An intelligent flight scheduling method based on transport capacity external adjustment

Abstract

The application relates to the field of electric digital processing, and discloses a flight intelligent scheduling method based on transport capacity external adjustment, which comprises the following steps: S1, constructing a network flow model; S2, assigning three labels of an execution arc of a flight string; S3-S5, constructing a first sub-network model; S6, constructing a second sub-network model; S7, taking the second label of the execution arc as the cost of the execution arc, and calculating the minimum cost path of the second sub-network model by using a Dijkstra algorithm; S8, judging whether the network flow reaches the maximum flow of the network, if yes, entering step S10, otherwise, entering step S9; S9, updating the maximum allowable delay time and entering step S6; S10, exporting the minimum maximum allowable delay time as the current maximum allowable delay time, and exporting an execution arc set meeting the condition as a flight optimization scheduling result. The application reduces the time and operation cost of an airline company due to aircraft exchange from the root.

Description

Technical Field

[0001] This invention relates to the field of electronic digital processing, and in particular to a method for intelligent flight scheduling based on external capacity allocation. Background Technology

[0002] Civil aviation, characterized by its rapid response capabilities, is a vital force for national defense transportation and economic security, and a crucial component of emergency response. When airlines receive requests for capacity redeployment, it reduces their capacity. Without adjustments, this can lead to numerous flight cancellations, causing difficulties for subsequent operations and impacting passenger travel.

[0003] After capacity is redeployed, due to the suddenness and uncertainty of the aircraft's deployment, operations control personnel typically cancel all flights involving the redeployed aircraft and resume flight schedules once the redeployed aircraft has completed its assigned tasks. While this adjustment allows for a quick response, directly cancelling flights, though simple, is detrimental to passenger travel. Therefore, minimizing flight cancellations is a critical issue that urgently needs to be addressed.

[0004] However, most existing technologies focus on optimizing the scheduling of irregular flights due to airport closures or restrictions, or insufficient capacity caused by weather conditions, and have not solved the problem of transferring capacity to other regions. Summary of the Invention

[0005] The purpose of this invention is to overcome the shortcomings of the prior art and provide a flight intelligent scheduling method based on external capacity allocation.

[0006] The objective of this invention is achieved through the following technical solution:

[0007] A first aspect of the present invention provides a method for intelligent flight scheduling based on external capacity allocation, characterized by comprising the following steps:

[0008] S1: Construct a network flow model in Python. The network flow model includes a source node, an execution node, an end node, an aircraft arc, and an execution arc. The source node represents the start of optimized scheduling, the execution node represents the aircraft starting to execute its corresponding task, and the end node represents the end of optimized scheduling. The arc from the source node to the execution node is the aircraft arc, representing the aircraft... The flight mission or emergency support mission is executed starting from the source node; each arc from the execution node to the end node corresponds to an aircraft. Execute flight series The execution arc; where K represents the set of aircraft. R represents the set of flight strings;

[0009] S2: Set all aircraft arc flow rates to 1; calculate aircraft... Execute flight series The delay time relative to the original scheduled flight plan is assigned to the corresponding value in the flight sequence. The first label of the execution arc In the middle; the flight will be determined to be contaminated. Is it the corresponding aircraft in the original scheduled flight plan? The result of the judgment is assigned to the flight string. The second label of the execution arc In the middle; the third label of the execution arc of the flight sequence Setting it to 1 represents a flight string. Only one plane can be Execute once;

[0010] S3: Introduce maximum permissible delay time and minimum allowable delay time Initialize according to the problem being solved Let be the maximum value of the execution arc delay time of the current network flow model. ;

[0011] S4: Initialize network traffic ;

[0012] S5: Remove the first label of the execution arc from the original network flow model. Not satisfied The execution arc remains unchanged, while other network attributes and content are preserved. Based on this, the first sub-network model is constructed.

[0013] S6: Remove the first label of the execution arc from the first sub-network model generated in step S5. Not satisfied The execution arc remains unchanged, while other network attributes and content are preserved. A second sub-network model is then constructed based on this.

