee meal time and comfortable interval

[0174]

[0175] In this example, the corresponding positions of the three employees are in the small sub-areas of the 20 area (20A, 20B, 20C), which simulates the execution of tasks by personnel near the T4 terminal (domestic departure). The output results show the working order of the three employees, such as Figure 5As shown in the diagram, employee 1 completed dual-aircraft support work at gates 201 and 202 during their first work session of the day (morning). Employee 2 performed two tasks for the same short-term transit flight during their first work session of the day (afternoon), at gate 208. The work sequence shown in the diagram does not involve long-distance commutes or four consecutive remote gate operations.

[0176] For important parameters such as work hour deviation, the proportion of comfortable meal time, and the maximum meal time offset, this example shows significant optimization compared to the situation in December 2022. The calculation formulas include: an expression for solving the average work hour deviation for each worker; and an expression for solving the total proportion of comfortable meal times for workers, where c... e This represents the number of times an employee can comfortably eat in a day; the expression for solving the maximum meal time offset is as follows:

[0177]

[0178]

[0179]

[0180] Detailed data as follows Figure 6 As shown, this includes determining lunch time shifts (including lunch for employees who eat two meals and lunch for employees who eat only one meal), and determining dinner time shifts (including dinner for employees who eat two meals and dinner for employees who eat only one meal).

[0181] The following model will be solved using actual data from Xiamen Gaoqi International Airport. The specific original employee data (partial) is shown in Table 12, which displays relevant variables and audit data of 30 employees regarding work hours, number of meals, meal times, and meal duration.

[0182] Table 12 Employee Work Schedule

[0183]

[0184]

[0185] The settings are based on the parameters involved in the solution process, as shown in Table 13.

[0186] Table 13 Model Parameter Table

[0187]

[0188]

[0189] For objective function 1, "Best objective 5.400000000000e+01", the optimal solution is 54. Through the output The relevant.csv file shows that 30 employees are assigned to 60 tasks, with double-machine guarantee times of 6, so deduct 6 double-machine guarantee penalty scores ε e The optimal solution value 54 is the total number of tasks completed by the employees on that day minus the number of double-machine guarantee times.

[0190] For objective function 2, "Optimal solution found, Best objective 8.220000000000e+02", the optimal solution is 822. Through the output The relevant.csv file shows that the total work hour deviation of 30 employees is 160, and the optimal solution value 822 is the total work hour deviation of all employees on that day.

[0191] From Figure 7 It can be seen that each task is assigned to a specific employee to execute. Extract the "26" employee in the figure, which is the 27th employee, and the task numbers he executes are 1, 30, 31, 43, and 58 (corresponding to 0, 29, 30, 42, 57 characters in the figure). Listing the specific information of these tasks can more intuitively see the employee's work process on that day, as shown in Table 14.

[0192] Table 14 Specific information of tasks executed by employee No. 27 on that day

[0193]

[0194] Through the execution sequence and execution area, the employee's work location and sequence transfer diagram on that day can be drawn, as shown in Figure 7 .

[0195] For important parameters such as work hour deviation, this example has obtained a more obvious optimization compared to the current situation in December 2022, while also considering two new indicators. The specific calculation formulas include: the expression for solving the average number of workers needed per hundred tasks; and the expression for solving the average time utilization.

[0196] 100x|E| / |I|

[0197]

[0198] Figure 8 The tornado chart for this work allocation optimization project comparison. Figure 8The total working hours deviation of the 30 employees presented was 822 minutes, with an average deviation of 27.4 minutes per employee, a significant decrease of 60% compared to 68.5 minutes in December 2022. According to the airport ground service hybrid skills dispatch model, the percentage of times meal times fell within the comfortable range (lunch 11:00-13:00, dinner 17:00-19:00) increased from 71.2% to 98.5%, an increase of 38.3%.

[0199] The shift in meal times during this dispatch was at most 17 minutes for employee 25 (lunch at 13:17), a significant decrease compared to the previous 88 minutes (dinner at 15:32). The average number of people required per task increased from 0.87 to 0.97, reflecting a reduction in the number of dual-aircraft support operations. The average time utilization rate decreased slightly from 85.6% to 82.3%, indicating a slight reduction in overall working time, meaning that employees completed the same tasks with a more balanced schedule and more comfortable work-rest balance. In conclusion, the airport can save on employee hiring costs and improve its reputation through this optimized dispatch model.

[0200] The above description is a detailed description of the preferred embodiments of the present invention. However, the embodiments are not intended to limit the scope of the patent application of the present invention. All equivalent changes or modifications made under the technical spirit disclosed in the present invention should fall within the patent scope covered by the present invention.

