Airport time configuration method and system based on large neighborhood search

Abstract

The invention belongs to the field of flight plan optimization, and provides an airport time configuration method and system based on large neighborhood search. The method comprises the following steps: constructing a moment configuration model; sorting the time applications based on preset influence factor priorities, and sequentially inserting the time applications into a pre-constructed null time allocation scheme according to a sorting result to obtain an initial solution of the time configuration model; randomly selecting a target neighborhood operator structure from the time window neighborhood operator structure, the chain type neighborhood operator structure and the airport combination neighborhood operator structure based on a roulette principle, and optimizing the initial solution based on the target neighborhood operator structure to obtain an intermediate solution; then introducing a large neighborhood algorithm to optimize the intermediate solution to obtain an approximate integer solution; and taking the approximate integer solution as a target solution of the time configuration model to obtain a flight configuration result and a flight plan. According to the invention, the continuity of the flight planning scheme and the interests of airlines are ensured.

Description

Technical Field

[0001] The present invention relates to the field of flight plan optimization, and in particular to an airport schedule configuration method based on large neighborhood search. Background Art

[0002] This section merely provides background information related to the present disclosure and is not necessarily prior art.

[0003] Airport slot allocation in the air transport system is of great significance to ensure the efficient and equitable use of limited airport infrastructure.

[0004] The slot allocation problem aims to allocate takeoff and landing slots to appropriate requests, given the runway capacity of one or more airports and airline requests for flight operations at specific slots, while taking into account various priorities, conflicts, and capacity requirements. In practice, the large number of slot requests, complex slot allocation rules, and the need for seamless coordination across airport networks all contribute to the high complexity of the slot allocation problem. This complexity is further exacerbated by the need to balance conflicting interests and priorities.

[0005] Existing technologies for addressing the highly complex problem of airport slot allocation suffer from the following shortcomings. Given the complexity of airport slot allocation, slot allocation models are typically applied only to a single airport, ignoring the strong complementarity between slot allocations for individual flights at both departure and destination airports, lacking practical application considerations. Most slot allocation models also fail to account for situations where slot requests exceed airport capacity. This means that no matter how slots are allocated, it may be impossible to allocate all requested slots. Therefore, these models are nearly incapable of addressing this scenario. Summary of the Invention

[0006] In view of the shortcomings of the existing technology, the present invention provides an airport schedule configuration method based on large neighborhood search to improve the technical problems of the existing technology that only a single airport is considered when performing schedule configuration and the request exceeds the capacity.

[0007] A method for configuring airport schedules based on large neighborhood search includes the following steps:

[0008] Step 1: Construct a schedule configuration model based on flight operation data; the schedule configuration model includes an objective function and constraints.

[0009] Furthermore, step 1 includes a flight time matrix, a time slot application data table, and airport capacity information.

[0010] Step 1 specifically includes:

[0011] Step 1-1: Collect the time slot applications submitted by each airline and extract the flight time matrix through the time slot applications; the time slot application includes the operating airline, take-off and landing airports, requested take-off time, requested landing time, application cycle, operating week, requested flight type, operating aircraft type and connecting flights.

[0012] Step 1-2: Construct a slot application matrix based on the slot application date and time, and construct slot application constraint information based on the slot application priority type and connecting flights.

[0013] Step 1-3: Collect airport capacity information, where the airport capacity information includes 60-minute take-off and landing capacity and 15-minute take-off and landing capacity. The 60-minute take-off and landing capacity and the 15-minute take-off and landing capacity include runway capacity and apron capacity, respectively.

[0014] Steps 1-4: Construct a time configuration model based on the collected data. The time configuration model is constructed using an operations research solver and includes two parts: an objective function and constraints. The objective function includes, but is not limited to, maximum time offset, total time offset, fairness index, number of unallocated flights, and their combinations. The constraints include, but are not limited to, airport take-off and landing flights, aircraft turnover constraints, etc.

[0015] Furthermore, the objective function in step 1 includes the maximum time offset Total time offset Dp max and unallocated flight frequencies∑ i∈R ∑ d∈ DM id N i .

[0016] Among them, i is the index of the moment application, d is the date index, For the time of application i, the positive offset is is the negative offset of the moment of application i, M id Indicates whether the application i is running on day d; Dp max express or The maximum value in N i Indicates whether the application i at the moment is allocated.

[0017] The constraints include capacity constraints, turnaround constraints and priority constraints; the capacity constraints include airport takeoff constraints, airport landing constraints and airport total takeoff and landing constraints; the turnaround constraints include maximum turnaround time constraints for connecting flights and minimum turnaround time constraints for connecting flights; the priority constraints use time offsets to constrain time applications of different priorities.

[0018] Furthermore, in the capacity constraint,

[0019] The airport takeoff frequency constraint is:

[0020]

[0021] The airport landing constraints are:

[0022]

[0023] The total take-off and landing constraints of the airport are:

[0024]

[0025] Where a is the index of the airport, c is the time period, d is the date index, and i is the index of the time application; F i The flight time requested for moment i, is the set of departure flights from airport a, L C The capacity time resolution is generally 1h or 15min. is the departure capacity of airport a during period c on day d, is the arrival capacity of airport a during period c on day d, is the total capacity of airport a during period c on day d; dep and tol represent the departure capacity and total capacity respectively; B it is the allocation time, which is 0 when the time application i is allocated is earlier than time t, otherwise it is 1; M id Indicates whether application i is running on day d. If application i is running on day d, it is 1, otherwise it is 0; N i Indicates whether the application i at the moment is allocated, if it is allocated, it is 0, if not allocated, it is 1.