[0014] S7: The second label of the execution arc To calculate the cost of the execution arc, Dijkstra's algorithm is used to compute the minimum cost path of the second sub-network model. The minimum cost path of the second sub-network model is solved using a mixed-integer linear programming model based on the optimization scheduling objective. This mixed-integer linear programming model is modeled with the objective function of minimizing the maximum allowable delay time and minimizing the costs of flight execution and delays. The objective function is as follows:

[0015] (1);

[0016] (2);

[0017] (3);

[0018] (4);

[0019] In equation (1), Indicates when the airplane Execute flight series The delay time is represented by Equation (1), which means minimizing the allowable delay time of the newly generated flight plan relative to the corresponding flight string in the original flight plan.

[0020] In equation (2), C represents the total operating cost. Indicates airplane Flights Operating costs Indicates airplane Whether to operate flight sequence , This indicates the cost for an airline to operate a flight. This represents the unit delay cost of a flight series. This indicates the corresponding flight sequence of the new flight plan relative to the original flight plan. Equation (2) represents the total operating cost obtained by minimizing the cost of the airline executing flights and adding the cost of delays caused by delays in the new flight schedule relative to the corresponding flight string of the original flight schedule;

[0021] In equation (3), Indicates flight sequence The travel time Indicates airplane Unit fuel consumption cost Indicates flight sequence Airport service costs;

[0022] In equation (4), the aircraft Execute flight series Delay time airplane earliest available time Subtract the corresponding flight sequence from the original flight plan The first flight Original departure time Perform calculations, where F represents the flight set; at the same time, in equation (4), since the aircraft that do not need to respond to scheduling are available all day, in order to avoid calculation errors of negative values, the non-negative processing is carried out in the way of equation (4);

[0023] Meanwhile, the objective function satisfies the following constraints:

[0024] (5);

[0025] (6);

[0026] (7);

[0027] The constraints in equation (5) represent the constraints for each aircraft. Only one flight sequence can be executed. The constraints in equation (6) represent the flight sequences within each original scheduled flight plan. All must be executed; the constraints in equation (7) represent the aircraft and flight connections The execution relationship between them, when Aircraft executing flight sequence hour It is 1 if it is true, otherwise it is 0.

[0028] S8: Determine network traffic Has the network's maximum traffic been reached? If the condition is met, proceed to step S10; otherwise, proceed to step S9.

[0029] S9: Update Proceed to step S6;

[0030] S10: Export current for The set of execution arcs that meet the conditions is exported as the result of flight optimization scheduling.

[0031] Furthermore, the method invokes the large-scale mathematical optimization solver Gurobi for computation.

[0032] The beneficial effects of this invention are:

[0033] In an exemplary embodiment of the present invention, a binary search algorithm is first used to calculate the minimum maximum allowable delay time. Then, the minimum maximum allowable delay time obtained in the previous step is used as a constraint to select execution arcs that meet the conditions to construct a weighted directed graph. The Dijkstra algorithm is combined to minimize the number of aircraft exchanges. Finally, the minimum cost recovery scheme is calculated.

[0034] Minimize the maximum allowable delay time Simultaneously, a weighted directed graph is constructed to determine whether an aircraft exchange has occurred using arc labels. For path costs, Dijkstra's algorithm is used to calculate the minimum-cost path, minimizing aircraft swaps during recovery. This exemplary embodiment achieves the recovery objective while reducing the time and operating costs incurred by airlines due to aircraft swaps at the source. Attached Figure Description

[0035] Figure 1 This is a schematic diagram of airlines' flight operations under the external transfer of capacity;

[0036] Figure 2 A diagram illustrating flight exchange and delay times;

[0037] Figure 3 The flowchart illustrates an exemplary embodiment of the present invention of a flight intelligent scheduling method based on external capacity allocation;

[0038] Figure 4 This is a schematic diagram of an iterative process for reducing the maximum allowable delay time of a system through binary search, provided in an exemplary embodiment of the present invention.

[0039] Figure 5 This is a schematic diagram comparing the delay times of different solutions provided in an exemplary embodiment of the present invention. Detailed Implementation

[0040] The technical solution of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0041] Furthermore, the technical features involved in the different embodiments of the present invention described below can be combined with each other as long as they do not conflict with each other.

[0042] First, it should be noted that, in order to reduce the decrease in flight execution rate caused by flight cancellations due to capacity relocation, the exemplary embodiments described below adopt the method of aircraft swapping and flight delays. Through scheduling optimization, the losses of airlines are minimized while reducing flight cancellations. Based on the above objectives, the model constructed in the exemplary embodiments described below follows the following assumptions: (1) It is assumed that the airport closes at 2:00 AM and opens at 6:00 AM. (2) Only flight recovery on the day of capacity relocation is considered. (3) Only aircraft recovery is considered. (4) Aircraft maintenance issues are not considered.

