A Large-Scale Flight Regulation Method Based on Multi-Objective Collaborative Optimization

Abstract

The present invention relates to a large-scale flight regulation method based on multi-objective collaborative optimization, belonging to the technical field of civil aviation management, and solves the problems of inappropriate existing grouping strategies and the failure to consider multi-objective collaborative optimization. It includes generating a delay time vector and a first delay time for each flight; updating the departure time and arrival time of the flight to obtain flight conflict information; grouping all flights according to the flight time overlap degree to obtain multiple flight groups; generating a sub-population from the delay time vectors of each flight in each flight group, successively crossing each sub-population to generate offspring, obtaining the delay variables of each flight and performing mutation, and after multiple evolutions, taking the solution with the maximum fitness in each optimized sub-population as the second delay time of the corresponding flight, updating the first delay time of each flight to the second delay time, and using the first delay time obtained after multiple cycles to regulate the departure time of the corresponding flight. The dual optimization of flight conflicts and delays is achieved.

Description

Technical Field

[0001] The present invention relates to the field of civil aviation management technology, and in particular to a large-scale flight control method based on multi-objective collaborative optimization. Background Art

[0002] With the continuous development of the aviation industry and the increasing demand for air travel, the congestion of existing airspace and the probability of conflicts between flights have also increased, causing serious safety threats and leading to problems such as flight delays. Resolving flight conflicts may inevitably cause flight delays, while minimizing flight delays may increase the probability of conflicts. Therefore, the problem is essentially an optimization problem with two conflicting objectives.

[0003] In the past, many researchers have conducted systematic research on this problem, mainly focusing on single-objective mathematical programming formulas. Among them, the basic 0-1 integer programming model was proposed in 1998. Its purpose is to minimize flight delays under airport capacity and department capacity constraints by optimizing the time slots of flight departure and arrival. It should be noted that airspace congestion and flight conflicts are often caused by multiple factors, such as flight routes, traffic density, etc., which are difficult to characterize with linear functions. With the continuous increase in the scale of the problem and the complexity of the objective function, traditional mathematical methods are not suitable for solving such problems and are difficult to adapt to the requirements of actual conditions.

[0004] Genetic algorithm, a random optimization technology, can efficiently solve multi-objective, high-dimensional complex optimization problems using heuristic algorithms. In recent years, the co-evolution method has been successfully applied to this type of problem. The main idea is to use a random grouping strategy to randomly divide the variables of large-scale problems into several equal groups, thereby decomposing the problem into several sub-components for a divide-and-conquer solution. The key step in this idea is the problem decomposition method, such as random grouping, or decomposing an n-dimensional problem into 2 n / 2-dimensional problems and solving them recursively in sequence. However, when the variables of the problem are interactive, the grouping method still needs to be improved. For many flights, due to the complexity of the interactions between them, the random grouping strategy is not very suitable, so it is necessary to explore new grouping schemes under the constraints. Summary of the invention

[0005] In view of the above analysis, an embodiment of the present invention aims to provide a large-scale flight control method based on multi-objective collaborative optimization, so as to solve the problem that the existing grouping strategy is not suitable and does not consider multi-objective collaborative optimization.

[0006] The embodiment of the present invention provides a large-scale flight control method based on multi-objective collaborative optimization, comprising the following steps:

[0007] Acquire flight data, and generate a delay time vector and a first delay time for each flight;

[0008] According to the first delay time of each flight, the departure time and arrival time of each flight are updated, and the flight conflict information is obtained by detecting the flight conflicts at each time; according to the overlap of the flight time of any two flights, all flights are grouped to obtain multiple flight groups;

[0009] A subpopulation is generated according to the delay time vector of each flight in each flight group, and each subpopulation is crossed to generate offspring in turn through a fast genetic algorithm, and the delay variable of each flight is obtained according to the flight conflict information, and mutation is performed based on the delay variable to complete a genetic evolution, and an optimized subpopulation is obtained after multiple evolutions;

[0010] The solution with the largest fitness in each optimized subpopulation is taken as the second delay time of the corresponding flight, and the first delay time of each flight is updated to the second delay time. The take-off time and arrival time of each flight are updated again, and flight conflicts are detected, grouped and genetically evolved, and the first delay time is updated. The first delay time obtained after multiple cycles is used to adjust the take-off time of the corresponding flight.

[0011] Based on the further improvement of the above method, the acquisition of flight conflict information by detecting flight conflicts at each time comprises:

[0012] Based on the condition that each flight has the same flight altitude and speed, the flight path is divided at fixed time intervals to obtain a state vector of each flight at each moment; based on the state vector of each flight, whether any two flights have a flight conflict at each moment is identified in turn, and flight conflict information is obtained by statistics, including the total number of conflicts at each moment and the total number of conflicts for each flight.