[0026] In the turnover constraint,

[0027] The maximum turnaround time constraint for the connecting flight is:

[0028]

[0029] The minimum turnaround time constraint for the connecting flight is:

[0030]

[0031] Among them, T max and T min are the maximum turnaround time and minimum turnaround time of connecting flights, respectively. ω is a maximum value used to remove the turnaround restrictions on unassigned flights. i is the earliest allocation time for application i, LR j is the earliest allocation time for applying for time j, UR i The latest allocation time for application i, URj The latest allocation time for applying for time j.

[0032] The priority constraints are:

[0033]

[0034] in,

[0035] Among them A it is the application time of time application i, which is related to the specific time t. If t is earlier than the application time, then A it is 1, otherwise it is 0; H it is the historical running time of application i, which is related to the specific time t. If t is earlier than the application time, then H it is 1, otherwise it is 0; R hc A collection of all moment requests for historical moment changes.

[0036] Step 2: Sort each time slot application based on the preset priority of the influencing factors, and insert each time slot application into the pre-built empty time slot allocation plan in sequence according to the sorting results to obtain the initial solution of the time slot configuration model obtained in step 1.

[0037] Furthermore, the initial solution includes a set of allocated flights obtained by taking the position with the smallest insertion penalty value as the target insertion position, and a set of unallocated flights consisting of time slot requests that cannot be inserted; the penalty value is determined by the number of capacity conflicts, the total offset, and the number of connecting flight violations.

[0038] Step 3: Based on the target neighborhood operator structure, the initial solution obtained in step 2 is optimized to obtain an intermediate solution. The intermediate solution is iteratively optimized by introducing a large neighborhood algorithm to obtain an approximate integer solution.

[0039] Furthermore, step 3 specifically includes:

[0040] Step 3-1: Divide all the time applications into two categories. One category is the time applications selected in the target neighborhood operator structure. The configuration variables B corresponding to this part of the application it No changes are made, and the remaining variables that are not in the target neighborhood operator structure are applied for the corresponding variables B at other times it Do fixed processing, that is, set B it The upper and lower bounds are both B in the last iteration result it to ensure that the values of these variables remain unchanged during the solution process.

[0041] Step 3-2: Iterate and solve based on the branch-and-bound algorithm to obtain the optimal linear upper bound and the corresponding approximate integer solution.

[0042] Step 3-3: If the preset number of iterations has not been reached but the preset iteration duration has been reached, and the output solution is not optimal and the solution gap is greater than the preset gap, the neighborhood is reduced based on the preset neighborhood reduction strategy and iterative optimization is performed again;

[0043] Step 3-4: Select a more suitable target neighborhood operator structure by improving the target neighborhood operator structure.

[0044] Furthermore, the neighborhood reduction strategy described in step 3-3 is specifically as follows: setting the weight τ of the selected application i at the moment i The initial value of the reduction parameter ρ is set to 1, and it is stipulated that if the solution cannot be found within the specified time and the solution gap is greater than the preset gap, the reduction parameter ρ'=ρ×α is modified to reduce the selected weight τ i '=τ i ×α, where α is the preset reduction coefficient and lowering coefficient;

[0045] If the accuracy of the output solution is higher than that of the previous iterative solution after reaching the preset number of iterative calculations, the output solution is used as the approximate integer solution. If the accuracy of the output solution is lower than that of the previous iterative solution, the neighborhood is reduced based on the preset neighborhood reduction strategy and then the iterative optimization is performed again. The comparison of accuracy in the method depends on the setting of the objective function value. If the objective function value is maximized, the larger the objective function value of the solution, the more accurate it is. Conversely, when the objective function value is minimized, the smaller the objective function value of the solution, the more accurate it is.

[0046] Furthermore, the method for selecting the target neighborhood operator structure in step 3 is to randomly select from pre-constructed target neighborhood operator structures based on the roulette principle, and the pre-constructed target neighborhood operator includes a time window neighborhood operator structure, a chain neighborhood operator structure and an airport combination neighborhood operator structure.

[0047] The time window neighborhood operator structure includes all flights of the target airport A in the period (t, t+r), and all flights of the associated airport B in the period (t+T ab , t+T ab +r) period and (t+T ab , t+T ab +r) all flights within the time period, where T ab represents the flight time from airport a to airport b, T ab It represents the flight time from airport b to airport a, t is the start time of the time window, and r is the step size of the time window.

[0048] The chained neighborhood operator structure is a selection set, which is a set consisting of a preset number of time applications selected from the candidate time applications, using the penalty value of the time application as a weight.

[0049] The airport combination neighborhood operator structure is composed of all time slot applications involving the combined airports.

[0050] Furthermore, the specific adjustment method for improving the target neighborhood operator structure in steps 3-4 is as follows:

[0051] When the accuracy of the approximate integer solution is equal to the accuracy of the intermediate solution, the weight of the target neighborhood operator structure is reduced.

[0052] When the accuracy of the approximate integer solution is higher than the accuracy of the intermediate solution, but the difference between the two is less than the target difference, the weight of the target neighborhood operator structure is kept unchanged.