[0043] With capacity being redeployed, how do airlines operate their flights? Figure 1 As shown. Figure 1 Each node (scheduling node, required node, task node, end node) represents an airport, and each arc (scheduling arc, execution arc, return arc) represents the flight activities of the aircraft.

[0044] At the dispatch node, the airline receives an emergency capacity redeployment order from its superiors, designating a portion of its aircraft to depart from the dispatch arc and proceed to the mission node to participate in emergency support tasks. The remaining aircraft continue their original scheduled flight plans along the execution arc. The mandatory nodes represent the airports involved in the original scheduled flight plan (mandatory nodes). Due to the redeployment of some aircraft for emergency tasks, available capacity is reduced, making it impossible to fully execute flight missions according to the original scheduled flight plan, resulting in a capacity shortage. To improve flight execution efficiency, the original scheduled flight plan must be readjusted and optimized. The return arc indicates that after completing the emergency support mission, the aircraft responding to the capacity redeployment return to their original flight plan to continue their flight missions or return to their departure airport (end node). The end node marks the end of the flight recovery process, and the system returns to a stable state.

[0045] This process involves not only dynamic adjustments to flight scheduling, but also comprehensive consideration of the optimal allocation of aircraft resources and the efficiency of mission execution, achieving a balance between emergency support missions and the original scheduled flight plans.

[0046] The recovery plan that delays the execution time of the scheduled flight sequence to the earliest available time of the aircraft in the original flight schedule may be the recovery plan that simultaneously minimizes the airline's losses and the total delay time of all flights. However, the delay time of the scheduled aircraft will be extended indefinitely with the earliest available time of the aircraft. Excessive delay time will lead to a significant decrease in passenger satisfaction and may even prevent the flight sequence from being completed.

[0047] In the following exemplary embodiments, the maximum permissible delay time is reduced through aircraft switching. Maximum allowable delay time This refers to the maximum allowable delay time while ensuring the feasibility of flight schedules and service quality. This is achieved by reducing the maximum permissible delay time. The goal is to obtain the recovery plan with the shortest delay time while smoothing out the delay time to avoid any single flight experiencing a long delay. For example... Figure 2 As shown, before the swap, both flight sequences A and B experienced delays because their earliest available times were later than the departure time of the first flight in their respective flight sequences, with flight sequence B experiencing a longer delay. After the aircraft swap, although both flight sequences were still delayed, the delay time was smoothly distributed between them, effectively solving the problem of excessively long delays in flight sequence B before the swap. Simultaneously, a recovery plan that minimizes airline recovery costs and meets recovery objectives was generated based on the modeling.

[0048] See Figure 3 , Figure 3 The flowchart illustrates an exemplary embodiment of the present invention, which describes a flight intelligent scheduling method based on external capacity allocation, comprising the following steps:

[0049] S1: Construct a network flow model in Python. The network flow model includes a source node, an execution node, an end node, an aircraft arc, and an execution arc. The source node represents the start of optimized scheduling, the execution node represents the aircraft starting to execute its corresponding task, and the end node represents the end of optimized scheduling. The arc from the source node to the execution node is the aircraft arc, representing the aircraft... The flight mission or emergency support mission is executed starting from the source node; each arc from the execution node to the end node corresponds to an aircraft. Execute flight series The execution arc (including the area within each execution node, and the path from the last execution node to the end node); where K represents the set of aircraft. R represents the set of flight strings;

[0050] S2: Set all aircraft arc flow rates to 1; calculate aircraft... Execute flight series The delay time relative to the original scheduled flight plan is assigned to the corresponding value in the flight sequence. The first label of the execution arc In the middle; the flight will be determined to be contaminated. Is it the corresponding aircraft in the original scheduled flight plan? The result of the judgment is assigned to the flight string. The second label of the execution arc In the middle; the third label of the execution arc of the flight sequence Setting it to 1 represents a flight string. Only one plane can be Execute once.

[0051] Specifically, in steps S1 and S2, in order to solve the model, a network flow model is constructed in the Python environment, and source nodes, execution nodes, and end nodes are set.

