A flight ring scheduling method combining stopover slack time and flight delay

Abstract

The present invention relates to the field of computer technology applications, and in particular to a flight ring scheduling method that combines stopover slack time with flight delays, comprising: obtaining an initial task node set based on a pending flight task information table; obtaining an initial pending task node list based on a greedy algorithm; obtaining a target task node sequence set if the number of iterations does not reach a preset iteration threshold or the objective function value corresponding to the current node sequence set record set meets a convergence condition, and updating the current task node sequence set based on a sequence destruction method and a sequence reconstruction method to obtain an updated current task node sequence set, and use it as the current task node sequence set. The present invention enables the obtained target task node sequence set to have a minimum resource requirement value. In actual flight scheduling, it can reduce airline operating costs, take into account aircraft utilization and delays, and improve aircraft operating efficiency.

Description

Technical Field

[0001] The present invention relates to the field of computer technology applications, and in particular to a flight ring scheduling method combining stopover slack time and flight delays. Background Art

[0002] In some application scenarios, tasks that need to be performed within a set time period need to be scheduled so that they can be completed as punctually as possible with minimal resources, such as flight scheduling. Therefore, how to complete tasks as punctually as possible with minimal resources is a topic worth exploring. Summary of the Invention

[0003] In view of the above technical problems, the technical solution adopted by the present invention is:

[0004] An embodiment of the present invention provides a flight ring scheduling method that combines stopover slack time and flight delays, the method comprising the following steps:

[0005] S100, based on the flight task information table to be processed, obtaining an initial task node set; wherein one node in the initial task node set corresponds to one flight task.

[0006] S200, based on the greedy algorithm, obtain an initial task node sequence set that meets the preset constraints from the initial task node set, and delete the task nodes in the initial task node sequence set from the initial task node set to obtain an initial list of task nodes to be processed; set the iteration counter c=1.

[0007] S300, if c≤c0 or the objective function value corresponding to the current node sequence set record set meets the convergence condition, execute S600, otherwise, execute S400; the initial value of the current node sequence set record set is the initial task node sequence set; c0 is the preset number threshold; wherein, the objective function value of each task node sequence is used to characterize the resource requirement value required by the task node sequence, and the resource requirement value of each task node sequence is determined based on the stay slack time and delay time corresponding to the corresponding task node.

[0008] S400: Based on the sequence destruction method and the sequence reconstruction method, the current task node sequence set is updated to obtain an updated task node sequence set and stored in the current node sequence set record set. The initial value of the current task node sequence set is the initial task node sequence set, and S500 is executed.

[0009] S500 , based on a preset update result receiving rule, select one from the current task node sequence set and the updated task node sequence set as a new current task node sequence set; set c=c+1, and execute S300 .

[0010] S600: taking the task node sequence set with the minimum resource requirement value in the current node sequence set record set as the target task node sequence set, and exiting the current control program.

[0011] The present invention has at least the following beneficial effects:

[0012] An embodiment of the present invention provides a flight loop scheduling method that combines stopover slack time and flight delays. During the acquisition of a task node sequence, the goal is to minimize the resource requirement value determined based on the stopover slack time and delay time corresponding to the task node. At least one sequence destruction method and a sequence reconstruction method are used to update the sequence set, thereby ensuring that the obtained target task node sequence set has the minimum resource requirement value. In actual flight scheduling, this method can reduce airline operating costs, balance aircraft utilization and delays, and improve aircraft operating efficiency.

[0013] It should be understood that the content described in this section is not intended to identify the key or important features of the embodiments of the present invention, nor is it intended to limit the scope of the present invention. Other features of the present invention will become readily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0015] Figure 1 A flowchart of a flight ring scheduling method combining stopover slack time and flight delays provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0016] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without making any creative efforts shall fall within the scope of protection of the present invention.

[0017] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the art to which this invention pertains. The terms used herein in the specification of the present invention are for the purpose of describing specific embodiments only and are not intended to limit the present invention. The term "and / or" as used herein includes any and all combinations of one or more of the associated listed items.

[0018] It should be noted that some exemplary embodiments are described as processes or methods depicted as flowcharts. Although the flowcharts describe the steps as sequential processes, many of the steps can be performed in parallel, concurrently, or simultaneously. In addition, the order of the steps can be rearranged. A process can be terminated when its operation is completed, but can also have additional steps not included in the accompanying drawings. A process can correspond to a method, function, procedure, subroutine, subprogram, etc.