[0013] Based on the further improvement of the above method, the state vector of each flight at each time includes a horizontal coordinate and a vertical coordinate, a flight altitude, and a corresponding time;

[0014] The step of sequentially identifying whether any two flights at each moment have a flight conflict according to the state vector of each flight includes:

[0015] At each time, add the square of the difference between the horizontal coordinates and the vertical coordinates of any two flights and then take the square root to get the interval distance;

[0016] If the interval distance is less than the distance threshold, a flight conflict will occur between the corresponding two flights; otherwise, no flight conflict will occur.

[0017] Based on the further improvement of the above method, the flight time overlap between any two flights is grouped to obtain multiple flight groups, including:

[0018] Based on the flight data, sort all flights and add the first flight to the first flight group;

[0019] Starting from the second flight, one flight is taken out in turn as the current flight, the intersection of the flight time periods of the current flight and the remaining flights is calculated, and the flight with the largest intersection length is obtained as the flight to be grouped;

[0020] If the flight to be grouped has been added to a flight group, the current flight is added to the same flight group as the flight to be grouped; otherwise, a new flight group is created, and the current flight is added to the new flight group;

[0021] The grouping is completed until all flights have been added to a flight group, and multiple flight groups are obtained.

[0022] Based on the further improvement of the above method, the delay time vector of each flight is {0, ts ,2× ts ,3× ts ,..., δ max} as the delay time selection range, randomly select S times, and obtain the delay time vector , where 1≤ i ≤ N , N is the total number of flights in the flight data, ts is the sampling time, δ max is the maximum flight delay time and can be ts Divisible.

[0023] Based on the further improvement of the above method, a subpopulation is generated according to the delay time vector of each flight in each flight group, including: taking the delay time vector of each flight in each flight group as each row of the matrix, taking the delay time of each flight in the same column as a chromosome of the current subpopulation, and taking each delay time as a gene.

[0024] Based on the further improvement of the above method, the step of obtaining the delay variable of each flight according to the flight conflict information includes:

[0025] Based on the total number of conflicts at each time in the flight conflict information, and according to the difference between the total number of conflicts between each two adjacent time points, the maximum slope value, the minimum slope value and the slope value at each time point are obtained;

[0026] According to the flight corresponding to each gene on the chromosome during the genetic evolution process, as the flight to be mutated, according to the updated take-off time of the flight to be mutated, the slope value of the flight to be mutated at the corresponding time is obtained;

[0027] If the slope value of the flight to be mutated is less than 0, the delay variable of the flight to be mutated is the inverse of the ratio of the slope value of the flight to be mutated to the minimum slope value; otherwise, the delay variable of the flight to be mutated is the ratio of the slope value of the flight to be mutated to the maximum slope value.

[0028] Based on a further improvement of the above method, the mutating based on the delay variable includes:

[0029] According to the total number of conflicts of each flight in the current subpopulation, obtain the maximum total number of conflicts;

[0030] Randomly select the conflict threshold within the range less than or equal to the maximum total number of conflicts;

[0031] It is determined in turn whether the total number of conflicts of the flight to be mutated corresponding to the current gene is greater than the conflict threshold. If it is greater, the mutation value is obtained according to the delay variable of the flight to be mutated, and the current gene value is updated to the mutation value; otherwise, the current gene value remains unchanged.

[0032] Based on the further improvement of the above method, the step of obtaining the mutation value according to the delay variable of the flight to be mutated includes:

[0033] When the delay variable of the flight to be mutated is greater than rand(0,1), a value is randomly selected within the range less than or equal to the current gene as the mutation value;

[0034] When the delay variable of the flight to be mutated is less than -rand(0,1), a value is randomly selected within the range of the current gene and the maximum flight delay time as the mutation value;

[0035] When the delay variable of the flight to be mutated meets the other conditions, a value is randomly selected within the range of less than or equal to the maximum flight delay time as the mutation value.

[0036] Based on the further improvement of the above method, the fitness is calculated according to the following formula:

[0037]

[0038] in, For the k Among the subpopulations l The fitness of a chromosome, m k For the k Total number of flights in the subpopulation; δmax is the maximum flight delay time, For the k Among the subpopulations l Chromosome No. j genes, NC kj For the k Among the subpopulations j The total number of conflicts for flights corresponding to the gene.

