A flight schedule coordination period division method, device and storage medium

Abstract

The present invention provides a flight schedule coordination period division method, device and storage medium, comprising: collecting flight operation data, calculating flight operation efficiency indicators, including: flight departure delay, flight arrival delay, flight execution rate, and flight regularity; constructing a time series with a day as the granularity according to the flight operation efficiency indicators; using a time series similarity measurement method to measure the similarity between monthly time series; establishing a flight schedule coordination period division calculation model according to the time series similarity measurement result, and dividing the flight schedule coordination period of an airport or an airport group; using a TOPSIS method with an entropy value to evaluate each division result, obtaining the optimal division result of the flight schedule coordination period of different research objects, and optimizing the current flight season division method to meet the management demand of flight time slot allocation.

Description

Technical Field

[0001] The invention relates to the field of air traffic season division, and in particular to a flight time coordination period division method, device and storage medium. Background Art

[0002] With the rapid development of my country's civil aviation industry, aviation demand has shown a rapid growth trend. However, my country's civil aviation industry is still facing two major challenges: insufficient resource reserves and a sharp increase in demand. Although my country has made remarkable achievements in air traffic management construction in recent years, the contradiction between the insufficient supply-side guarantee capacity of civil aviation and the incompatibility with the needs of the rapid development of the civil aviation market still exists. The service quality with flight regularity as the core needs to continue to improve. The reasonable division of flight schedule coordination cycles will help promote the reasonable allocation of flight schedule resources and improve flight punctuality.

[0003] At present, there are few studies on the flight schedule coordination cycle, that is, the flight season, in China. Abde-laty et al. detected the daily, weekly and seasonal airport arrival delay cycle patterns at Orlando Airport, and on this basis, used multinomial logistic regression and binary logistic regression models to study the impact of different factors on average daily delays and single flight delays. Inniss et al. proposed a 1-parameter arrival capacity distribution model, combined with the seasonal clustering method, and used the integer programming model to divide the seasonal cycle. Taking San Francisco International Airport as an example, the bad weather in the morning was converted into the probability distribution of arrival capacity. This model is suitable for airport capacity modeling with specific weather types occurring at specific times, and cannot reflect the impact of predicted weather on capacity. The method has a large amount of calculation and is difficult to use. Based on Inniss, Zhang Jing designed a time series seasonal division method based on genetic algorithms, identifying months with similar capacity probability distribution characteristics in the time series as one season. Finally, the seasonal division of common weather effects at San Francisco Airport in the United States was obtained.

[0004] It can be seen that some scholars have noticed the significance of reasonably dividing the flight schedule coordination cycle in promoting the rational allocation of flight schedule resources, and have conducted research using airport capacity as a research indicator and San Francisco Airport in the United States as a research object. However, the above studies mostly take foreign airports as examples, and have not yet involved research on the division of my country's flight coordination cycle. Moreover, the above studies are all based on a single dimensional standard for division, ignoring the significance of different indicators for the division of flight seasons.

[0005] Therefore, how to overcome the above problems is currently in urgent need of resolution. Summary of the invention

[0006] Purpose of the invention: The technical problem to be solved by the present invention is to provide a method, device and storage medium for dividing a flight schedule coordination period in view of the deficiencies of the prior art. The method comprises:

[0007] Collect flight operation data and calculate various indicators of flight operation efficiency;

[0008] Taking the day as the time granularity, construct the monthly time series of flight operation efficiency indicators;

[0009] Use the Dynamic Time Warping (DTW) algorithm to measure the similarity of time series;

[0010] According to the time series similarity measurement results, a flight schedule coordination period division model is established to divide the flight schedule coordination period;

[0011] TOPSIS (Technique for Order Preference by Similarity to an Ideal Solution) with entropy value is used to evaluate the solved flight schedule division periods and determine the optimal flight coordination schedule period.