[0019] First, the concepts used in the embodiments of the present invention are described.

[0020] In this embodiment of the present invention, the minimum turnaround time refers to the shortest interval between an aircraft landing on its previous flight and departing on its next flight. This time is primarily determined by the time it takes for passengers to disembark and embark, the time required for cabin cleaning, and ground support operations (including refueling, baggage and cargo loading and unloading).

[0021] When the actual connecting time between two flights exceeds the minimum stopover time, the remaining time is called stopover slack time. In actual operations, this time plays a crucial role in absorbing delays and preventing them from spreading across flights. Different flight scheduling schemes generate varying amounts of slack time. If too much slack time is reserved, aircraft will remain at the airport for extended periods, reducing aircraft utilization and impacting the airline's operational profitability. However, if too little slack time is reserved, it can exacerbate delay propagation and impact flight operations. Therefore, properly setting stopover slack time is crucial for airlines' operational efficiency and cost control.

[0022] Flight delay is an important factor affecting flight scheduling. In short-term flight scheduling, flight delay can be divided into independent delay and affected delay. Among them, independent delay refers to the delay of a flight due to its own reasons. This delay is usually unrelated to other flights and may be caused by technical failure, crew problems or other internal factors. Impact delay refers to the delay of a flight that further affects the punctuality of subsequent flights. This delay usually occurs in the flight network. The delay of the previous flight will cause the subsequent flights to be unable to take off as planned, thus causing a chain reaction. The actual flight delay is composed of the sum of the independent delay of the flight itself and the affected delay of the previous flight. The calculation of flight delay refers to the existing literature, such as the flight delay calculation method disclosed in literature 1 (Zhang Haifeng, Hu Minghua; Airline Short-term Flight Scheduling Model and Algorithm [J]; Journal of Nanjing University of Aeronautics and Astronautics, 2015, 47(04): 553-558). Among them, independent delays are approximately independent of each other and normally distributed. In actual application, if an aircraft fails to arrive at the airport on time due to an independent delay, and the delay time cannot be fully absorbed by the stopover slack time, the excess time will be regarded as the affected delay time, which will affect the operation of subsequent flights.

[0023] The embodiment of the present invention provides a flight ring scheduling method that combines stopover slack time and flight delay, such as Figure 1 As shown, the method includes the following steps:

[0024] S100: Obtain an initial task node set based on the flight task information table to be processed.

[0025] In an embodiment of the present invention, a flight mission information table may be an information table requiring flight scheduling, and may be a flight mission information table for a specific day of a specific destination airline. The flight mission information table may include multiple flight mission information, each of which includes a corresponding flight number, departure airport, planned departure time, arrival airport, planned arrival time, passenger volume, and independent delay distribution.

[0026] Among them, one node in the initial task node set corresponds to one flight task, that is, the number of nodes in the initial task node set is equal to the number of flight tasks in the flight task information table to be processed, and each node in the task node set can be represented by the flight number of the corresponding flight task.

[0027] S200, based on the greedy algorithm, obtains an initial task node sequence set that meets preset constraints from the initial task node set, and deletes the task nodes in the initial task node sequence set from the initial task node set to obtain an initial list of task nodes to be processed.

[0028] In an embodiment of the present invention, the preset constraint conditions may include the following conditions:

[0029] Constraint 1: Each executor can only execute one mission node sequence, that is, each aircraft can only fly one flight route.

[0030] Constraint 2: The departure point corresponding to the first task node in the task node sequence and the destination corresponding to the last task node must be the same, that is, the departure airport of the first flight and the arrival airport of the last flight must be the same.

[0031] Constraint 3: The time interval between two adjacent task nodes in the task node sequence must be greater than or equal to the minimum dwell time, or minimum transit time. This means the time interval between adjacent flights must meet the airport's minimum transit time requirement. The minimum dwell time can be determined based on the actual conditions of the airport.

[0032] Constraint 4: The number of executors corresponding to all task node sequences cannot exceed the number of provided executors, that is, the number of aircraft used in flight scheduling cannot exceed the number of provided aircraft.

[0033] In the embodiment of the present invention, the departure time and arrival time of the task node are both converted into the form of minutes, and the specific conversion method can be the existing method.