[0053] When the accuracy of the approximate integer solution is higher than the accuracy of the intermediate solution, and the difference between the two is not less than a target difference, the weight of the target neighborhood operator structure is increased.

[0054] Step 4: The approximate integer solution obtained in step 3 is used as the target solution of the time configuration model. The target solution is the flight configuration result. The flight configuration result and flight plan are obtained through the target solution.

[0055] An airport schedule configuration system based on large neighborhood search includes a model building module, an initial solution acquisition module, an approximate integer solution acquisition module, and a configuration scheme acquisition module.

[0056] The model building module builds a time configuration model based on flight operation data, wherein the flight operation data includes a flight time matrix, a time application data table, and airport capacity information, and the time configuration model includes an objective function and constraint conditions.

[0057] The initial solution acquisition module sorts each time application based on a preset priority of influencing factors, and inserts each time application into a pre-built empty time allocation scheme in sequence according to the sorting result to obtain an initial solution of the time configuration model.

[0058] The approximate integer solution acquisition module randomly selects a target neighborhood operator structure from the pre-constructed time window neighborhood operator structure, chain neighborhood operator structure and airport combination neighborhood operator structure based on the roulette principle, optimizes the initial solution based on the target neighborhood operator structure to obtain an intermediate solution, and iteratively optimizes the intermediate solution through a large neighborhood algorithm to obtain an approximate integer solution.

[0059] The configuration solution acquisition module uses the approximate integer solution as the target solution of the time configuration model, and then obtains the flight configuration result and flight plan.

[0060] Beneficial effects:

[0061] It can be seen from the above technical solutions that the technical solution of the present invention provides an airport schedule configuration method based on large neighborhood search, so as to improve the technical defects of the existing technology that only a single airport is considered when performing schedule configuration and the request overcapacity situation cannot be solved. Specifically, this technical solution first proposes a type of schedule configuration model that considers a multi-airport network and constructs three operators according to its characteristics for solution. It dynamically considers the impact of flight duration on the airports, and the proposed mixed integer programming model combines multiple constraints such as airport capacity and airline turnover. Secondly, this technical solution also introduces the option of non-allocation to make the solution of the problem more in line with the actual situation. At the same time, it comprehensively considers the advantages of the high dimensionality of the solution space of the plan formulation, and correspondingly proposes a large neighborhood algorithm and three operator structures to solve this type of problem efficiently and accurately, solving the problem of extremely high complexity in the optimization process. BRIEF DESCRIPTION OF THE DRAWINGS

[0062] The present invention will be further described below in conjunction with the accompanying drawings and specific embodiments, and the above and / or other advantages of the present invention will become more apparent.

[0063] Figure 1 Flowchart of the airport schedule configuration method based on large neighborhood search described in this embodiment.

[0064] Figure 2 A flowchart of constructing the time configuration model for this embodiment.

[0065] Figure 3 Schematic diagram of the time window neighborhood operator structure described in this embodiment.

[0066] Figure 4 Schematic diagram of the chain neighborhood operator structure described in this embodiment.

[0067] Figure 5 This is a flowchart for solving the approximate integer solution described in this embodiment.

[0068] Figure 6 This is a structural block diagram of the airport schedule configuration system based on large neighborhood search described in this embodiment.

[0069] Figure 7 This is a structural block diagram of the electronic device described in this embodiment. DETAILED DESCRIPTION

[0070] To make the purpose, technical solutions and advantages of the embodiments of the present invention more clear, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention.

[0071] Existing approaches to the highly complex problem of airport slot allocation only consider the single-airport problem. While this simplifies the problem, it overlooks the strong complementarity between slot allocations for individual flights at the departure and destination airports, resulting in a failure to meet actual needs. Furthermore, it fails to consider the conditions under which slot requests exceed airport capacity, making it impossible to address such situations. Based on this, this embodiment provides an airport slot allocation method based on a large neighborhood search, seeking the optimal approach to slot allocation from a global perspective to ensure efficient and equitable use of limited airport infrastructure.

[0072] The following is a detailed introduction to the airport schedule configuration method based on large neighborhood search described in this embodiment with reference to the accompanying drawings.

[0073] like Figure 1 As shown, the method includes the following steps:

[0074] Step 1: Construct a schedule configuration model based on flight operation data; the schedule configuration model includes an objective function and constraints.

[0075] Step 2: Sort each time slot application based on the preset priority of the influencing factors, and insert each time slot application into the pre-built empty time slot allocation plan in sequence according to the sorting results to obtain the initial solution of the time slot configuration model obtained in step 1.

[0076] Step 3: Based on the target neighborhood operator structure, the initial solution obtained in step 2 is optimized to obtain an intermediate solution. The intermediate solution is iteratively optimized by introducing a large neighborhood algorithm to obtain an approximate integer solution.

[0077] Step 4: The approximate integer solution obtained in step 3 is used as the target solution of the time configuration model, and the flight configuration result and flight plan are obtained through the target solution.

[0078] Furthermore, the flight operation data in step 1 includes a flight time matrix, a time application data table, and airport capacity information; Figure 2 As shown, step 1 specifically includes:

[0079] Step 1-1: Collect the time slot applications submitted by each airline and extract the flight time matrix through the time slot applications; the time slot application includes the operating airline, take-off and landing airports, requested take-off time, requested landing time, application cycle, operating week, requested flight type, operating aircraft type and connecting flights.