[0039] Compared with the prior art, the present invention can achieve at least one of the following beneficial effects:

[0040] 1. Based on the scenario of a large number of flights, overlapping time periods, and complex routes, a control scheme was designed using the ideas of dynamic grouping and co-evolution to meet the optimization needs of flight conflicts and delays. The method of dynamic grouping based on the time overlap of flights enables highly coupled flights to participate in the optimization as a whole, and optimizes the interactive flights into one group to the maximum extent, so as to reduce the risk of premature convergence; co-evolution reasonably decomposes the original high-dimensional optimization problem into several lower-dimensional optimization problems for centralized solution, obtains high-quality solutions, and realizes the dual optimization of flight conflicts and delays;

[0041] 2. Grouping by time overlap only requires calculation and sorting of overlapping time periods for every two flights, which greatly reduces the time required for grouping and improves the speed of each round of calculation;

[0042] 3. A mutation operator combined with a local search mechanism is used, which can perform mutations that meet a certain trend near the original value, achieving the effect of "cutting peaks and filling valleys", making the distribution of conflicts more balanced, facilitating the next round of optimization, and thus reducing conflicts.

[0043] In the present invention, the above-mentioned technical solutions can also be combined with each other to achieve more preferred combination solutions. Other features and advantages of the present invention will be described in the subsequent description, and some advantages can become obvious from the description, or can be understood by practicing the present invention. The purpose and other advantages of the present invention can be realized and obtained through the contents particularly pointed out in the description and the drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] The accompanying drawings are only used for the purpose of illustrating specific embodiments and are not to be considered as limiting the present invention. In the entire drawings, the same reference symbols represent the same components;

[0045] Figure 1 This is a flow chart of a large-scale flight control method based on multi-objective collaborative optimization in Example 1 of the present invention;

[0046] Figure 2 This is a schematic diagram of flight routes and conflicts in Example 1 of the present invention;

[0047] Figure 3 Schematic diagram of conflict detection based on state vector comparison in Embodiment 1 of the present invention;

[0048] Figure 4 This is a flow chart of flight grouping based on flight time overlap in Embodiment 1 of the present invention;

[0049] Figure 5 This is a schematic diagram showing the change of the overall number of flight conflicts over time in Example 1 of the present invention;

[0050] Figure 6 This is a comparison chart of the implementation effects of the flight control method in Example 2 of the present invention. DETAILED DESCRIPTION

[0051] The preferred embodiments of the present invention are described in detail below in conjunction with the accompanying drawings, wherein the accompanying drawings constitute a part of this application and are used together with the embodiments of the present invention to illustrate the principles of the present invention, but are not used to limit the scope of the present invention.

[0052] A specific embodiment of the present invention discloses a large-scale flight control method based on multi-objective collaborative optimization. In this method, first, the conflict situations of each flight are obtained based on the flight information, and then the flights are dynamically grouped according to the overlap in the take-off and landing time between the flights, so as to achieve the division and conquer of the original problem; secondly, a fast genetic algorithm is used to cross and mutate the chromosomes in the subpopulations corresponding to each group, and fitness selection is performed at the gene level. When the subpopulations have evolved to a certain generation, the method of fitness evaluation of chromosomes is used to select the optimal solution from each subpopulation, and further merge them into the optimal solution to the original problem. Under the premise of reasonable division and conquer, the conflicts and delays of each flight are weighed to achieve the optimization of flight take-off slots. Figure 1 As shown, the method comprises the following steps:

[0053] S11: Acquire flight data, and generate a delay time vector and a first delay time for each flight;

[0054] It should be noted that the flight data obtained includes: flight number, departure time, arrival time, delay time, departure airport coordinates, landing airport coordinates and flight path. The maximum flight delay time of all flights in the flight data is obtained by statistics. δ max .

[0055] According to the sampling time of flight data ts, a delay time vector is generated for each flight, which is a delay time vector composed of multiple times that are randomly selected from the same number and divisible by the sampling time within the range less than or equal to the maximum flight delay time.

[0056] Specifically, {0, ts ,2× ts ,3× ts ,..., δ max} as the delay time selection range, here δ max / ts+ Randomly select one at a time and repeat S times, get flights F i The delay time vector . Among them, 1≤ i ≤ N , N is the total number of flights in the flight data, δ max Can be ts Divisible by 80≤ S ≤100, can be randomly selected multiple times at the same time.

[0057] For example, ts 0.5 minutes, δ max For 90 minutes, S is 80.

[0058] The first delay time of each flight is initially set to 0, and is continuously updated through subsequent steps. The final first delay time is used to adjust the take-off time of the corresponding flight to achieve optimization of the flight take-off slot.