[0034] Furthermore, the initial task node sequence set can be obtained through the following steps:

[0035] S201, sort the nodes in the initial task node set according to the order of departure time from early to late to obtain a sorted task node set; set the counter k=1.

[0036] S202, if k≤Q, execute S203, otherwise, execute S208; Q is the number of task nodes in the initial task node sequence set.

[0037] S203, taking the kth task node in the sorted task node set as the current starting node; and executing S204.

[0038] S204, obtaining a task node that is allowed to be a successor node of the current starting task node from the sorted task node set. If the successor node is obtained, execute S205; otherwise, set k=k+1 and execute S202.

[0039] Among them, the task node allowed to be the successor node of the current starting task node is the task node whose time interval with the current starting task node is greater than or equal to the minimum stay time and whose departure airport is the same as the arrival airport of the current starting task node.

[0040] S205: If a successor node is obtained, the successor node is used as the target successor node of the current starting task node, and S206 is executed. If multiple successor nodes are obtained, the successor node with the earliest departure time is used as the target successor node of the current starting task node, and S206 is executed.

[0041] S206: If the arrival airport of the target successor node is the same as the departure airport of the current starting task node, execute S207; if the arrival airport of the target successor node is different from the departure airport of the current starting task node, use the target successor node as the current starting task node and execute S204;

[0042] S207: Based on the kth task node and the corresponding successor node, obtain the initial task node sequence corresponding to the task node and add it to the current task node sequence record set; set k=k+1, and execute S202. The initial value of the current task node sequence record set is empty.

[0043] S208 , obtaining an initial task node sequence set based on the current initial task node sequence record set, that is, the initial task node sequence set includes all initial task node sequences in the current initial task node sequence record set.

[0044] Those skilled in the art know that if all task nodes in the initial task node set are in the initial task node sequence set, the initial list of to-be-processed task nodes is empty.

[0045] S300 , if c≤c0 or the objective function value corresponding to the current node sequence set record set meets the convergence condition, execute S600 , otherwise, execute S400 .

[0046] The initial value of the current node sequence set record set is the initial task node sequence set. c0 is the preset number threshold. The objective function value of each task node sequence is used to represent the resource demand value required by the task node sequence. The resource demand value required by each task node sequence is determined based on the dwell slack time and delay time corresponding to the corresponding task node.

[0047] In an embodiment of the present invention, c0 can be set based on actual needs. The objective function value corresponding to the current node sequence set record set includes the objective function values ​​corresponding to all node sequence sets in the current node sequence set record set. The objective function value corresponding to each node sequence set includes the objective function value corresponding to each node sequence in the node sequence set.

[0048] In an embodiment of the present invention, the convergence condition may be that the objective function value tends to be stable and no longer changes.

[0049] In this embodiment of the present invention, the dwell slack time of a task node is the stopover slack time of the flight task, which is equal to the dwell time of the task node minus the minimum dwell time. The dwell time of the task node is the planned stopover time of the flight task. The planned stopover time of a flight task is equal to the time interval between the planned arrival time of the flight task and the planned departure time of the subsequent flight task.

[0050] In an embodiment of the present invention, the delay time of a task node is equal to the actual arrival time corresponding to the task node minus the corresponding planned arrival time. The actual arrival time of a task node is equal to the sum of the planned departure time of the task node, the affected delay time of the predecessor task node of the task node, and the independent delay time of the task node. The independent delay time of each task node can be determined based on the corresponding independent delay distribution, and the specific determination method can be the existing technology, for example, it can be equal to the expectation in the independent delay distribution. The affected delay time of the predecessor task node of each task node is determined based on the delay time of the predecessor task node and the corresponding stay slack time of the task node. If the difference obtained by subtracting the stay slack time corresponding to the task node from the delay time of the predecessor task node is greater than 0, the difference is the affected delay time, otherwise, the affected delay time is 0. In one illustrative embodiment, for example, if the planned transit time between two flight missions is 40 minutes and the minimum transit time is 30 minutes, then the dwell slack time is (40-30) minutes = 10 minutes. If the preceding flight is delayed by 20 minutes, a delay of (20-10) minutes = 10 minutes will be applied to the following flight. The 20 minutes represent the preceding flight's delay, and the 10 minutes not fully offset by the slack time represent the affected delay time affecting the following flight. In this embodiment of the present invention, the objective function value of each task node sequence is equal to the sum of the resource requirement values ​​of all corresponding task nodes. The resource requirement value of each task node is equal to a first resource requirement value determined based on the corresponding dwell slack time and a second resource requirement value determined based on the delay time.