[0012] The collecting of flight operation data and calculation of various flight operation efficiency indicators include: selecting and collecting flight scheduled departure time, flight scheduled arrival time, number of planned flights per unit time, number of flights actually executed per unit time, and number of normal flights per unit time according to the air traffic operation level similarity measurement requirements; and then calculating various flight operation efficiency indicators according to the following formula:

[0013] Flight departure delay index, i.e. the difference between the actual departure time and the planned departure time of flights at an airport / airport group, is calculated as:

[0014]

[0015] Where depdelay is the departure delay of a single flight, ATD is the actual departure time of the flight, and ETD is the scheduled departure time of the flight;

[0016] The flight arrival delay index is the difference between the actual arrival time and the planned arrival time of flights at an airport / airport group. The calculation formula is:

[0017]

[0018] Where arrdelay is the arrival delay of a single flight, ATA is the actual arrival time of the flight, and ETA is the planned arrival time of the flight;

[0019] The flight execution rate index is the ratio of the number of flights actually executed by an airport / airport group to the number of flights planned to be executed. The calculation formula is:

[0020]

[0021] Where e-rate is the flight execution rate within the statistical time granularity, f actual To count the number of flights actually executed within the time granularity, f schedule To count the number of flights planned to be executed within the statistical time granularity;

[0022] The flight regularity index, that is, the ratio of the number of regular flights to the number of scheduled flights at an airport / airport group, is calculated as follows:

[0023]

[0024] Where r-rate is the flight regularity within the statistical time granularity, f regular is the number of normal flights within the statistical time granularity, f schedule To count the number of flights planned to be executed within the time granularity. According to the Civil Aviation Administration of China's "Civil Aviation Flight Normal Statistics Method (2013)", a flight that meets one of the following conditions can be considered a normal flight:

[0025] The aircraft is unblocked within 5 minutes before or after the scheduled departure time and takes off according to the forward flight process without any special circumstances such as taxiing back, aborted takeoff, return flight or alternate landing;

[0026] Block the ship no later than the planned arrival time.

[0027] The method of constructing the monthly time series of the operating efficiency index with the day as the time granularity includes: calculating the daily average value of each flight operating efficiency index with the day as the time granularity, and constructing the monthly time series of each flight operating efficiency index. For example, the time series of flight departure delay time in January is: X = {x1, x2..., x 31}, where x1 is the national average departure delay time on January 1.

[0028] The method of measuring the similarity of time series by using the dynamic time warping algorithm includes: selecting the monthly time series data of each indicator of flight operation efficiency, constructing a time series distance matrix, each element in the matrix represents the distance between two corresponding points in the monthly time series of each indicator of flight operation efficiency, and calculating the distance between each time series by the following formula according to the dynamic time warping distance measurement method:

[0029]

[0030] Where D dtw (i,j) is the DTW distance between sequences X and Y, xi is the i-th point in time series X, y j represents the jth point in the time series Y, i and j are the order of the points in the time series X and Y respectively, Dist(x i ,y j ) represents the time series distance matrix x i With y j The distance between.

[0031] The method of establishing a flight schedule coordination period division model based on the time series similarity measurement result to divide the flight schedule coordination period includes:

[0032] Step a1, defining parameters;

[0033] Step a2, constructing the objective function;

[0034] Step a3, setting constraints, and finally forming a flight schedule coordination period division model by the objective function and constraints;

[0035] Step a1 includes: defining the following parameters:

[0036] m: number of divided seasons;

[0037] n s : The number of months included in the sth season;

[0038] X sab : The DTW distance between month a and month b in season s, i.e. the distance between two points measured using the dynamic time warping algorithm

[0039] X ijkl : The DTW distance between the jth month in the i-th season and the lth month in the k-th season.

[0040] Step a2 includes: the objective function is:

[0041] The first objective function is to minimize the average intra-seasonal distance: n s The number of types of seasons and pairwise combinations;

[0042] The second objective function is to maximize the average seasonal distance:

[0043] The third objective function is to minimize the ratio of the average intra-season distance to the inter-season distance:

[0044] Step a3 includes: The constraints are:

[0045] The first constraint is to divide the system into at least 2 seasons: m ≥ 2;

[0046] The second constraint is that each quarter contains at least two months: n s ≥2;

[0047] The third constraint is that each division must be a complete division of 12 months:

[0048] The TOPSIS method with entropy value is used to evaluate the solved flight time division periods to determine the optimal flight coordination time period, including:

[0049] Step b1, substituting the constraint conditions into the objective function and calculating the optimal value of the objective function;

[0050] Step b2, using entropy method to define the weights of each objective function;

[0051] Step b3, using the TOPSIS method with entropy value to evaluate each Pareto optimal solution;

[0052] Wherein, step b1 includes: using a heuristic algorithm to solve, such as a non-dominated genetic algorithm with an elite strategy to solve the objective function:

[0053] Each solution individual is set to a 0 or 1 vector of length 12, where 1 represents the first month in a season and 0 represents joining the previous season. The optimal Pareto solution set of the objective function is found through the genetic algorithm. The iteration number is generally set according to the population size, which is set to 1000 here. The iteration stops when the set iteration number is reached.