[0059] S12: updating the departure time and arrival time of each flight according to the first delay time of each flight, and obtaining flight conflict information by detecting flight conflicts at each time; grouping all flights into multiple flight groups according to the flight time overlap between any two flights;

[0060] It should be noted that the protection zone for aircraft in flight is defined as a cylindrical area with a radius of 5 nautical miles and an altitude of 2,000 feet. When the protection zones of two flights overlap, a conflict will occur. Figure 2 The following is a schematic diagram of flight routes and conflicts. F i The flight route is (A, W1, W3, W4, D), flight F jThe flight path of the two flights is (B, W2, W3, W5, C). Assuming that the two flights fly at the same speed and at the same altitude, W3 is a waypoint where a potential conflict may occur.

[0061] The departure time and arrival time of each flight are updated, that is, the corresponding first delay time is added to the departure time and arrival time of each flight.

[0062] By detecting flight conflicts at each time, flight conflict information is obtained, including:

[0063] ① Based on the condition that each flight has the same flight altitude and speed, the flight path is divided into fixed time intervals to obtain the state vector of each flight at each time;

[0064] It should be noted that the state vector includes the horizontal coordinate x and the vertical coordinate y , flight altitude h , and the corresponding time t , expressed as ( x , y , h , t ). By dividing the flight path into fixed time intervals according to the flight path and take-off time, we can get the state vector corresponding to each flight at each time, for example, at intervals of 10 seconds.

[0065] ② According to the state vector of each flight, identify in turn whether there is a flight conflict between any two flights at each time, and obtain flight conflict information statistically, including the total number of conflicts at each time and the total number of conflicts for each flight.

[0066] Specifically, according to the state vector of each flight, we identify whether any two flights have a flight conflict at each moment, including:

[0067] At each time, add the square of the difference between the horizontal coordinates and the vertical coordinates of any two flights and then take the square root to get the interval distance;

[0068] If the interval distance is less than the distance threshold, the corresponding two flights will have a flight conflict, otherwise, no flight conflict will occur.

[0069] For example, the flight altitude is h 0 , t Flight Schedule F i The state vector ( x i , y i , h 0 ,t ) and flights F j 's state vector ([[]] x j , y j , h 0 , t ), if , it means that two flights will conflict, otherwise they will not. In Figure 3 , the points connected by the dotted line correspond to the same time instant, and the part in the circle is the conflict area.

[0070] All flights are pairwise detected for flight conflicts at each moment, and finally the total number of conflicts at each moment and the total number of conflicts for each flight are obtained.

[0071] Then, the flights are dynamically grouped according to the overlap degree of flight times, and multiple flight groups are obtained. The flow chart is as Figure 4 shown, including:

[0072] ① Based on flight data, sort all flights and add the first flight to the first flight group;

[0073] It should be noted that for each flight in the flight data, the overlap degree of the flight time with other flights will be identified. Therefore, the sorting method is not limited in this embodiment. Exemplarily, it can be sorted according to the flight number or departure time.

[0074] ② Starting from the second flight, take out one flight in turn as the current flight, calculate the intersection of the flight time period of the current flight and the remaining flights, and obtain the flight with the largest intersection length as the flight to be grouped;

[0075] It should be noted that the overlap degree of flight times is represented by the intersection length of the flight time periods. Starting from flight F 2 , calculate the overlap period length between flight F i and the remaining flights F j (2 ≤ i ≤ N， 1 ≤ j ≤ N and j ≠ i,N is the total number of flights), that is, find the intersection of the flight time periods of the two flights to obtain the time overlap degree t ij = len (( D i , A i ) ∩ ([[]]D j , A j )),in, D i For flights F i Updated departure time, A i For flights F i Updated arrival time, D j For flights F j Updated departure time, A j For flights F j Updated arrival time.

[0076] Each flight will be connected to the remaining N -1 flight formed N -1 corresponding time overlap, where the largest time overlap corresponds to two flights ( F jmax , F i ) F jmax These are the flights to be grouped.

[0077] ③ If the flight to be grouped has been added to a flight group, add the current flight to the same flight group as the flight to be grouped; otherwise, create a new flight group and add the current flight to the new flight group;

[0078] ④ Until all flights have been added to a flight group, the grouping is completed and multiple flight groups are obtained.

[0079] The grouping is done based on the time overlap of the aircraft, so that each flight is grouped with the flight with the largest time overlap as much as possible. Time overlap means that flight conflicts may occur during this period, so this dynamic grouping essentially dynamically groups flights that may conflict, and realizes the grouping of flights with interactive relationships so that they can be genetically evolved as a subpopulation later.