[0051] In a specific embodiment of the present invention, the first resource requirement value R1 of each task node satisfies the following conditions:

[0052] R1=LC+Cd+Cz+CF.

[0053] Where LC is the parking fee corresponding to each mission node, that is, the fee that the aircraft needs to pay when parking at the airport. The parking fee can be determined based on the parking fee charging standard of the corresponding airport. Cd is the aircraft depreciation fee corresponding to each mission node, Cd=P z (1-S)Ts / N y N d L d , where P zis the sales price of the aircraft corresponding to the task node, i.e., the total cost of purchasing the aircraft. S is the residual value of the aircraft model, for example, it can be 5% of the corresponding aircraft sales price. Ts is the stay slack time corresponding to the task node. N y N is the service life of the aircraft corresponding to the mission node. d is the number of days in a year that the aircraft corresponding to the mission node is used, L d is the daily utilization rate of the aircraft corresponding to the mission node. Cz is the flight allowance paid to the crew corresponding to the mission node, which can be obtained based on existing methods. CF is the fuel consumption corresponding to the mission node, CF = PF × U × Ts. PF is the unit price of fuel, and U is the average hourly fuel consumption of the aircraft's auxiliary power unit during the aircraft's ground holding phase.

[0054] The second resource requirement value R2 of each task node meets the following conditions:

[0055] R2=Tc×△P.

[0056] Tc is the delay time of each task node, and △P is the resource demand value per unit delay time, which can be specifically the fee to be paid to each passenger for every minute of delay. It can be determined based on actual conditions. For example, the fee to be paid to each passenger for every minute of delay is 1 yuan.

[0057] It is known to those skilled in the art that, for the first node of a task node sequence, since there is no preceding node, the objective function value of the node is determined only based on the corresponding stay slack time; and for the last node of a task node sequence, since there is no subsequent node, the objective function value of the node is determined only based on the corresponding delay time.

[0058] S400: Based on the sequence destruction method and the sequence reconstruction method, the current task node sequence set is updated to obtain an updated task node sequence set and added to the current node sequence set record set. The initial value of the current task node sequence set is the initial task node sequence set, and S500 is executed.

[0059] In an embodiment of the present invention, the sequence destruction method is used to perform a destruction operation on the current task node sequence set, and the sequence reconstruction method is used to perform a reconstruction operation on the task node sequence set after the destruction operation.

[0060] S500 , based on a preset update result receiving rule, select one from the current task node sequence set and the updated task node sequence set as a new current task node sequence set; set c=c+1, and execute S300 .

[0061] In an embodiment of the present invention, the preset update result receiving rules include:

[0062] If the objective function value corresponding to the updated task node sequence set is less than the current optimal objective function value, the updated task node sequence set is selected as the new current task node sequence set.

[0063] In an embodiment of the invention, the current optimal objective function value is the minimum of the objective function values ​​corresponding to all task node sequence sets in the current node sequence set record set. If the objective function value corresponding to the updated task node sequence set is less than the current optimal objective function value, it is definitely less than the objective function value corresponding to the current task node sequence set. Therefore, the updated task node sequence set is selected as the new current task node sequence set.

[0064] If the objective function value corresponding to the updated task node sequence set is greater than or equal to the objective function value corresponding to the current task node sequence set, a new current task node sequence set is determined based on the objective function value corresponding to the current task node sequence set and the objective function value corresponding to the updated task node sequence set.

[0065] S600: taking the task node sequence set with the minimum resource requirement value in the current node sequence set record set as the target task node sequence set, and exiting the current control program.

[0066] Furthermore, S400 specifically includes:

[0067] S401, obtaining the objective function value corresponding to each task node sequence in the current task node sequence set, and deleting the task node sequence with the largest objective function value from the current task node sequence set, and adding it to the current list of pending task nodes.

[0068] S402 : Based on the current priorities of each sequence destruction method and each sequence reconstruction method, a sequence destruction method and a sequence reconstruction method to be currently selected are obtained as a current sequence destruction method and a current sequence reconstruction method, respectively.

[0069] In this embodiment of the present invention, the current priority of each sequence destruction method satisfies the following conditions:

[0070] P(d) i =WD i / ∑ m j=1 WD j .