[0080] Step 1-2: Construct a slot application matrix based on the slot application date and time, and construct slot application constraint information based on the slot application priority type and connecting flights.

[0081] Step 1-3: Collect airport capacity information, where the airport capacity information includes 60-minute take-off and landing capacity and 15-minute take-off and landing capacity. The 60-minute take-off and landing capacity and the 15-minute take-off and landing capacity include runway capacity and apron capacity, respectively.

[0082] Steps 1-4: Construct a time configuration model based on the collected data. The time configuration model is constructed using an operations research solver and includes two parts: an objective function and constraints. The objective function includes, but is not limited to, maximum time offset, total time offset, fairness index, number of unallocated flights, and their combinations. The constraints include, but are not limited to, airport take-off and landing flights, aircraft turnover constraints, etc.

[0083] The objective function in step 1 includes the maximum time offset Total time offset Dp max and unallocated flight frequencies∑ i∈R ∑ d∈ DM id N i In specific implementation, a certain weight coefficient is set to balance the relationship among the three, and the weight coefficient can be evaluated and selected according to actual needs.

[0084] Among them, i is the index of the moment application, d is the date index, For the time of application i, the positive offset is is the negative offset of the moment of application i, M id Indicates whether the application i is running on day d; Dp max express or The maximum value in N i Indicates whether the application i at the moment is allocated.

[0085] The constraints include capacity constraints, turnover constraints and priority constraints.

[0086] Specifically, the capacity constraint includes an airport take-off constraint, an airport landing constraint, and an airport total take-off and landing constraint.

[0087] The airport takeoff frequency constraint is:

[0088]

[0089] The airport landing constraints are:

[0090]

[0091] The total take-off and landing constraints of the airport are:

[0092]

[0093] Where a is the index of the airport, c is the time period, d is the date index, and i is the index of the time application; F i The flight time requested for moment i, is the set of departure flights from airport a, L C The capacity time resolution is generally 1h or 15min. is the departure capacity of airport a during period c on day d, is the arrival capacity of airport a during period c on day d, is the total capacity of airport a during period c on day d; dep and tol represent the departure capacity and total capacity respectively; B it is the allocation time, which is 0 when the time application i is allocated is earlier than time t, otherwise it is 1; M id Indicates whether the application i is running on day d. If the application i is running on day d, it is 1, otherwise it is 0; N i Indicates whether the application i at the moment is allocated, if it is allocated, it is 0, if not allocated, it is 1.

[0094] The turnaround constraints include a maximum turnaround time constraint for connecting flights and a minimum turnaround time constraint for connecting flights.

[0095] Specifically, the maximum turnaround time constraint for the connecting flight is:

[0096]

[0097] The minimum turnaround time constraint for the connecting flight is:

[0098]

[0099] Among them, T max and T min are the maximum turnaround time and minimum turnaround time of connecting flights, respectively. ω is a maximum value used to remove the turnaround restrictions on unassigned flights. i is the earliest allocation time for application i, LR j is the earliest allocation time for applying for time j, UR i The latest allocation time for application i, UR j The latest allocation time for applying for time j.

[0100] The priority constraint uses time offset to constrain time applications of different priorities.

[0101] Specifically, the definition:

[0102]

[0103]

[0104] Then the priority constraint is:

[0105]

[0106] Among them A it is the application time of time application i, which is related to the specific time t. If t is earlier than the application time, then A it is 1, otherwise it is 0; H it is the historical running time of application i, which is related to the specific time t. If t is earlier than the application time, then H it is 1, otherwise it is 0; R hc A collection of all moment requests for historical moment changes.

[0107] Step 2 specifically involves sorting each time slot application based on a preset priority of influencing factors, and inserting each time slot application into a pre-built empty time slot allocation scheme in sequence according to the sorting result to obtain an initial solution of the time slot configuration model.

[0108] In specific implementation, all slot applications are collected and sorted according to the slot application priority type, slot application date, and connecting flight. The slot application priority type takes precedence over the slot application date, and the slot application date takes precedence over the connecting flight.

[0109] All flight applications are prioritized in the following order: historical priority, historical schedule changes, and new carriers. For flights with the same priority, they are sorted by number of operations, with the longest number of operations ranking higher. For example, a flight operating on Wednesdays and Fridays, if its application runs between July 1, 2024, and August 1, 2024, has a number of operations of 9. For connecting flights a and b, the later-ranked flight will be used to ensure simultaneous insertion and avoid conflicts.

[0110] The initial solution includes a set of allocated flights, obtained by targeting the insertion location with the lowest insertion penalty, and a set of unallocated flights, consisting of slot requests for which insertion was not possible. The penalty value is determined by the number of capacity conflicts, the total offset, and the number of connecting flight violations.

[0111] Step 3 specifically includes:

[0112] Step 3-1: Divide all the time applications into two categories. One category is the time applications selected in the target neighborhood operator structure. The configuration variables B corresponding to this part of the application it No changes are made, and the remaining variables that are not in the target neighborhood operator structure are applied for the corresponding variables B at other times it Do fixed processing, that is, set B itThe upper and lower bounds are both B in the last iteration result it to ensure that the values of these variables remain unchanged during the solution process.

[0113] Step 3-2: Iterate and solve based on the branch-and-bound algorithm to obtain the optimal linear upper bound and the corresponding approximate integer solution.