[0080] In addition, using time overlap grouping only requires calculating and sorting the overlapping time periods of two flights, so its corresponding time complexity is O( N 2 ),in N is the total number of flights, and in the prior art, the time complexity of grouping according to conflict relations is O(T× N 2), where T is the time step for calculating conflicts. It can be seen that the grouping according to time overlap adopted in this embodiment greatly reduces the time required for grouping and improves the speed of each round of calculation.

[0081] S13: Generate a subpopulation according to the delay time vector of each flight in each flight group, and use a fast genetic algorithm to cross each subpopulation in turn to generate offspring, obtain the delay variable of each flight according to the flight conflict information, and perform mutation based on the delay variable to complete a genetic evolution, and obtain the optimized subpopulation after multiple evolutions;

[0082] It should be noted that a genetic evolution is completed once for each flight group through steps S131-S134.

[0083] S131: construct subpopulations and obtain chromosomes;

[0084] Specifically, the delay time vectors of each flight in each flight group are organized into a subpopulation, including: taking the delay time vector of each flight in each flight group as each row of the matrix, taking the delay time of each flight in the same column as a chromosome of the current subpopulation, and taking each delay time as a gene.

[0085] For k For group, ,in For the k The first j flights, m k For the k The total number of flights in the group, then k Subpopulation corresponding to the group subpop k It is expressed as:

[0086]

[0087] Formula (1)

[0088] in, For flights The corresponding delay time vector, 2≤ k ≤ s , s is the total number of subpopulations, is a chromosome, representing a possible solution, For the k The first delay time in the delay time vector for the first flight in the group.

[0089] S132: cross subpopulations to produce offspring;

[0090] It should be noted that for The genes on this chromosome , its fitness is:

[0091] Formula (2)

[0092] in, NC kj For the k The first j The total number of conflicts for flights, i.e. k Among the subpopulations j The total number of conflicts in flights corresponding to genes, δ max is the maximum flight delay time.

[0093] This crossover method does not calculate the fitness of the entire chromosome, but introduces the fitness of each gene in the chromosome, and considers reducing ground delays and conflicts for each flight during the crossover process. The parameters in the formula are determined at the beginning, and it has the characteristic of quickly finding a better solution.

[0094] The chromosomes of the parent subpopulation are randomly paired to produce two daughter chromosomes, where the probability of crossover is c, 0.6≤c≤0.9, preferably c=0.7. A number between (0,1) is randomly generated. If the number is less than c, crossover occurs, otherwise no crossover occurs.

[0095] During crossover, comparisons are made on the parent chromosome and The alleles of and No. j Alleles, including:

[0096] ① When When j Alleles inherited from No. j Alleles.

[0097] ② When When j Alleles inherited from No. j bit gene.

[0098] ③When When the chromosomes of the two daughters j The alleles are:

[0099] Formula (3)

[0100] Formula (4)

[0101] in, floor () is a function that rounds down. ,and u j ∈U(0,1), randomly selected, η >0 is the distribution index, preferably, η= 1.

[0102] The crossover operator used here simulates binary single-point crossover. j It is determined dynamically and randomly, so the operator can jump out of the local optimal solution and has good global search ability. It has excellent performance for the sparse individual space of high-dimensional target optimization problems.

[0103] S133: Obtain the delayed variable and make a mutation;

[0104] If crossover occurs in step S132, the newly generated offspring chromosome is mutated; if crossover does not occur, the parent chromosome is directly mutated.

[0105] Specifically, the delay variable of each flight obtained based on the flight conflict information is used to mark whether the flight should take off early or late. The acquisition method includes:

[0106] ① Based on the total number of conflicts at each time in the flight conflict information, the maximum slope value, the minimum slope value and the slope value at each time are obtained according to the difference between the total number of conflicts between each two adjacent times;

[0107] Exemplarily, the total number of conflicts at each moment obtained in step S12 is plotted into a conflict number curve varying with time, as shown in the schematic diagram. Figure 5 As shown, the maximum slope value is marked K max , minimum slope value K min .

[0108] ② According to the flight corresponding to each gene on the chromosome during the genetic evolution process, as the flight to be mutated, according to the updated take-off time of the flight to be mutated, the slope value of the flight to be mutated at the corresponding time is obtained;

[0109] For example, one gene on a chromosome corresponds to flight F i , according to its updated departure time, Figure 5 Get the slope value from K i .

[0110] ③If the slope value of the flight to be mutated is less than 0, the delay variable of the flight to be mutated is the inverse of the ratio of the slope value of the flight to be mutated to the minimum slope value; otherwise, the delay variable of the flight to be mutated is the ratio of the slope value of the flight to be mutated to the maximum slope value.