[0071] P(d) i is the current priority of the i-th sequence destruction method, WD i is the current weight of the i-th sequence destruction method, i ranges from 1 to m, m is the number of sequence destruction methods, WD jis the current weight of the j-th sequence destruction method, and the value of j ranges from 1 to m.

[0072] In this embodiment of the present invention, the current priority of each sequence reconstruction method satisfies the following conditions:

[0073] P(r) u =WR u / ∑ n v=1 WR v .

[0074] P(r) u is the current priority of the u-th sequence reconstruction method, WR u is the current weight of the u-th sequence reconstruction method, u ranges from 1 to n, n is the number of sequence reconstruction methods, WR v is the current weight of the v-th sequence reconstruction method, and the value of v ranges from 1 to n.

[0075] In the embodiment of the present invention, the initial weight of each sequence destruction method is the same, that is, 1 / m. The initial weight of each sequence reconstruction method is the same, that is, 1 / n.

[0076] Furthermore, WD i The following conditions are met:

[0077] WD i = (1-φ)WD i last +φ(g1 i / t i );

[0078] Where φ is the update coefficient, which can be a value between (0, 1) or an empirical value. i last is the weight of the i-th sequence destruction method in the previous iteration, g1 i is the cumulative score of the ith sequence destruction method, t i is the number of times the i-th sequence destruction method is selected, that is, the number of times it is selected during the iteration process.

[0079] Furthermore, WR u The following conditions are met:

[0080] WR u = (1-φ)WR u last +φ(g2 u / t u );

[0081] Among them, WR u lastis the weight of the u-th sequence reconstruction method in the previous iteration, g2 u is the cumulative score of the u-th sequence reconstruction method, t u is the number of times the u-th sequence reconstruction method is selected.

[0082] Furthermore, the cumulative score of any sequence destruction method or any sequence reconstruction method meets the following conditions:

[0083] If, during a certain iteration, the objective function value corresponding to the updated task node sequence set obtained based on the sequence destruction method or the sequence reconstruction method is equal to the current optimal objective function value, the current cumulative score of the sequence destruction method or the sequence reconstruction method is increased by a first score; the current optimal objective function value is the minimum of the objective function values ​​corresponding to all task node sequence sets corresponding to the current node sequence set record set;

[0084] If the objective function value corresponding to the updated task node sequence set obtained based on the sequence destruction method or the sequence reconstruction method is greater than the current optimal objective function value, the current cumulative score of the sequence destruction method or the sequence reconstruction method is set to increase by a second score;

[0085] If the objective function value corresponding to the updated task node sequence set obtained based on a certain sequence destruction method or sequence reconstruction method is less than the current optimal objective function value, but the updated task node sequence set is selected as the new current task node sequence set, the current cumulative score of the sequence destruction method or the sequence reconstruction method is set to increase the third score.

[0086] In an embodiment of the present invention, the first score is less than the second score, but greater than the third score. The specific scores of the first score to the third score can be set based on actual needs. The initial value of the current cumulative score of the sequence destruction method or the sequence reconstruction method is 0. In an embodiment of the present invention, whether the task node sequence set obtained based on a certain sequence destruction method or sequence reconstruction method is selected as the new current task node sequence set is determined based on a preset update result reception rule.

[0087] S403, for any task node sequence in the current task node sequence set, the task node sequence is updated using the current sequence destruction method and the current sequence reconstruction method; an updated task node sequence set corresponding to the current task node sequence set is obtained, and stored in the current node sequence set record set.

[0088] Among them, S403 specifically includes: using the current sequence destruction method to perform a destruction operation on the task node sequence to obtain the destroyed task node sequence as the current task node sequence to be reconstructed, and using the current sequence reconstruction method to perform a reconstruction operation on the current task node sequence to be reconstructed to obtain the reconstructed task node sequence.

[0089] Among them, the destruction operation is used to delete at least one task node in the task node sequence that needs to be destroyed from the corresponding task node sequence and add it to the current list of pending task nodes, and the reconstruction operation is used to use the task nodes in the current list of pending task nodes to reconstruct the task node sequence after the destruction operation is executed.

[0090] Furthermore, in an illustrative embodiment of the present invention, a first sequence destruction method and a second sequence destruction method may be included, wherein the first sequence destruction method is used to randomly select a task node from the current task node sequence as a target destruction task node, and add the target destruction task node and its subsequent task nodes to the current list of task nodes to be processed, and the second sequence destruction method is used to add the task node with the maximum residence time in the current task node sequence and its subsequent task nodes to the current list of task nodes to be processed.