[0114] The strategies for iteratively optimizing the intermediate solutions in step 3-2 to obtain approximate integer solutions include:

[0115] If the preset number of iterations has not been reached but the preset iteration duration has been reached, and the output solution is not optimal and the solution gap is greater than the preset gap, the neighborhood is reduced based on the preset neighborhood reduction strategy and iterative optimization is performed again;

[0116] If the accuracy of the output solution is higher than that of the previous iterative solution after reaching the preset number of iterative calculations, the output solution is used as the approximate integer solution. If the accuracy of the output solution is lower than that of the previous iterative solution, the neighborhood is reduced based on the preset neighborhood reduction strategy and then the iterative optimization is performed again. The comparison of accuracy in the method depends on the setting of the objective function value. If the objective function value is maximized, the larger the objective function value of the solution, the more accurate it is. Conversely, when the objective function value is minimized, the smaller the objective function value of the solution, the more accurate it is.

[0117] The method for selecting the target neighborhood operator structure in step 3 is to randomly select from pre-constructed target neighborhood operator structures based on the roulette principle. The pre-constructed target neighborhood operator includes a time window neighborhood operator structure, a chain neighborhood operator structure, and an airport combination neighborhood operator structure.

[0118] The time window neighborhood operator structure includes all flights of the target airport A in the period (t, t+r), and all flights of the associated airport B in the period (t+T ab , t+T ab +r) period and (t+T ab , t+T ab +r) all flights within the time period, where T ab represents the flight time from airport a to airport b, T ab represents the flight time from airport b to airport a, t is the start time of the time window, and r is the step size of the time window. Figure 3 As an example, s represents the time window width. Taking Airport A as the base, assume the common flight time from Airport A to Airport B is 2 hours, and the common flight time from Airport B to Airport A is 3 hours. The common time is the mode of flight times for the same takeoff and landing airports in the timetable application. In this example, if the 10:00-12:00 segment at Airport A is selected, then the 7:00-9:00 and 12:00-14:00 segments at Airport B should also be selected. Similar procedures apply to other airports.

[0119] like Figure 4 As shown in the example, the chain neighborhood operator structure is a selection set. The selection set is a set consisting of a preset number of time applications selected from the alternative time applications, using the penalty value of the time application as a weight; a certain number of times are randomly selected based on the penalty value generated by the time application as a weight, and the selected time is used as the initial time selection set. A time is selected from the alternative time applications and added to the selection set. This process is repeated until the number of times in the set reaches a certain value. The alternative time is a time application that may be replaced or interchanged with the base time application. The above selection set constitutes a chain neighborhood operator structure. R1 is a random application selected from the alternative time application.

[0120] The airport combination neighborhood operator structure uses the combination of different airports as a division standard and applies all time slots related to the combined airports as the airport combination neighborhood operator structure.

[0121] Step 3: Based on the roulette principle, a target neighborhood operator structure is randomly selected from the pre-constructed time window neighborhood operator structure, chain neighborhood operator structure, and airport combination neighborhood operator structure, and the initial solution is optimized based on the target neighborhood operator structure to obtain an intermediate solution; then, a large neighborhood algorithm is introduced to iteratively optimize the intermediate solution to obtain an approximate integer solution.

[0122] Combine Figure 5 As shown in FIG. 1 , as a specific implementation, the approximate integer solution is obtained by the following method:

[0123] Step 3-1: Use the moment application in the selected target neighborhood operator structure as a variable, and fix the other moment applications that are not in the target neighborhood operator structure.

[0124] Step 3-2: Iterate and solve based on the branch-and-bound algorithm to obtain the optimal linear upper bound and the corresponding approximate integer solution.

[0125] The branch and bound algorithm is used to solve the main problem, which is more conducive to solving the complex problem of time configuration in this embodiment.

[0126] Taking into account the convergence problem of the solution during the iterative calculation of the large neighborhood algorithm, the following process is also included to perform adaptive adjustment of the algorithm:

[0127] Step 3-3: If the preset number of iterative calculations has not been reached, but the preset iterative calculation time has been reached, and the output solution is not optimal and the solution gap is greater than the preset gap, the neighborhood is reduced based on the preset neighborhood reduction strategy and iterative optimization is performed again.

[0128] In specific implementation, the neighborhood reduction strategy is: set the weight τ of the selected application i at the time iThe initial value of the reduction parameter ρ is set to 1, and it is stipulated that if the solution cannot be found within the specified time and the solution gap is greater than the preset gap, the reduction parameter ρ'=ρ×α is modified to reduce the selected weight τ i '=τ i ×α, where α is the preset reduction coefficient and lowering coefficient.

[0129] If the output solution does not change after reaching the preset number of iterative calculations or the solution gap is not greater than the preset gap, and the optimal approximate integer solution cannot be solved using the output solution as the approximate integer solution and the solution gap is greater than the preset gap, the neighborhood is reduced based on the preset neighborhood reduction strategy and the iterative optimization is performed again.

[0130] In this embodiment, the preset gap is 1%, and the reduction parameter and the reduction coefficient are both 0.8.

[0131] Step 3-4: To select a more appropriate target neighborhood operator structure, improve the target neighborhood operator structure in the following ways:

[0132] If the accuracy of the approximate integer solution is equal to the accuracy of the intermediate solution, reducing the weight of the target neighborhood operator structure;

[0133] If the accuracy of the approximate integer solution is higher than the accuracy of the intermediate solution, but the difference between the two is less than the target difference, the weight of the target neighborhood operator structure remains unchanged;

[0134] If the accuracy of the approximate integer solution is higher than the accuracy of the intermediate solution, and the difference between the two is not less than the target difference, the weight of the target neighborhood operator structure is increased.