[0111] Specifically, if the flight to be mutated F i Slope value K i If it is less than 0, it means that the total number of flight conflicts corresponding to the flight to be mutated is decreasing when it takes off. flag i =- K i / K min , otherwise, the delay variable flag i = K i / K max .

[0112] It can be seen that the delay variable flag i ∈[-1,1].

[0113] Mutations based on delayed variables, including:

[0114] ①According to the total number of conflicts of each flight in the current subpopulation, obtain the maximum total number of conflicts;

[0115] ② Randomly select the conflict threshold within the range less than or equal to the maximum total number of conflicts;

[0116] It should be noted that in [0, NC kmax ] randomly selects the conflict threshold NC kh Used to determine whether a mutation should be made. NC kmax For the k The maximum number of conflicts in a group of flights, NC kh For the k The conflict threshold for the group.

[0117] ③ Determine in turn whether the total number of conflicts of the flight to be mutated corresponding to the current gene is greater than the conflict threshold. If it is, obtain the mutation value according to the delay variable of the flight to be mutated, and update the current gene value to the mutation value; otherwise, the current gene value remains unchanged.

[0118] Specifically, the flight to be mutated corresponding to the current gene is flight F i , the total number of conflicts is NC ki :

[0119] if NC ki > NC kh ,, represents the flight F i The total number of conflicts exceeds the conflict threshold, and a mutation should occur according to the flight F i Delay variable flag i Get mutation values, including:

[0120] When the delay variable of the flight to be mutated is greater than rand(0,1), a value is randomly selected within the range less than or equal to the current gene as the mutation value;

[0121] That is: when flag i >rand(0,1), then A value is randomly selected as the new delay time of the flight, indicating that the flight should take off appropriately earlier. For the k The first i Flight No. l The corresponding gene in each chromosome, that is, the delay time for take-off.

[0122] When the delay variable of the flight to be mutated is less than -rand(0,1), a value is randomly selected within the range of the current gene and the maximum flight delay time as the mutation value;

[0123] That is: when flag i <-rand(0,1), then A value is randomly selected as the new delay time of the flight, indicating that the flight should delay its take-off appropriately.

[0124] When the delay variable of the flight to be mutated meets the other conditions, a value is randomly selected within the range of less than or equal to the maximum flight delay time as the mutation value.

[0125] That is: the rest of the cases are in [0, δ max ] generates new delay times randomly, with no tendency to be early or late.

[0126] if NC ki ≤ NC kh , representing the flight F iThe total number of conflicts does not exceed the conflict threshold, no mutation occurs, and the gene value does not change.

[0127] In the genetic mutation process, considering that the overall number of conflicts increases when the flight departs, it means that the flight should take off appropriately in advance to avoid the upcoming peak of conflicts; the overall number of conflicts decreases when the flight departs, which means that the flight can be delayed appropriately, which can further make the number of conflicts more balanced and avoid concentrated conflicts. At the same time, the setting of the threshold makes it more likely that flights with more conflicts will change their delay time, while having less impact on flights with fewer conflicts, thereby effectively affecting the takeoff schedule of flights with more conflicts. Overall, the mutation operator achieves the effect of "peak shaving and valley filling" by making certain adjustments near the original value, making the distribution of conflicts more balanced, facilitating the next optimization, and thus reducing conflicts.

[0128] S134: Determine whether to continue genetic evolution based on the evolutionary generation;

[0129] Specifically, if the current evolutionary generation is less than the preset maximum evolutionary generation of the subpopulation, the evolutionary generation is increased by 1, and the chromosome composition population obtained in step S133 is used as the subpopulation, and the process returns to step S132 to start another crossover and mutation. Otherwise, the genetic evolution is completed for the current multiple flight groups to obtain multiple optimized subpopulations, and step S14 is executed. Exemplarily, the maximum evolutionary generation of the subpopulation is set to 80 generations.

[0130] S14: Take the solution with the largest fitness in each optimized subpopulation as the second delay time of the corresponding flight, update the first delay time of each flight to the second delay time, update the take-off time and arrival time of each flight again, perform flight conflict detection, grouping and genetic evolution, and update the first delay time. The first delay time obtained after multiple cycles is used to adjust the take-off time of the corresponding flight.