[0091] The first-order destruction method does not significantly improve solution quality, but it can increase solution diversity, escape local optima, and demonstrate the random nature of neighborhood search. The second-order destruction method destroys flight combinations with long stopover times, helping to escape local optima. In this embodiment of the present invention, the solution corresponds to a set of task node sequences generated during the iteration process.

[0092] In an exemplary embodiment of the present invention, a first sequence reconstruction method and a second sequence reconstruction method may be included. The first sequence reconstruction method is used to select task nodes from the current list of pending task nodes based on a greedy algorithm to reconstruct the damaged task node sequence, and the second sequence reconstruction method is used to select task nodes from the current list of pending task nodes based on a dynamic programming algorithm to reconstruct the damaged task node sequence so as to minimize the objective function value of the reconstructed task node sequence.

[0093] Furthermore, the first sequence reconstruction method is used to perform the following steps:

[0094] S10: For the tail node of the current intermediate task node sequence, select a task node that is allowed to be the successor node of the tail node from the current list of pending task nodes. If the corresponding successor node is obtained, execute S11; otherwise, execute S13. The initial value of the current intermediate task node sequence is the current sequence of task nodes to be reconstructed that has been destroyed.

[0095] S11: If a successor node is obtained, the successor node is used as the target successor node of the tail node, and S12 is executed. If multiple successor nodes are obtained, the successor node with the earliest departure time is used as the target successor node of the tail node, and S12 is executed.

[0096] S12, if the arrival airport of the target successor node is the same as the departure airport of the first node of the current intermediate task node sequence, that is, the first node, the current intermediate task node sequence is used as the reconstructed task node sequence of the current task node sequence to be reconstructed; if the arrival airport of the target successor node is not the same as the departure airport of the first node of the current intermediate task node sequence, the target successor node is used as the tail node of the current intermediate task node sequence, and S10 is executed.

[0097] S13, taking the initial value of the current intermediate task node sequence as the reconstructed task node sequence corresponding to the current intermediate task node sequence.

[0098] Furthermore, the second sequence reconstruction method is used to perform the following steps:

[0099] S20, for the tail node of the current intermediate task node sequence, select a task node that is allowed to be the successor node of the tail node from the current list of pending task nodes. If at least one successor node is obtained, execute S21, otherwise, execute S23.

[0100] S21: Obtain the objective function values ​​corresponding to the tail node and each successor node to obtain multiple objective function values. The successor node corresponding to the minimum of the multiple objective function values ​​is used as the target successor node for the tail node, and S22 is executed. The objective function values ​​corresponding to the tail node and each successor node are equal to the sum of the resource requirement value of the tail node and the resource requirement value of the successor node. Since the successor node does not have a stopover time, the resource requirement value of the successor node is determined based on the delay time of the successor node. The specific determination method can be found in the above description.

[0101] S22, if the arrival airport of the target successor node is the same as the departure airport of the head node of the current intermediate task node sequence, the current intermediate task node sequence is used as the reconstructed task node sequence of the current task node sequence to be reconstructed; if the arrival airport of the target successor node is not the same as the departure airport of the head node of the current intermediate task node sequence, the target successor node is used as the tail node of the current intermediate task node sequence and S20 is executed.

[0102] S23, taking the initial value of the current intermediate task node sequence as the reconstructed task node sequence corresponding to the current intermediate task node sequence.

[0103] Determining a new current task node sequence set based on the objective function value corresponding to the current task node sequence set and the objective function value corresponding to the updated task node sequence set specifically includes:

[0104] Based on the objective function value corresponding to the current task node sequence set and the objective function value corresponding to the updated task node sequence set, the current selection value P of the updated task node sequence set is obtained.

[0105] If P≥P0, the updated task node sequence set is selected as the new current task node sequence set, otherwise, the current task node sequence set is selected as the new current task node sequence set. P0 is a set selection value threshold, which can be an empirical value, such as 0.8.

[0106] Furthermore, P satisfies the following conditions:

[0107] P=e -（F（new）-F（cur）） / T（cur） .

[0108] Among them, F(new) is the objective function value corresponding to the updated task node sequence set, F(cur) is the objective function value corresponding to the current task node sequence set, and T(cur) is the current adjustment coefficient.