[0052] The arc from the source node to the execution node is the airplane arc, representing the airplane. Flight missions or emergency support missions are executed starting from the scheduling node, with each flight arc assigned a flow rate of 1. Each arc from the execution node to the end node corresponds to a specific aircraft. Execute a flight The execution arcs are defined, and each arc has an arc weight. Each arc is assigned three labels: the first label... Second tag Third tag .

[0053] To be an airplane Execute flight series At that time, the flight schedule Compared to the original scheduled flight schedule, the flight schedule was more flexible. The delay time (the calculation rules are shown in the following formula (5)). Is it to determine flight sequence Is it the corresponding aircraft in the original scheduled flight plan? Execution. If the flight plan is revised, the aircraft... The same flight schedule will still be executed. Then record for Otherwise, record it as 1. Let be the capacity of the arc, and let This indicates that each arc can only hold one unit of capacity, meaning that a flight sequence can only be executed once by one aircraft. It should be noted that this third label... This is a constraint that restricts a flight sequence to be run by only one aircraft to prevent the problem of a flight sequence being run by multiple aircraft simultaneously. When constructing a network flow model, an arc can only be matched with one aircraft.

[0054] In previous solutions to multi-objective programming models, different weights were often defined for different objective functions to determine their priority and influence factors in actual calculations. However, in subsequent steps S3-S10, it is difficult to define reasonable weights for the two objective functions (Equations (1) and (2)). Therefore, a hierarchical solution approach is adopted, combining the binary search algorithm and Dijkstra's algorithm to solve the multi-objective model. Specifically:

[0055] S3: Introduce maximum permissible delay time and minimum allowable delay time Initialize according to the problem being solved Let be the maximum value of the execution arc delay time of the current network flow model. ;

[0056] S4: Initialize network traffic ;

[0057] S5: Remove the first label of the execution arc from the original network flow model. Not satisfied The execution arc remains unchanged, while other network attributes and content are preserved. Based on this, the first sub-network model is constructed.

[0058] S6: Remove the first label of the execution arc from the first sub-network model generated in step S5. Not satisfied The execution arc remains unchanged, while other network attributes and content are preserved. A second sub-network model is then constructed based on this.

[0059] S7: The second label of the execution arc To calculate the cost of the execution arc, Dijkstra's algorithm is used to compute the minimum cost path for the second sub-network model. The second label of the execution arc. The definition is to determine flight sequence Is it the corresponding aircraft in the original scheduled flight plan? The label is set to 0 or 1. By assigning this label and then using Dijkstra's algorithm to calculate the minimum cost path of the second sub-network model, the number of aircraft exchanges can be minimized.

[0060] In the second sub-network model, the minimum cost path is solved using a mixed-integer linear programming model based on the optimization scheduling objective. The mixed-integer linear programming model is modeled with the objective function of minimizing the maximum allowable delay time and minimizing the cost of flight execution and delay. The objective function is as follows:

[0061] (1);

[0062] (2);

[0063] (3);

[0064] (4);

[0065] In equation (1), Indicates when the airplane Execute flight series The delay time is represented by Equation (1), which means minimizing the allowable delay time of the newly generated flight plan relative to the corresponding flight string in the original flight plan.

[0066] In equation (2), C represents the total operating cost. Indicates airplane Flights Operating costs Indicates airplane Whether to operate flight sequence , This indicates the cost for an airline to operate a flight. This represents the unit delay cost of a flight series. This indicates the corresponding flight sequence of the new flight plan relative to the original flight plan. Equation (2) represents the total operating cost obtained by minimizing the cost of the airline executing flights and adding the cost of delays caused by delays in the new flight schedule relative to the corresponding flight string of the original flight schedule;

[0067] In equation (3), Indicates flight sequence The travel time Indicates airplane Unit fuel consumption cost Indicates flight sequence Airport service costs;

[0068] In equation (4), the aircraft Execute flight series Delay time airplane earliest available time Subtract the corresponding flight sequence from the original flight plan The first flight Original departure time Perform calculations, where F represents the flight set; at the same time, in equation (4), since the aircraft that do not need to respond to scheduling are available all day, in order to avoid calculation errors of negative values, the non-negative processing is carried out in the way of equation (4);

[0069] Meanwhile, the objective function satisfies the following constraints:

[0070] (5);

[0071] (6);

[0072] (7);

[0073] The constraints in equation (5) represent the constraints for each aircraft. Only one flight sequence can be executed. The constraints in equation (6) represent the flight sequences within each original scheduled flight plan. All must be executed; the constraints in equation (7) represent the aircraft and flight connections The execution relationship between them, when Aircraft executing flight sequence hour It is 1 if it is true, otherwise it is 0.