Claims

1. An airport ground service dispatch optimization method considering time windows and airport conditions, characterized in that, Comprise the following steps: S1. Establish an airport ground service mixed skill dispatching model based on a team-oriented problem with a time window, and describe the airport ground service scheduling problem through the model; S2. Establish a target function for multi-objective optimization of the airport ground service mixed skill dispatching model, and determine the number of completed ground service tasks and the balance of employee working hours as optimization objectives; S3. According to the actual situation of the airport, set the constraint conditions of the airport ground service mixed skill dispatching model, and set the parameters of the airport ground service mixed skill dispatching model; S4. Based on Python, write a program to read a data file and then call the Gurobi solver to obtain the dispatching optimization scheme of the airport ground service; The target function for multi-objective optimization of the airport ground service mixed skill dispatching model is: In the formula, |E| represents the number of employees, formula (1) maximizes the number of completed tasks, that is, the total number of tasks completed by employees minus the total score of the double-machine guarantee penalty; Formula (2) minimizes the difference between the working hours of employees and the average of all employee working hours, satisfying the working hour balance among employees; The constraint conditions of the airport ground service mixed skill dispatching model are: where k represents the kth task; g represents the gth task; t represents the tth available time period; represents whether employee e has no audit task to perform after performing audit task i at time period t; represents whether employee e has no audit task to perform after performing audit task i at time period t; represents whether employee e has no audit task to perform after performing audit task i at time period t; represents whether employee e has no audit task to perform after performing audit task i at time period t; represents whether employee e has no audit task to perform after performing audit task i at time period t; represents whether employee e has no audit task to perform after performing audit task i at time period t; j represents whether employee e has no audit task to perform after performing audit task i at time period t; represents whether employee e has no audit task to perform after performing audit task i at time period t; represents whether employee e has no audit task to perform after performing audit task i at time period t; represents whether employee e has no audit task to perform after performing audit task i at time period t; k represents whether employee e has no audit task to perform after performing audit task i at time period t; j represents whether employee e has no audit task to perform after performing audit task i at time period t; E denotes a set of employees, e denotes the e-th employee, I denotes a set of tasks, i denotes the i-th task, j denotes the j-th task; T e denotes a set of time periods in which employee e can perform a task; denotes whether employee e performs audit task i at time period t; denotes whether dual machine safeguard task ij is successfully performed; ε e ε represents the penalty score of employee e for performing a double-crewed task once;d e ε represents the total working hours of employee e for performing a task in a day; ε represents whether employee e is qualified for audit task i; ε represents whether employee e performs audit task j after performing audit task i in time period t;b ij ε represents whether audit task i and audit task j belong to the same short-haul transit flight;O i ε represents the start time of audit task i;v i ε represents the duration of audit task i; τ ij denotes the time for the staff to move from the location of audit task i to the location of audit task j; denotes the start working time of staff e at time period t; denotes the end working time of staff e at time period t;c ij denotes whether audit task i and audit task j meet the condition of double-machine guarantee;γ e denotes whether staff e has the qualification of double-machine guarantee;ξ i denotes whether audit task i is located at a remote location; denotes the hth start eating time of staff e; h denotes the hth eating of staff;H e denotes the set of eating times of staff e;t h denotes the previous available time period corresponding to the hth eating;Γ h denotes the time length required for the hth eating in a day;Θ h denotes the earliest start time of the hth eating;Φ h denotes the latest start time of the hth eating.

2. The method of claim 1, wherein, The airport ground service scheduling problem is described through the model, specifically: During the ground service process, it is necessary to determine whether the employee e has the qualification for the audit task i whether the employee e has the dual-machine guarantee qualification γ e , whether the audit task i and the audit task j belong to the same short-time transit flight b ij , whether the audit task i and the audit task j meet the condition c of dual-machine guarantee ij , set the penalty score ε of the employee e performing a dual-machine guarantee task according to the optimization target e ; In the airport ground service task, the completion of the task and the arrangement of the staff are included, and during the airport ground service task, the start time O of the audit task i needs to be determined i , whether the audit task i is located at the remote position ξ i , the duration v of the audit task i is determined i ; For staff scheduling, the time τ for staff to move from the location of audit task i to the location of audit task j needs to be considered ij , the set of time periods T in which staff e can perform tasks e , the start work time of staff e at time period t , the end work time of staff e at time period t , the duration Γ of the hth meal in a day h , the set of meal times H of staff e e , the previous available time period t corresponding to the hth meal h , the earliest start time Θ of the hth meal h , the latest start time Φ of the hth meal h , the total working hours d of staff e in a day e .

3. The method of claim 2, wherein, airport ground service mixed skill dispatching model, which combines task completion and scheduling optimization to make decisions on staff task arrangement, considers whether the staff e performs the audit task j after completing the audit task i at the time period t whether the staff e performs the audit task i at the time period t whether the dual-machine guarantee task ij is successfully performed the hth time the staff e starts eating

Images