[0131] Specifically, each optimized subpopulation selects the chromosome with the largest fitness among its chromosomes. is the solution obtained in this round of iteration, where the fitness is calculated according to the following formula:

[0132] Formula (5)

[0133] in, For the k Among the subpopulations l The fitness of a chromosome, m k For the k Total number of flights in the subpopulation; δ maxis the maximum flight delay time, For the k Among the subpopulations l Chromosome No. j genes, NC kj For the k Among the subpopulations j The total number of conflicts for flights corresponding to the gene.

[0134] Each gene value on the chromosome with the largest fitness is the latest delay time of the corresponding flight, which is used as the second delay time, and the first delay time of each flight is updated to the second delay time. If the current overall evolutionary generation is less than the preset overall maximum evolutionary generation, then the overall evolutionary generation is increased by 1, and the process returns to execute steps S12 and S13 again to detect, group and genetically evolve flight conflicts; otherwise, the optimization and control process is completed, and the first delay time finally obtained is used to control the take-off time of the corresponding flight.

[0135] It should be noted that, in order to ensure that the obtained solution is relatively stable, a larger evolutionary generation can be used. Preferably, the overall maximum evolutionary generation is set to 800 generations.

[0136] Based on the national air traffic flow data from 9:00 to 12:00 on October 1, 2018, flight data was obtained, and the traditional NSGA-Ⅱ, NSGA-Ⅲ, RVEA algorithms and the control algorithm in this embodiment were used respectively to obtain the following: Figure 6 The results are shown. Figure 6 In the figure, the horizontal axis represents the total number of conflicts, and the vertical axis represents the deviation between the arrival time at the destination and the planned time. It can be seen that through the control scheme of this embodiment (the Recombination algorithm represented by the dots), the number of conflicts is significantly reduced and the delay time is also greatly reduced.

[0137] Compared with the prior art, the large-scale flight control method based on multi-objective collaborative optimization provided in this embodiment is a control scheme designed for the optimization needs of flight conflicts and delays using the ideas of dynamic grouping and collaborative evolution. It uses time overlap to group, which greatly reduces the time required for grouping, improves the speed of each round of calculation, and enables highly coupled flights to participate in the optimization as a whole, and optimizes and groups flights with interacting relationships as much as possible to reduce the risk of premature convergence; collaborative evolution reasonably decomposes the original high-dimensional optimization problem into several lower-dimensional optimization problems for centralized solution, and adopts a mutation operator combined with a local search mechanism, which can perform mutations that meet a certain trend near the original value, achieving the effect of "peak shaving and valley filling", making the distribution of conflicts more balanced, thereby reducing conflicts, obtaining high-quality solutions, and achieving dual optimization of flight conflicts and delays.

[0138] Those skilled in the art will appreciate that all or part of the processes of the above-mentioned embodiments can be implemented by instructing related hardware through a computer program, and the program can be stored in a computer-readable storage medium, wherein the computer-readable storage medium is a disk, an optical disk, a read-only storage memory, or a random access memory, etc.

[0139] The above description is only a preferred specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any changes or substitutions that can be easily conceived by any technician familiar with the technical field within the technical scope disclosed by the present invention should be covered within the protection scope of the present invention.

Claims

1. A large-scale flight control method based on multi-objective collaborative optimization, It is characterized in that The steps include: Acquire flight data, and generate a delay time vector and a first delay time for each flight; According to the first delay time of each flight, the departure time and arrival time of each flight are updated, and the flight conflict information is obtained by detecting the flight conflicts at each time; according to the overlap of the flight time of any two flights, all flights are grouped to obtain multiple flight groups; A subpopulation is generated according to the delay time vector of each flight in each flight group, wherein the delay time of the same column of each flight is used as a chromosome of the current subpopulation, and each delay time is used as a gene; a fast genetic algorithm is used to generate offspring by introducing the fitness of each gene in the chromosome for each subpopulation in turn, and a delay variable of each flight is obtained according to the flight conflict information, and mutation is performed based on the delay variable to complete a genetic evolution, and an optimized subpopulation is obtained after multiple evolutions; The solution with the largest fitness in each optimized subpopulation is taken as the second delay time of the corresponding flight, and the first delay time of each flight is updated to the second delay time. The departure time and arrival time of each flight are updated again, and flight conflicts are detected, grouped, and genetically evolved, and the first delay time is updated. The first delay time obtained after multiple cycles is used to adjust the departure time of the corresponding flight. The fitness of each gene is calculated by the following formula: in, is the lth chromosome in the kth subpopulation The jth gene on The fitness of NC kj is the total number of conflicts in the flights corresponding to the jth gene in the kth subpopulation, δ max is the maximum flight delay time.