[0135] In specific implementation, when increasing or decreasing the weight, it is sufficient to multiply the original weight by a fixed increase coefficient or decrease coefficient.

[0136] In step 4, the approximate integer solution obtained in step 3 is used as the target solution of the time configuration model, and then the flight configuration result and flight plan are obtained.

[0137] In summary, the method described in this embodiment establishes multiple constraints within the time configuration model and combines three operator structures with a large neighborhood algorithm during the solution process, making the resulting configuration effective for multi-airport networks. Furthermore, the introduction of a non-allocation option makes the solution more realistic. It also comprehensively considers the advantages of the high-dimensional solution space for planning, addressing the significant complexity inherent in the optimization solution.

[0138] The above program can be executed in a processor or stored in a memory (or computer-readable storage medium). Computer-readable media include permanent and non-permanent, removable and non-removable media that can implement information storage by any method or technology. Information can be computer-readable instructions, data structures, program modules, or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disc read-only memory (CD-ROM), digital versatile disc (DVD) or other optical storage, magnetic cassettes, magnetic tape, disk storage or other magnetic storage devices, or any other non-transmission medium that can be used to store information that can be accessed by a computing device. As defined herein, computer-readable media does not include temporary computer-readable media such as modulated data signals and carrier waves.

[0139] These computer programs can also be loaded onto a computer or other programmable data processing device so that a series of operational steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for executing on the computer or other programmable device to implement the process. Figure 1 a process or multiple processes and / or boxes Figure 1 The steps of the functions specified in one or more blocks can be implemented by different modules corresponding to different steps.

[0140] Based on this, this embodiment also provides an airport schedule configuration system based on large neighborhood search. Figure 6 As shown, the system includes:

[0141] The model building module is used to build a time configuration model based on the flight time matrix, the time application data table and the airport capacity information; wherein the time configuration model includes an objective function and a number of constraints.

[0142] The objective function includes the maximum time offset Total time offset Dp max and unallocated flight frequencies∑ i∈R ∑ d∈D M id N i ; Where i is the index of the moment application, d is the date index, For the time of application i, the positive offset is is the negative offset of the moment of application i, M id Indicates whether the application i is running on day d; Dp max express or The maximum value in N i Indicates whether the application i at the moment is allocated.

[0143] The constraints include capacity constraints, turnaround constraints and priority constraints; among them, the capacity constraints include: airport take-off constraints, airport landing constraints and airport total take-off and landing constraints; the turnaround constraints include: maximum turnaround time constraints for connecting flights and minimum turnaround time constraints for connecting flights; the priority constraints use time offsets to constrain time applications of different priorities.

[0144] The initial solution acquisition module is used to sort each time application based on the preset influencing factor priority, and insert each time application into the pre-built empty time allocation plan in sequence according to the sorting result to obtain the initial solution of the time configuration model.

[0145] The initial solution includes a set of allocated flights obtained by targeting the insertion location with the minimum insertion penalty, and a set of unallocated flights consisting of time slot requests that cannot be inserted. The penalty value is determined by the number of capacity conflicts, the total offset, and the number of connecting flight violations.

[0146] An approximate integer solution acquisition module is used to randomly select a target neighborhood operator structure from pre-constructed time window neighborhood operator structures, chain neighborhood operator structures, and airport combination neighborhood operator structures based on the roulette principle, and optimize the initial solution based on the target neighborhood operator structure to obtain an intermediate solution; then, a large neighborhood algorithm is introduced to iteratively optimize the intermediate solution to obtain an approximate integer solution.

[0147] The time window neighborhood operator structure includes all flights of the target airport A in the period (t, t+r), and all flights of the associated airport B in the period (t+T ab , t+T ab +r) period and (t+T ab , t+T ab +r) all flights within the time period, where T ab represents the flight time from airport a to airport b, T ab It represents the flight time from airport b to airport a, t is the start time of the time window, and r is the step size of the time window.

[0148] The chained neighborhood operator structure is a selection set, wherein the selection set is a set consisting of a preset number of time applications selected from the candidate time applications using the penalty value of the time application as a weight.

[0149] The airport combination neighborhood operator structure is composed of all time slot applications involving the combined airports.

[0150] The configuration solution acquisition module is used to use the approximate integer solution as the target solution of the time configuration model, and then obtain the flight configuration result and flight plan.

[0151] The system is used to implement the steps of the above method.

[0152] The approximate integer solution acquisition module further includes:

[0153] A weight optimization unit is configured to reduce the weight of the target neighborhood operator structure when the accuracy of the approximate integer solution is equal to the accuracy of the intermediate solution; maintain the weight of the target neighborhood operator structure unchanged when the accuracy of the approximate integer solution is higher than the accuracy of the intermediate solution but the difference between the two is less than a target difference; and increase the weight of the target neighborhood operator structure when the accuracy of the approximate integer solution is higher than the accuracy of the intermediate solution and the difference between the two is not less than the target difference.

[0154] This embodiment also provides an electronic device. Figure 7 As shown, it includes at least one processor, the processor is coupled to a memory, the memory stores a computer program, and the computer program is configured to execute the above method when executed by the processor.