[0054] Step b2 includes:

[0055] Perform data standardization and unify the dimensions of each objective function value in each Pareto solution set. The formula is:

[0056] Positive indicators, that is, the larger the value, the better the objective function value:

[0057] Reverse indicator, that is, the smaller the value, the better the objective function value:

[0058] where x ij is the index value after standardization. is the original index value, is the maximum value of the j-th type objective function, is the minimum value of the objective function of the jth class.

[0059] Calculate the weight of each solution and establish the weight matrix. The formula is:

[0060]

[0061] p ij is the proportion of the i-th Pareto solution in the j-th objective function;

[0062] Calculate the entropy value of each objective function, the formula is:

[0063]

[0064] where e jis the entropy value of the j-th type objective function, and m is the number of Pareto solutions;

[0065] Step b3 includes: setting the weight of each objective function according to the entropy value of each objective function, the formula is:

[0066]

[0067] According to the jth weight w j , construct a weighted matrix for each objective function value, the formula is:

[0068]

[0069] is the weighted value;

[0070] In the weighted matrix, find the optimal solution Worst solution The formula is:

[0071]

[0072] Calculate the optimal distance Worst distance The formula is:

[0073]

[0074] Calculate the relative closeness C of each Pareto solution to the optimal objective function i :

[0075]

[0076] The present invention also provides a flight schedule coordination period division device, comprising:

[0077] The data collection module is used to collect flight operation data and calculate various indicators of flight operation efficiency;

[0078] The flight operation efficiency index construction module is used to construct the monthly time series of flight operation efficiency index with the day as the time granularity;

[0079] The time series similarity measurement module is used to measure the similarity of time series using the dynamic time warping algorithm;

[0080] The flight schedule coordination period division module is used to establish a flight schedule coordination period division model to divide the flight schedule coordination period according to the time series similarity measurement results;

[0081] The optimal flight coordination time period calculation module is used to evaluate the solved flight time division periods using the TOPSIS method with entropy value to determine the optimal flight coordination time period.

[0082] The present invention also provides a storage medium storing a computer program or instruction, and when the computer program or instruction is executed, the method described is implemented.

[0083] The beneficial effect of the present invention is that the present invention relates to a flight schedule coordination cycle division method based on similarity. The flight schedule coordination cycle division method based on similarity includes: collecting flight operation data, calculating flight operation efficiency indicators, including: flight departure delays, flight arrival delays, flight execution rate, flight regularity; constructing a time series with a daily granularity based on the flight operation efficiency indicators; using a time series similarity measurement method to measure the similarity between monthly time series; establishing a flight schedule coordination cycle division calculation model based on the time series similarity measurement results, and dividing the flight schedule coordination cycle of airports and airport groups; using the TOPSIS method with entropy value to evaluate each division result, and obtain the optimal division result of the flight schedule coordination cycle of different research objects, and optimize the current flight season division method to meet the management needs of flight time slot allocation. BRIEF DESCRIPTION OF THE DRAWINGS

[0084] The present invention will be further described in detail below in conjunction with the accompanying drawings and specific embodiments, and the above and / or other advantages of the present invention will become more clear.

[0085] Figure 1 is a flow chart of the method of the present invention.

[0086] Figure 2a , Figure 2b Schematic diagram of the dynamic time rule algorithm principle.

[0087] Figure 3 Schematic diagram of the principle of solving the model of non-dominated genetic algorithm with elite strategy.

[0088] Figure 4 Solve the problem for a non-dominated genetic algorithm with an elitist strategy.