[0109] In an embodiment of the present invention, T(cur) satisfies the following condition: T(cur) = T(last) × ε, where T(last) is the adjustment coefficient from the previous iteration and ε is a preset coefficient, which is a value between (0, 1). The initial value of the adjustment coefficient is equal to ω × F0, where F0 is the objective function value corresponding to the initial set of task node sequences, and ω is a hyperparameter. An embodiment of the present invention also provides an electronic device comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, the instructions being configured to execute the method described in an embodiment of the present invention.

[0110] An embodiment of the present invention further provides a computer-readable storage medium storing computer-executable instructions, wherein the computer instructions are used to execute the method described in the embodiment of the present invention.

[0111] It should be understood that the various forms of the processes shown above can be used to reorder, add, or delete steps. For example, the steps described in the present invention can be performed in parallel, sequentially, or in a different order, as long as the desired results of the technical solutions disclosed in the present invention can be achieved. This is not limited herein.

[0112] The above specific embodiments do not limit the scope of protection of the present invention. Those skilled in the art will appreciate that various modifications, combinations, sub-combinations, and substitutions may be made based on design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention are intended to be included within the scope of protection of the present invention.

Claims

1. A flight ring scheduling method combining stopover slack time and flight delay, characterized in that: The method comprises the following steps: S100, based on the flight task information table to be processed, obtaining an initial task node set; wherein one node in the initial task node set corresponds to one flight task; S200, based on the greedy algorithm, obtain an initial task node sequence set that meets preset constraints from the initial task node set, and delete the task nodes in the initial task node sequence set from the initial task node set to obtain an initial list of task nodes to be processed; set the iteration counter c=1; S300, if c≤c0 or the objective function value corresponding to the current node sequence set record set meets the convergence condition, execute S600, otherwise, execute S400; the initial value of the current node sequence set record set is the initial task node sequence set; c0 is a preset number threshold; wherein the objective function value of each task node sequence is used to represent the resource demand value required by the task node sequence, and the resource demand value of each task node sequence is determined based on the stay slack time and delay time corresponding to the corresponding task node; S400, based on the sequence destruction method and the sequence reconstruction method, the current task node sequence set is updated to obtain an updated task node sequence set, and the updated task node sequence set is added to the current node sequence set record set; the initial value of the current task node sequence set is the initial task node sequence set, and S500 is executed; S500, based on a preset update result receiving rule, select one from the current task node sequence set and the updated task node sequence set as a new current task node sequence set; set c=c+1, and execute S300; S600: taking the task node sequence set with the minimum resource requirement value in the current node sequence set record set as the target task node sequence set, and exiting the current control program.

2. The method according to claim 1, characterized in that S400 specifically includes: S401, obtaining the objective function value corresponding to each task node sequence in the current task node sequence set, and deleting the task node sequence with the largest objective function value from the current task node sequence set, and adding it to the current list of pending task nodes; S402, based on the current priorities of each sequence destruction method and each sequence reconstruction method, obtaining the sequence destruction method and sequence reconstruction method currently to be selected as the current sequence destruction method and the current sequence reconstruction method respectively; S403: For any task node sequence in the current task node sequence set, update the task node sequence using the current sequence destruction method and the current sequence reconstruction method; obtain an updated task node sequence set corresponding to the current task node sequence set, and store it in the current node sequence set record set; Among them, S403 specifically includes: using the current sequence destruction method to perform a destruction operation on the task node sequence to obtain a destroyed task node sequence as the current task node sequence to be reconstructed, and using the current sequence reconstruction method to perform a reconstruction operation on the current task node sequence to be reconstructed to obtain a reconstructed task node sequence; Among them, the destruction operation is used to delete at least one task node in the task node sequence that needs to be destroyed from the corresponding task node sequence and add it to the current list of pending task nodes, and the reconstruction operation is used to use the task nodes in the current list of pending task nodes to reconstruct the task node sequence after the destruction operation is executed.

3. The method according to claim 2, characterized in that The current priority of each sequence destruction method satisfies the following conditions: P(d) i =WD i / ∑ m j=1 WD j ; P(d) i is the current priority of the i-th sequence destruction method, WD i is the current weight of the i-th sequence destruction method, i ranges from 1 to m, m is the number of sequence destruction methods, WD j is the current weight of the j-th sequence destruction method, and the value of j ranges from 1 to m; The current priority of each sequence reconstruction method satisfies the following conditions: P(r) u =WR u / ∑ n v=1 WR v ; P(r) u is the current priority of the u-th sequence reconstruction method, WR u is the current weight of the u-th sequence reconstruction method, u ranges from 1 to n, n is the number of sequence reconstruction methods, WR v is the current weight of the v-th sequence reconstruction method, and the value of v ranges from 1 to n.