[0074] S8: Determine network traffic Has the network's maximum traffic been reached? If the condition is met, proceed to step S10; otherwise, proceed to step S9.

[0075] S9: Update Proceed to step S6;

[0076] S10: Export current for The set of execution arcs that meet the conditions is exported as the result of flight optimization scheduling.

[0077] Specifically, the basic model is built in steps S1 to S5, including: the maximum allowable delay time in step S3. and minimum allowable delay time Initialization, network traffic in step S4 The initialization of the first sub-network model in step S5 involves checking the delay time of the arcs to prevent the occurrence of redundant arcs.

[0078] In steps S6 to S10, a cyclical process is performed to calculate the minimum-cost recovery plan. Specifically:

[0079] The purpose of step S6 is: the first step of the binary search loop iteration, which gradually approaches the possible solution by updating and narrowing the solution space and adjusting the upper limit of the delay time, which is related to the accuracy of the algorithm;

[0080] The purpose of step S7 is to use Dijkstra's algorithm to find the minimum cost path, with the aim of reducing the number of aircraft swaps and maintaining operational stability.

[0081] The purpose of step S8 is to check whether the maximum flow has been reached, i.e., whether all aircraft have been allocated, which is one of the termination conditions.

[0082] The purpose of step S9 is to update the minimum and maximum delay times. The lower bound is determined, and the iteration continues to ensure that the optimal solution is found;

[0083] The purpose of step S10 is to gradually approach the optimal solution (minimize delay time and minimize aircraft swaps) through iterative loops (S6→S9), and finally output the scheduling scheme with minimum and maximum delay time and recovery efficiency in S10.

[0084] First, the maximum allowable delay time is calculated by using a binary search algorithm in a cyclic manner. Then, in step S7, the minimum allowable delay time obtained in the previous step is used as a constraint to select execution arcs that meet the conditions to construct a weighted directed graph. The Dijkstra algorithm is then used to minimize the number of aircraft exchanges. Finally, the minimum cost recovery scheme is calculated.

[0085] In summary, in this exemplary embodiment, minimizing the maximum permissible delay time... (Thus reducing flight delays) In step S7, a weighted directed graph is constructed to determine the arc labels for whether an aircraft exchange has occurred. For path costs, Dijkstra's algorithm is used to calculate the minimum-cost path, minimizing aircraft swaps during recovery. This exemplary embodiment achieves the recovery objective while reducing the time and operating costs incurred by airlines due to aircraft swaps at the source.

[0086] It should be noted that network traffic is a current technology and will not be discussed further here.

[0087] In subsequent exemplary embodiments, it is preferable to introduce the commercial solver Gurobi to obtain the optimal solution that satisfies the recovery objectives: minimizing airline losses, minimizing delay time, and minimizing unnecessary aircraft changes.

[0088] To verify the effectiveness of the above modeling and algorithm in handling the aircraft recovery problem of airlines under capacity relocation, the flight schedule of Kunming Airlines was used for verification. The experiment was conducted using Python 2024.1.4, and the large-scale mathematical optimization solver Gurobi 11.0.3 was used to solve the model. The program ran on a computer equipped with an AMD R9 7845HX processor and 16GB of RAM.

[0089] 1. Parameter settings and data description

[0090] This exemplary embodiment involves 22 aircraft, 22 flight sequences, 77 flights, and 30 airports. The partial flight schedules involved in the optimized scheduling and the corresponding aircraft availability times are shown in Table 1. In this example, the turnaround time between any two flights within any flight sequence is greater than 45 minutes, and all flight operation times are between 6:00 AM and 2:00 AM the following day.

[0091] In Table 1, the aircraft number represents the aircraft serial number. For example, aircraft number 10 corresponds to aircraft number 10. Available time represents the time an aircraft can use to execute a flight sequence; the flight sequence number is an internal management number assigned to a "flight schedule chain" after several flights operated by the same aircraft on the same day (or several consecutive days). It is the "aircraft mission sequence number" used by airlines for scheduling and aircraft dispatching; the flight number is used to "identify a specific flight on a specific day"; the departure airport and arrival airport are the English abbreviations of the corresponding airports; the departure time and landing time represent the aircraft's estimated departure time and estimated landing time, respectively.