[0089] Figure 5 Graph of the iterative process for solving a non-dominated genetic algorithm with an elitist strategy. DETAILED DESCRIPTION

[0090] Example

[0091] See also Figure 1 The present invention provides a method for dividing a flight schedule coordination cycle based on similarity. The method for dividing a flight schedule coordination cycle based on similarity comprises the following steps:

[0092] S110: Collect flight operation data and calculate various flight operation efficiency indicators.

[0093] S120: Construct a time series of flight operation efficiency with daily granularity.

[0094] S130: Calculate the similarity of the time series of each month based on the time series of the operating efficiency of each month by using a time series similarity measurement method.

[0095] S140: Based on the time series similarity measurement result, a flight schedule coordination period division model is used to obtain a division result.

[0096] S150: Using the TOPSIS method with entropy value to evaluate each partition result and determine the optimal partition result.

[0097] In this embodiment, step S110 includes: flight departure delay, flight arrival delay, flight execution rate, flight regularity; when the average value of each flight operation efficiency index is calculated at the daily granularity, there are a small number of missing values. Considering the similarity between adjacent days of flight operation conditions, the previous value filling method is used to fill the missing data.

[0098] In this embodiment, step S130 uses a dynamic time warping algorithm to measure the distance between each time series. The schematic diagram of the algorithm principle is shown in FIG. Figure 2a , Figure 2b As shown, Figure 2a The horizontal axis is the subscript of the midpoint of the two time series, and the vertical axis is the DTW distance. Figure 2b The horizontal and vertical coordinates are the subscripts of the midpoints of the two time series, and the calculation formula is:

[0099]

[0100] In this embodiment, step S140 includes the following steps:

[0101] S141: Parameter definition:

[0102] m divides the number of seasons;

[0103] n s The number of months included in season s;

[0104] X sab The DTW distance between month a and month b in season s;

[0105] X ijkl The DTW distance between month j in season i and month l in season k.

[0106] S142: Construct objective function:

[0107] Objective function 1 - Minimum average intra-season distance:

[0108] Objective function 2 - Maximum average seasonal distance:

[0109] Objective function 3: The ratio of the average intra-season distance to the inter-season distance is minimized:

[0110] S143: Set constraints:

[0111] Constraint 1: Divide into at least 2 seasons: m ≥ 2

[0112] Constraint 2—Each quarter must contain at least two months: n s ≥2

[0113] Constraint 3 - Each division must be a complete division of 12 months:

[0114] S144: Substitute the constraint conditions into the objective function and calculate the optimal value of the objective function.

[0115] Set each solution individual to a 0-1 vector of length 12, where 1 represents the first month in a season and 0 represents joining the previous season. Use the genetic algorithm to find the optimal Pareto solution set of the objective function, and stop iterating when the set iteration number is reached. The algorithm principle is as follows Figure 3 As shown, the iterative process is as follows Figure 5 The final result is as shown in Figure 4 shown.

[0116] HV is the hypervolume index, which indicates the area enclosed by the non-dominated solution set obtained by the algorithm and the reference point. It is one of the indicators for evaluating the convergence and diversity of the algorithm. The larger its value, the better the algorithm performance. The calculation method is: Where n represents the number of individuals in the Pareto optimal solution set of the algorithm; R i is a rectangle constructed with each element in the Pareto optimal solution and the reference point as diagonal vertices.

[0117] S150: Use TOPSIS with entropy value to evaluate the Pareto solution and determine the solution closest to the best result.

[0118] S151: Data standardization processing, unifying the dimensions of each objective function value in each Pareto solution set.

[0119] Positive indicators, that is, the larger the value, the better the objective function value: Reverse indicator, that is, the smaller the value, the better the objective function value:

[0120] S152: Calculate the weight of each solution and establish a weight matrix.

[0121]

[0122] S153: Calculate the entropy value of each objective function.

[0123]

[0124] S154: According to the entropy value of each objective function, set the weight of each objective function.

[0125]

[0126] S155: Construct a weighting matrix.

[0127]

[0128] S156: Find the best and worst solutions.

[0129]

[0130] S157: Calculate the optimal and the worst distances.

[0131]

[0132] S158: Calculate relative proximity.