[0155] This embodiment further provides a computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, the above method is implemented.

[0156] Since the system, electronic equipment and storage medium are all built based on the method and used to execute the method, the efficiency of the time configuration plan is greatly improved in actual application, while ensuring the continuity of the flight plan and protecting the interests of the airline as much as possible.

[0157] The present invention provides a method and concept for airport schedule configuration based on a large neighborhood search. There are numerous methods and approaches for implementing this technical solution. The foregoing description is merely a preferred embodiment of the present invention. It should be noted that those skilled in the art may make various improvements and modifications without departing from the principles of the present invention, and such improvements and modifications are also within the scope of protection of the present invention. Any components not specified in this embodiment may be implemented using existing technologies.

Claims

1. A method for configuring airport schedules based on large neighborhood search, characterized in that: The steps include: Step 1: Build a schedule configuration model based on flight operation data; Step 2: Sort each time slot application based on the preset priority of the influencing factors, and insert each time slot application into the pre-built empty time slot allocation plan in sequence according to the sorting results, so as to obtain the initial solution of the time slot configuration model obtained in step 1; Step 3: Based on the target neighborhood operator structure, the initial solution obtained in step 2 is optimized to obtain an intermediate solution. The intermediate solution is iteratively optimized by introducing a large neighborhood algorithm to obtain an approximate integer solution. Step 4: The approximate integer solution obtained in step 3 is used as the target solution of the time configuration model, and the flight configuration result and flight plan are obtained through the target solution.

2. The airport schedule configuration method based on large neighborhood search according to claim 1, characterized in that: The flight operation data in step 1 includes the flight time matrix, the time application data table, and the airport capacity information. Step 1 specifically includes: Step 1-1: Collect the time slot applications submitted by each airline and extract the flight time matrix based on the time slot applications; the time slot applications include the operating airline, take-off and landing airports, requested take-off time, requested landing time, application cycle, operating week, requested flight type, operating aircraft type, and connecting flights; Step 1-2: Construct a slot application matrix based on the slot application date and time, and construct the slot application constraint information based on the slot application priority type and connecting flights; Step 1-3: Collect airport capacity information, including 60-minute takeoff and landing capacity and 15-minute takeoff and landing capacity. The 60-minute takeoff and landing capacity and the 15-minute takeoff and landing capacity include runway capacity and apron capacity, respectively. Steps 1-4: Construct a time configuration model based on the collected data. The time configuration model is constructed using an operations research solver and includes two parts: an objective function and constraints. The objective function includes, but is not limited to, maximum time offset, total time offset, fairness index, number of unallocated flights, and their combinations. The constraints include, but are not limited to, airport take-off and landing flights, aircraft turnover constraints, etc.

3. The airport schedule configuration method based on large neighborhood search according to claim 2, characterized in that: The objective function in step 1 includes the maximum time offset Total time offset DP max and unallocated flight frequencies∑ i∈R ∑ d∈D M id N i ; Among them, i is the index of the moment application, d is the date index, For the time of application i, the positive offset is is the negative offset of the moment of application i, M id Indicates whether the application i is running on day d; Dp max express or The maximum value in N i Indicates whether the application i at the moment is allocated; The constraints include capacity constraints, turnaround constraints and priority constraints; the capacity constraints include airport takeoff constraints, airport landing constraints and airport total takeoff and landing constraints; the turnaround constraints include maximum turnaround time constraints for connecting flights and minimum turnaround time constraints for connecting flights; the priority constraints use time offsets to constrain time applications of different priorities.

4. The airport schedule configuration method based on large neighborhood search according to claim 3 is characterized in that: In the capacity constraint, The airport takeoff frequency constraint is: The airport landing constraints are: The total take-off and landing constraints of the airport are: Where a is the index of the airport, c is the time period, d is the date index, and i is the index of the time application; F i The flight time requested for moment i, is the set of departure flights from airport a, L C The capacity time resolution is generally 1h or 15min. is the departure capacity of airport a during period c on day d, is the arrival capacity of airport a during period c on day d, is the total capacity of airport a during period c on day d; dep and tol represent the departure capacity and total capacity respectively; B it is the allocation time, which is 0 when the time application i is allocated is earlier than time t, otherwise it is 1; M id Indicates whether application i is running on day d. If application i is running on day d, it is 1, otherwise it is 0; N i Indicates whether the application i at the moment is allocated, if it is allocated, it is 0, if it is not allocated, it is 1; In the turnover constraint, The maximum turnaround time constraint for the connecting flight is: The minimum turnaround time constraint for the connecting flight is: Among them, T max and T min are the maximum turnaround time and minimum turnaround time of connecting flights, respectively. ω is a maximum value used to remove the turnaround restrictions on unassigned flights. i is the earliest allocation time for application i, LR j is the earliest allocation time for applying for time j, UR i The latest allocation time for application i, UR j The latest allocation time for applying for moment j; The priority constraints are: in, Among them A it is the application time of time application i, which is related to the specific time t. If t is earlier than the application time, then A it is 1, otherwise it is 0; H it is the historical running time of application i, which is related to the specific time t. If t is earlier than the application time, then H it is 1, otherwise it is 0; R hc A collection of all moment requests for historical moment changes.