[0092] In this scenario, the airline receives an emergency capacity redeployment request, demanding that aircraft numbers 10 and 16 participate in emergency support. After the emergency support mission is completed... and The return times are 2:35 PM and 4:00 PM on the same day (corresponding to available time).

[0093] Table 1. Partial Flight Schedules and Corresponding Aircraft Availability

[0094]

[0095] In this exemplary embodiment, the aircraft used is a B737. The values ​​of various parameters involved in the cost calculation in the model are shown in Table 2. Due to differences in airports and flight sequences, some parameters vary. For ease of calculation, the values ​​are all average values.

[0096] Table 2 Parameter values ​​of the model

[0097]

[0098] Table 2 mainly discloses The corresponding unit delay cost of the flight sequence Corresponding to aircraft unit fuel consumption cost Corresponding to flight sequence Airport service costs. Among them, and The unit is "yuan·h" -1 That is, "15,000 yuan per hour" or "17,500 yuan per hour". The unit is "yuan per flight string" -1 "80,000 yuan per flight" means "each flight costs 80,000 yuan".

[0099] 2. Analysis of Experimental Results

[0100] To evaluate the ability of the intelligent flight scheduling method based on external capacity adjustment in the above exemplary embodiment to optimize the scheduling of airline flights under external capacity adjustment, three schemes are compared.

[0101] Option 1 is the traditional handling method for airlines receiving capacity transfers: all flights involving the transferred aircraft are cancelled, and the flight schedule is resumed after the transferred aircraft returns; Option 2 is to directly extend the flight sequence corresponding to the transferred aircraft until the transferred aircraft becomes available; Option 3 is an optimized recovery scheme based on the method of the above exemplary embodiments.

[0102] Based on the propagation of delays and the requirement of at least 45 minutes of layover time between each flight, this exemplary embodiment is solved according to the airport's opening time of 6:00 AM to 2:00 AM the following day.

[0103] The solution result for Scheme 3 is: and Perform aircraft swapping and execute corresponding flight sequences. and Perform aircraft swapping to execute the corresponding flight sequence. Scheme 3 uses an iterative process of binary search to reduce the system's maximum allowable delay time, as follows: Figure 4 As shown, Figure 4 The horizontal axis represents the number of iterations. Figure 4The vertical axis represents the delay time (in minutes). Figure 4 Showing , The changes during the iteration process, among which (Corresponding to step S9). In the initial stage (the first 5 iterations), Rising rapidly Rapidly declining, The initial fluctuations were significant; subsequently, the three curves gradually stabilized, indicating that the algorithm gradually converged, the delay time range narrowed, and it eventually approached the optimal solution. This reflects the efficiency and stability of the binary search method in the optimization process.

[0104] The solution results for each scheme are shown in Table 3:

[0105] Table 3 Solution results of the example

[0106]

[0107] Experimental results show that the number of flight cancellations in Scheme 2 and Scheme 3 is less than that in Scheme 1, with cancellation rates of 8%, 2%, and no cancellations, respectively. Scheme 2 has a 7.8% lower cancellation rate than Scheme 1. Compared to Scheme 2, which does not involve aircraft swaps, Scheme 3 has no flight cancellations. In Scheme 1, all flights involving the transfer of capacity were cancelled, resulting in indefinite delays. In Scheme 2, the total delay time was 2090 minutes. Figure 5 Based on the total delay times of the six delayed flights in Option 2, after reducing the maximum allowable delay time using the algorithm, the total delay time for Option 3 is 1055 minutes. Figure 5 The total delay time for the 8 delayed flights in Option 3 is calculated as follows: Option 3 reduces the total delay time by 49.5% compared to Option 2. The delay times for flights in Options 2 and 3 are as follows: Figure 5 As shown, the maximum allowable delay time for Option 2 is 370 minutes, and the maximum allowable delay time for Option 3 is 225 minutes. Furthermore, the delay time for each flight segment in Option 3 is smoother, and there is no situation where the delay time for any flight segment is too long.

[0108] In summary, taking this exemplary embodiment as an example, the solution generated by the method described above reduces the flight cancellation rate by 10.3% compared to traditional processing methods. This demonstrates that the constructed model can effectively improve flight execution rates when resuming flights under the influence of airline capacity reallocation.