[0133]

[0134] In summary, the present invention relates to a flight schedule coordination cycle division method based on similarity. The flight schedule coordination cycle division method based on similarity includes: collecting flight operation data, calculating flight operation efficiency indicators, including: flight departure delays, flight arrival delays, flight execution rate, and flight regularity; constructing a time series with a daily granularity based on the flight operation efficiency indicators; using a time series similarity measurement method to measure the similarity between monthly time series; establishing a flight schedule coordination cycle division calculation model based on the time series similarity measurement results, and dividing the flight schedule coordination cycle of airports and airport groups; using the TOPSIS method with entropy value to evaluate each division result, and obtain the optimal division result of the flight schedule coordination cycle of different research objects, and optimize the current flight season division method to meet the management needs of flight time slot allocation.

[0135] The present invention also provides a flight schedule coordination period division device, comprising:

[0136] The data collection module is used to collect flight operation data and calculate various indicators of flight operation efficiency;

[0137] The flight operation efficiency index construction module is used to construct the monthly time series of flight operation efficiency index with the day as the time granularity;

[0138] The time series similarity measurement module is used to measure the similarity of time series using the dynamic time warping algorithm;

[0139] The flight schedule coordination period division module is used to establish a flight schedule coordination period division model to divide the flight schedule coordination period according to the time series similarity measurement results;

[0140] The optimal flight coordination time period calculation module is used to evaluate the solved flight time division periods using the TOPSIS method with entropy value to determine the optimal flight coordination time period.

[0141] The present invention also provides a storage medium storing a computer program or instruction, and when the computer program or instruction is executed, the method described is implemented.

[0142] As described above, the apparatus according to the embodiment of the present application can be implemented in various terminal devices, such as a server of a distributed computing system. In one example, the apparatus according to the embodiment of the present application can be integrated into the terminal device as a software module and / or a hardware module. For example, the apparatus can be a software module in the operating system of the terminal device, or can be an application developed for the terminal device; of course, the apparatus can also be one of the many hardware modules of the terminal device.

[0143] Alternatively, in another example, the apparatus and the terminal device may also be separate terminal devices, and the apparatus may be connected to the terminal device via a wired and / or wireless network, and transmit the interactive information in accordance with an agreed data format.

[0144] The present invention provides a method, device and storage medium for dividing a flight schedule coordination period. There are many methods and ways to implement the technical solution. The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principle of the present invention. These improvements and modifications should also be regarded as the protection scope of the present invention. All components not specified in this embodiment can be implemented by existing technologies.

Claims

1. A method for dividing a flight schedule coordination period, characterized in that: include: Collect flight operation data and calculate various indicators of flight operation efficiency; According to the requirements of similarity measurement of air traffic operation level, the planned departure time of flights, the planned arrival time of flights, the number of planned flights per unit time, the number of flights actually executed per unit time and the number of normal flights per unit time are selected for collection; Taking the day as the time granularity, construct the monthly time series of flight operation efficiency indicators; Use dynamic time warping algorithm to measure time series similarity; According to the time series similarity measurement results, a flight schedule coordination period division model is established to divide the flight schedule coordination period, including: Step a1, defining parameters; Step a2, constructing the objective function; Step a3, setting constraints; Wherein, step a1 includes: defining the following parameters: m: number of divided seasons; n s : The number of months included in the sth season; X sab : The DTW distance between month a and month b in season s; X ijkl : The DTW distance between the jth month in the i-th season and the lth month in the k-th season; Step a2 includes: the objective function is: The first objective function is to minimize the average intra-seasonal distance: n s The number of types of seasons and pairwise combinations; The second objective function is to maximize the average seasonal distance: The third objective function is to minimize the ratio of the average intra-season distance to the inter-season distance: Step a3 includes: The constraints are: The first constraint is to divide the system into at least 2 seasons: m ≥ 2; The second constraint is that each quarter contains at least two months: n s ≥2; The third constraint is that each division must be a complete division of 12 months: The TOPSIS method with entropy value is used to evaluate the solved flight time division cycle to determine the optimal flight coordination time cycle, including: Step b1, substituting the constraint conditions into the objective function and calculating the optimal value of the objective function; Step b2, using entropy method to define the weights of each objective function; Step b3, using the TOPSIS method with entropy value to evaluate each Pareto optimal solution; Wherein, step b1 includes: setting each solution individual to a 0 or 1 vector of length 12, where 1 represents the first month in a season and 0 represents joining the previous season; The optimal Pareto solution set of the objective function is found through the genetic algorithm. The iteration is stopped when the set iteration number is reached. HV is the hypervolume index, which represents the area enclosed by the non-dominated solution set obtained by the algorithm and the reference point. The calculation method is: Where n represents the number of individuals in the Pareto optimal solution set of the algorithm; R i is a rectangle constructed with each element in the Pareto optimal solution and the reference point as diagonal vertices.