2. According to the large-scale flight control method based on multi-objective collaborative optimization according to claim 1, It is characterized in that The acquisition of flight conflict information by detecting flight conflicts at each time includes: Based on the condition that each flight has the same flight altitude and speed, the flight path is divided at fixed time intervals to obtain a state vector of each flight at each moment; based on the state vector of each flight, whether any two flights have a flight conflict at each moment is identified in turn, and flight conflict information is obtained by statistics, including the total number of conflicts at each moment and the total number of conflicts for each flight.

3. According to the large-scale flight control method based on multi-objective collaborative optimization according to claim 2, It is characterized in that The state vector of each flight at each time includes a horizontal coordinate and a vertical coordinate, a flight altitude, and a corresponding time; The step of sequentially identifying whether any two flights at each moment have a flight conflict according to the state vector of each flight includes: At each time, add the square of the difference between the horizontal coordinates and the vertical coordinates of any two flights and then take the square root to get the interval distance; If the interval distance is less than the distance threshold, a flight conflict will occur between the corresponding two flights; otherwise, no flight conflict will occur.

4. According to claim 3, the large-scale flight control method based on multi-objective collaborative optimization, It is characterized in that According to the overlap of flight time between any two flights, multiple flight groups are obtained by grouping, including: Based on the flight data, sort all flights and add the first flight to the first flight group; Starting from the second flight, one flight is taken out in turn as the current flight, and the intersection of the flight time periods of the current flight and the remaining flights is calculated. The flight with the largest intersection length is obtained as the flight to be grouped; If the flight to be grouped has joined a flight group, the current flight is added to the same flight group as the flight to be grouped; otherwise, a new flight group is created and the current flight is added to the new flight group; This process continues until all flights have joined a flight group, at which point the grouping ends and multiple flight groups are obtained.

5. The large-scale flight regulation method based on multi-objective collaborative optimization according to claim 4, characterized in that The delay time vector of each flight is {0,ts,2×ts,3×ts,...,δ max } as the selection range of delay time, randomly select S times from it, and get the delay time vector Among them, 1≤i≤N, N is the total number of flights in the flight data, ts is the sampling time, δ max is the maximum flight delay time and is divisible by ts.

6. The large-scale flight regulation method based on multi-objective collaborative optimization according to claim 5, characterized in that obtaining the delay variable of each flight according to the flight conflict information includes: Based on the total number of conflicts at each moment in the flight conflict information, the maximum slope value, the minimum slope value, and the slope value at each moment are obtained according to the difference in the total number of conflicts between every two adjacent moments; According to the flight corresponding to each gene on the chromosome during the genetic evolution process as the flight to be mutated, and according to the updated takeoff time of the flight to be mutated, the slope value of the flight to be mutated at the corresponding moment is obtained; If the slope value of the flight to be mutated is less than 0, the delay variable of the flight to be mutated is the opposite of the ratio of the slope value of the flight to be mutated to the minimum slope value; otherwise, the delay variable of the flight to be mutated is the ratio of the slope value of the flight to be mutated to the maximum slope value.

7. The large-scale flight regulation method based on multi-objective collaborative optimization according to claim 6, characterized in that mutating based on the delay variable includes: According to the total number of conflicts of each flight in the current sub-population, the maximum total number of conflicts is obtained; A conflict threshold is randomly selected within the range less than or equal to the maximum total number of conflicts; It is sequentially determined whether the total number of conflicts of the flight to be mutated corresponding to the current gene is greater than the conflict threshold. If it is greater, a mutation value is obtained according to the delay variable of the flight to be mutated, and the current gene value is updated to the mutation value; otherwise, the current gene value remains unchanged.

8. The large-scale flight regulation method based on multi-objective collaborative optimization according to claim 7, characterized in that obtaining the mutation value according to the delay variable of the flight to be mutated includes: When the delay variable of the flight to be mutated is greater than rand(0,1), a value is randomly selected within the range less than or equal to the current gene as the mutation value; When the delay variable of the flight to be mutated is less than -rand(0,1), a value is randomly selected within the range between the current gene and the maximum flight delay time as the mutation value; When the delay variable of the flight to be mutated meets the remaining conditions, a value is randomly selected within the range less than or equal to the maximum flight delay time as the mutation value.

9. The large-scale flight regulation method based on multi-objective collaborative optimization according to claim 5 or 8, characterized in that The fitness is calculated according to the following formula: in, is the fitness of the lth chromosome in the kth subpopulation, m k is the total number of flights in the kth subpopulation; δ max is the maximum flight delay time, is the jth gene on the lth chromosome in the kth subpopulation, NC kj is the total number of conflicts in the flights corresponding to the jth gene in the kth subpopulation.

Images