[0109] Obviously, the above embodiments are merely illustrative examples for clear explanation and are not intended to limit the implementation. Those skilled in the art can make other variations or modifications based on the above description. It is neither necessary nor possible to exhaustively list all possible implementations. However, obvious variations or modifications derived therefrom are still within the scope of protection of this invention.

Claims

1. A method for intelligent flight scheduling based on external capacity allocation, characterized in that: Includes the following steps: S1: Construct a network flow model in Python. The network flow model includes a source node, an execution node, an end node, an aircraft arc, and an execution arc. The source node represents the start of optimized scheduling, the execution node represents the aircraft starting to execute its corresponding task, and the end node represents the end of optimized scheduling. The arc from the source node to the execution node is the aircraft arc, representing the aircraft... The flight mission or emergency support mission is executed starting from the source node; each arc from the execution node to the end node corresponds to an aircraft. Execute flight series The execution arc; where K represents the set of aircraft. R represents the set of flight strings; S2: Set all aircraft arc flow rates to 1; calculate aircraft... Execute flight series The delay time relative to the original scheduled flight plan is assigned to the corresponding value in the flight sequence. The first label of the execution arc In the middle; the flight will be determined to be contaminated. Is it the corresponding aircraft in the original scheduled flight plan? The result of the judgment is assigned to the flight string. The second label of the execution arc In the middle; the third label of the execution arc of the flight sequence Setting it to 1 represents a flight string. Only one plane can be Execute once; S3: Introduce maximum permissible delay time and minimum allowable delay time Initialize according to the problem being solved Let be the maximum value of the execution arc delay time of the current network flow model. ; S4: Initialize network traffic ; S5: Remove the first label of the execution arc from the original network flow model. Not satisfied The execution arc remains unchanged, while other network attributes and content are preserved. Based on this, the first sub-network model is constructed. S6: Remove the first label of the execution arc from the first sub-network model generated in step S5. Not satisfied The execution arc remains unchanged, while other network attributes and content are preserved. A second sub-network model is then constructed based on this. S7: The second label of the execution arc To calculate the cost of the execution arc, Dijkstra's algorithm is used to compute the minimum cost path of the second sub-network model. The minimum cost path of the second sub-network model is solved using a mixed-integer linear programming model based on the optimization scheduling objective. This mixed-integer linear programming model is modeled with the objective function of minimizing the maximum allowable delay time and minimizing the costs of flight execution and delays. The objective function is as follows: (1)； (2)； (3)； (4)； In equation (1), Indicates when the airplane Execute flight series The delay time is represented by Equation (1), which means minimizing the allowable delay time of the newly generated flight plan relative to the corresponding flight string in the original flight plan. In equation (2), C represents the total operating cost. Indicates airplane Flights Operating costs Indicates airplane Whether to operate flight sequence , This indicates the cost for an airline to operate a flight. This represents the unit delay cost of a flight series. This indicates the corresponding flight sequence of the new flight plan relative to the original flight plan. Equation (2) represents the total operating cost obtained by minimizing the cost of the airline executing flights and adding the cost of delays caused by delays in the new flight schedule relative to the corresponding flight string of the original flight schedule; In equation (3), Indicates flight sequence The travel time Indicates airplane Unit fuel consumption cost Indicates flight sequence Airport service costs; In equation (4), the aircraft Execute flight series Delay time airplane earliest available time Subtract the corresponding flight sequence from the original flight plan The first flight Original departure time Perform calculations, where F represents the flight set; at the same time, in equation (4), since the aircraft that do not need to respond to scheduling are available all day, in order to avoid calculation errors of negative values, the non-negative processing is carried out in the way of equation (4); Meanwhile, the objective function satisfies the following constraints: (5)； (6)； (7)； The constraints in equation (5) represent the constraints for each aircraft. Only one flight sequence can be executed. The constraints in equation (6) represent the flight sequences within each original scheduled flight plan. All must be executed; the constraints in equation (7) represent the aircraft and flight connections The execution relationship between them, when Aircraft executing flight sequence hour It is 1 if it is true, otherwise it is 0. S8: Determine network traffic Has the network's maximum traffic been reached? If the condition is met, proceed to step S10; otherwise, proceed to step S9. S9: Update Proceed to step S6; S10: Export current for The set of execution arcs that meet the conditions is exported as the result of flight optimization scheduling.

2. The intelligent flight scheduling method based on external capacity allocation according to claim 1, characterized in that: The method invokes the large-scale mathematical optimization solver Gurobi for computation.

Images