5. The airport schedule configuration method based on large neighborhood search according to claim 4 is characterized in that: The priority in step 2 specifically includes the application priority type, the slot application date, and the connecting flight. The slot application priority type has a higher priority than the slot application date, and the slot application date has a higher priority than the connecting flight. The initial solution includes a set of allocated flights obtained by taking the position with the minimum insertion penalty as the target insertion position, and a set of unallocated flights consisting of time slot requests that cannot be inserted; the penalty value is determined by the number of capacity conflicts, the total offset, and the number of connecting flight violations.

6. The airport schedule configuration method based on large neighborhood search according to claim 5, characterized in that: Step 3 specifically includes: Step 3-1: Divide all the time applications into two categories. One category is the time applications selected in the target neighborhood operator structure. The configuration variables B corresponding to this part of the application it No changes are made, and the remaining variables that are not in the target neighborhood operator structure are applied for the corresponding variables B at other times it Do fixed processing, that is, set B it The upper and lower bounds are both B in the last iteration result it to ensure that the values of these variables remain unchanged during the solution process; Step 3-2: Iterate and solve based on the branch-and-bound algorithm to obtain the optimal linear upper bound and the corresponding approximate integer solution; Step 3-3: If the preset number of iterations has not been reached but the preset iteration duration has been reached, and the output solution is not optimal and the solution gap is greater than the preset gap, the neighborhood is reduced based on the preset neighborhood reduction strategy and iterative optimization is performed again; Step 3-4: Select a more suitable target neighborhood operator structure by improving the target neighborhood operator structure.

7. The airport schedule configuration method based on large neighborhood search according to claim 6, characterized in that: The specific neighborhood reduction strategy described in step 3-3 is: set the initial value of the weight τ selected by the application i at time i to be 1, set the initial value of the reduction parameter ρ to be 1, and stipulate that if the solution cannot be found within the specified time and the solution gap is greater than the preset gap, modify the reduction parameter ρ'=ρ×α and reduce the selected weight τ i '=τ i ×α, where α is the preset reduction coefficient and lowering coefficient; If the accuracy of the output solution is higher than that of the previous iterative solution after reaching the preset number of iterative calculations, the output solution is used as the approximate integer solution. If the accuracy of the output solution is lower than that of the previous iterative solution, the neighborhood is reduced based on the preset neighborhood reduction strategy and then the iterative optimization is performed again. The comparison of accuracy in the method depends on the setting of the objective function value. If the objective function value is maximized, the larger the objective function value of the solution, the more accurate it is. Conversely, when the objective function value is minimized, the smaller the objective function value of the solution, the more accurate it is.

8. The airport schedule configuration method based on large neighborhood search according to claim 7, characterized in that: In step 3, the method for selecting the target neighborhood operator structure is to randomly select it from pre-constructed target neighborhood operator structures based on the roulette principle. The pre-constructed target neighborhood operator includes a time window neighborhood operator structure, a chain neighborhood operator structure, and an airport combination neighborhood operator structure. The time window neighborhood operator structure includes all flights of the target airport A in the period (t, t+r), and all flights of the associated airport B in the period (t+T ab , t+T ab +r) period and (t+T ab , t+T ab +r) all flights within the time period, where T ab represents the flight time from airport a to airport b, T ab represents the flight time from airport b to airport a, t is the start time of the time window, and r is the step size of the time window; The chained neighborhood operator structure is a selection set, wherein the selection set is a set consisting of a preset number of time applications selected from the candidate time applications, using the penalty value of the time application as a weight; The airport combination neighborhood operator structure is composed of all time slot applications involving the combined airports.

9. The airport schedule configuration method based on large neighborhood search according to claim 8, characterized in that: The specific adjustment method for improving the target neighborhood operator structure in steps 3-4 is: When the accuracy of the approximate integer solution is equal to the accuracy of the intermediate solution, reducing the weight of the target neighborhood operator structure; When the accuracy of the approximate integer solution is higher than the accuracy of the intermediate solution, but the difference between the two is less than the target difference, the weight of the target neighborhood operator structure is kept unchanged; When the accuracy of the approximate integer solution is higher than the accuracy of the intermediate solution, and the difference between the two is not less than a target difference, the weight of the target neighborhood operator structure is increased.

10. An airport schedule configuration system based on large neighborhood search, characterized in that: Using any of the airport schedule configuration methods based on large neighborhood search in claims 1 to 9, the system includes a model building module, an initial solution acquisition module, an approximate integer solution acquisition module, and a configuration scheme acquisition module; The model building module builds a time configuration model based on flight operation data, wherein the flight operation data includes a flight time matrix, a time application data table, and airport capacity information, and the time configuration model includes an objective function and constraints; The initial solution acquisition module sorts each time application based on a preset priority of influencing factors, and inserts each time application into a pre-built empty time allocation scheme in sequence according to the sorting result to obtain an initial solution of the time configuration model; The approximate integer solution acquisition module randomly selects a target neighborhood operator structure from a pre-constructed time window neighborhood operator structure, a chain neighborhood operator structure, and an airport combination neighborhood operator structure based on the roulette principle, optimizes the initial solution based on the target neighborhood operator structure to obtain an intermediate solution, and iteratively optimizes the intermediate solution using a large neighborhood algorithm to obtain an approximate integer solution; The configuration solution acquisition module uses the approximate integer solution as the target solution of the time configuration model, and then obtains the flight configuration result and flight plan.

Images