4. The method according to claim 3, characterized in that WD i The following conditions must be met: WD i = (1-φ) WD i last +φ(g1) i / t i ); Among them, φ is the update coefficient, WD i last is the weight of the i-th sequence destruction method in the previous iteration, g1 i is the cumulative score of the ith sequence destruction method, t i is the number of times the i-th sequence destruction method is selected; WR u The following conditions must be met: WR u =(1-φ)WR u last +φ(g2 u / t u ); Among them, WR u last is the weight of the u-th sequence reconstruction method in the previous iteration, g2 u is the cumulative score of the u-th sequence reconstruction method, t u is the number of times the u-th sequence reconstruction method is selected.

5. The method according to claim 4, characterized in that The cumulative score of any sequence destruction method or any sequence reconstruction method meets the following conditions: If, during a certain iteration, the objective function value corresponding to the updated task node sequence set obtained based on the sequence destruction method or the sequence reconstruction method is equal to the current optimal objective function value, the current cumulative score of the sequence destruction method or the sequence reconstruction method is increased by a first score; The current optimal objective function value is the minimum of the objective function values ​​corresponding to all task node sequence sets corresponding to the current node sequence set record set; If the objective function value corresponding to the updated task node sequence set obtained based on the sequence destruction method or the sequence reconstruction method is greater than the current optimal objective function value, the current cumulative score of the sequence destruction method or the sequence reconstruction method is set to increase by a second score; If the objective function value corresponding to the updated task node sequence set obtained based on a certain sequence destruction method or sequence reconstruction method is less than the current optimal objective function value, but the updated task node sequence set is selected as the new current task node sequence set, the current cumulative score of the sequence destruction method or the sequence reconstruction method is set to increase the third score.

6. The method according to claim 1, characterized in that S500 specifically includes: If the objective function value corresponding to the updated task node sequence set is less than the current optimal objective function value, the updated task node sequence set is selected as the new current task node sequence set; the current optimal objective function value is the minimum of the objective function values ​​corresponding to all task node sequence sets corresponding to the current node sequence set record set; If the objective function value corresponding to the updated task node sequence set is greater than or equal to the objective function value corresponding to the current task node sequence set, a new current task node sequence set is determined based on the objective function value corresponding to the current task node sequence set and the objective function value corresponding to the updated task node sequence set.

7. The method according to claim 6, characterized in that The determining of a new current task node sequence set based on the objective function value corresponding to the current task node sequence set and the objective function value corresponding to the updated task node sequence set specifically includes: Based on the objective function value corresponding to the current task node sequence set and the objective function value corresponding to the updated task node sequence set, obtaining the current selection value P of the updated task node sequence set; If P≥P0, the updated task node sequence set is selected as the new current task node sequence set; otherwise, the current task node sequence set is selected as the new current task node sequence set, and P0 is the set selection value threshold.

8. The method according to claim 7, characterized in that P satisfies the following conditions: P=e -（F（new）-F（cur）） / T（cur） ; Among them, F(new) is the objective function value corresponding to the updated task node sequence set, F(cur) is the objective function value corresponding to the current task node sequence set, and T(cur) is the current adjustment coefficient.

9. The method according to claim 2, characterized in that It includes a first sequence destruction method and a second sequence destruction method, wherein the first sequence destruction method is used to randomly select a task node from the current task node sequence as a target destruction task node, and add the target destruction task node and its subsequent task nodes to the current list of task nodes to be processed; the second sequence destruction method is used to add the task node with the maximum residence time in the current task node sequence and its subsequent task nodes to the current list of task nodes to be processed.

10. The method according to claim 2, characterized in that The method includes a first sequence reconstruction method and a second sequence reconstruction method. The first sequence reconstruction method is used to select task nodes from the current list of task nodes to be processed based on a greedy algorithm to reconstruct the damaged task node sequence. The second sequence reconstruction method is used to select task nodes from the current list of task nodes to be processed based on a dynamic programming algorithm to reconstruct the damaged task node sequence.

Images