2. The method according to claim 1, characterized in that The flight operation efficiency indicators are calculated according to the following formula: Flight departure delay index, calculated as: Where depdelay is the departure delay of a single flight, ATD is the actual departure time of the flight, and ETD is the scheduled departure time of the flight; Flight arrival delay index, calculated as: Where arrdelay is the arrival delay of a single flight, ATA is the actual arrival time of the flight, and ETA is the planned arrival time of the flight; Flight execution rate index, calculated as follows: Where e-rate is the flight execution rate within the statistical time granularity, f actual To count the number of flights actually executed within the time granularity, f schedule To count the number of flights planned to be executed within the statistical time granularity; Flight regularity index, calculated as follows: Where r-rate is the flight regularity within the statistical time granularity, f regular is the number of normal flights within the statistical time granularity, f schedule The number of flights planned to be executed within the statistical time granularity.

3. The method according to claim 2, characterized in that The constructing of monthly time series of operation efficiency indicators with day as the time granularity includes: calculating daily average values ​​of flight operation efficiency indicators with day as the time granularity, and constructing monthly time series of flight operation efficiency indicators.

4. The method according to claim 3, characterized in that The method of measuring the similarity of time series by using the dynamic time warping algorithm includes: selecting the monthly time series data of each indicator of flight operation efficiency, constructing a time series distance matrix, each element in the matrix represents the distance between two corresponding points in the monthly time series of each indicator of flight operation efficiency, and calculating the distance between each time series by the following formula according to the dynamic time warping distance measurement method: Where D dtw (i,j) is the DTW distance between sequences X and Y, xi is the i-th point in time series X, y j represents the jth point in the time series Y, i and j are the order of the points in the time series X and Y respectively, Dist(x i ,y j ) represents the time series distance matrix x i With y j The distance between.

5. The method according to claim 4, characterized in that Step b2 includes: Perform data standardization and unify the dimensions of each objective function value in each Pareto solution set. The formula is: Positive indicators, that is, the larger the value, the better the objective function value: Reverse indicator, that is, the smaller the value, the better the objective function value: where x ij is the index value after standardization. is the original index value, is the maximum value of the j-th type objective function, is the minimum value of the objective function of the jth class; Calculate the weight of each solution and establish the weight matrix. The formula is: p ij is the proportion of the i-th Pareto solution in the j-th objective function; Calculate the entropy value of each objective function, the formula is: where e j is the entropy value of the j-th type objective function, and m is the number of Pareto solutions; Step b3 includes: setting the weight of each objective function according to the entropy value of each objective function, the formula is: According to the jth weight w j , construct a weighted matrix for each objective function value, the formula is: is the weighted value; In the weighted matrix, find the optimal solution Worst solution The formula is: Calculate the optimal distance Worst distance The formula is: Calculate the relative closeness C of each Pareto solution to the optimal objective function i :

6. A flight schedule coordination period division device implemented according to the method of any one of claims 1 to 5, characterized in that: include: The data collection module is used to collect flight operation data and calculate various indicators of flight operation efficiency; The flight operation efficiency index construction module is used to construct the monthly time series of flight operation efficiency index with the day as the time granularity; The time series similarity measurement module is used to measure the similarity of time series using the dynamic time warping algorithm; The flight schedule coordination period division module is used to establish a flight schedule coordination period division model to divide the flight schedule coordination period according to the time series similarity measurement results; The optimal flight coordination time period calculation module is used to evaluate the solved flight time division periods using the TOPSIS method with entropy value to determine the optimal flight coordination time period.

7. A storage medium, characterized in that: A computer program or instruction is stored, and when the computer program or instruction is executed, the method according to any one of claims 1 to 5 is implemented.

Images