Airport flight plan strategy delay prediction method based on delay disassembly and assembly

Abstract

The embodiment of the invention discloses an airport flight plan strategy delay prediction method based on delay disassembly and assembly, relates to the technical field of air control and planning, and can timely and effectively predict the delay of arrival and departure of an airport flight plan. The airport flight plan strategy delay prediction method based on delay disassembly and assembly is used for obtaining an airport capacity probability graph by using an airport flight plan and historical flight operation data, and obtaining an airport flight arrival and departure delay prediction value according to the airport flight plan, airport operation mode modeling and the airport capacity probability graph. The method is suitable for evaluating the rationality and effectiveness of airport flight time / planning at the strategic stage.

Description

Technical Field

[0001] The invention relates to the technical field of air traffic control and planning, and in particular to a delay prediction method for airport flight plan strategy based on delay decomposition. Background Art

[0002] With the increase in total civil aviation transport volume, flight delays have once again become the main contradiction restricting the development of the civil aviation transport industry. With the continuous increase in aviation market demand, airports have exposed the constraints of insufficient guarantee capacity, and flight delays have become inevitable under the influence of multiple factors.

[0003] From a global perspective, the problem of flight delays will exist for a long time and will become a roadblock to the development of the civil aviation industry. Therefore, predicting the level of flight delays and adjusting the corresponding flight plans, as well as preparing and warning for flight delays in both strategy and tactics, have become key issues in the development of the civil aviation industry. It is of great significance to optimize flight schedules and improve flight schedule operation efficiency, so that the civil aviation industry can optimize and change existing flight plans earlier and propose corresponding strategic plans and response measures.

[0004] Among these, how to timely and effectively predict the delays in airport flight schedules has become a topic that needs to be studied. Summary of the invention

[0005] The embodiment of the present invention provides a method for predicting airport flight plan strategic delay based on delay decomposition, which can timely and effectively predict the delays of airport flight plan arrival and departure.

[0006] To achieve the above object, the embodiments of the present invention adopt the following technical solutions:

[0007] The airport flight planning strategy delay prediction method based on delay decomposition includes:

[0008] S1. Using the airport flight operation data, obtain the clustering results of the airport's incoming and outgoing flights, wherein the flight operation data includes: flight calculated off-block time data and flight actual landing time data. Flight calculated off-block time (calculated off-block time) is a professional term in this field, which refers to the target off-block time TOBT calculated by the airport's minimum guarantee capacity after the aircraft passes the planned off-block time SOBT, and then inputs TOBT into the traffic management system. The COBT time calculated by the system is the flight calculated off-block time data described in the claims. In actual applications, for the airport flight arrival and departure operation data, the Kmeans clustering method is used to cluster the off-block flow profile and the actual landing flow profile, and the clustering centers under various K values ​​are obtained as clustering results.

[0009] S2. Using the clustering results of the airport's inbound and outbound flights, the airport's inbound and outbound capacity probability map is obtained through clustering result evaluation indicators and confidence interval judgment methods; wherein the capacity probability map is represented by a set of capacity cycle curves, which are used to generate random airport departure capacity and airport arrival capacity in different time windows. Specifically, the capacity probability map is a set of mutually fused and separated broken lines in the same coordinate system, wherein the fused broken lines represent fusion scenes formed by the fusion of different scenes, and the independent broken lines represent independent scenes; wherein each scene has a corresponding formation probability, the probability size is equal to the proportion of its corresponding cluster in the number of observations, and the fusion scene probability is the difference between 1 and the remaining independent scenes. In a specific time window, when an independent scene is selected, the inbound and outbound capacity is the flow size value corresponding to the independent scene in the specific time window; when a fusion scene is selected, the inbound and outbound capacity is the weighted value of the flow size value corresponding to each independent scene in the fusion scene in the specific time window, wherein the weight is represented by the quotient of the formation probability of each independent scene in the fusion scene and the sum of the formation probabilities.

[0010] S3. Using the arrival and departure capacity probability map, establish an airport arrival and departure native delay prediction model; wherein, the airport flight schedule delay can be broken down into native delay and flow management delay for analysis. The native delay is a flight delay caused by the airport's own support capabilities, airport facility structure and operation mode. Flow management delay is a delay caused by human factors or some accidents during the airport flow management process. Both types of delays will eventually lead to delays and lags in the airport flight schedule, so they will be recorded and archived during the daily management of the airport.

[0011] S4. Use the airport traffic management data to establish an airport flow control delay prediction model, wherein the airport traffic management data includes: flight departure airport and destination airport, flight target off-block time data and flight calculated off-block time data; specifically, the airport traffic generally refers to the actual take-off traffic of airport flights. The actual take-off traffic profile of the airport flights is expressed as follows: divide the statistical time period into 15 minutes as the time granularity, and count the actual number of flights taking off in each time window as the actual take-off traffic profile of the flights in the time window; the actual block traffic profile of the airport flights is expressed as follows: divide the statistical time period into 15 minutes as the time granularity, and count the actual number of flights in each time window as the actual block traffic profile of the flights in the time window.

[0012] S5. According to the original arrival and departure delay prediction model and the airport flow control delay prediction model, an airport flight schedule delay prediction model is established by using an airport flight delay propagation judgment formula;

[0013] S6. According to the airport flight plan delay prediction model, the airport flight plan to be processed is predicted and processed, and the arrival and departure delay prediction result of the airport flight plan to be processed is obtained.

[0014] Specifically, in S3, it also includes: before using the arrival and departure capacity probability map to establish the airport's arrival and departure native delay prediction model, loading the queuing network model corresponding to the airport operation mode, wherein the airport operation mode includes: runway service mode, airport departure operation mode and airport arrival operation mode; specifically, the queuing network model corresponding to the airport operation mode can be divided into multiple types according to function, such as: flight flow characteristic model, runway take-off time interval distribution model, runway landing time interval distribution model, parking stand wheel block withdrawal time interval distribution model, aircraft taxiing mode model, etc.

[0015] Furthermore, the native arrival and departure delay prediction model includes: a departure native delay prediction model and an arrival native delay prediction model; in the departure delay prediction model, the airport departure capacity is used to predict the flight stand waiting time and the queuing network model is used to predict the flight runway head waiting time; in the arrival delay prediction model, the airport arrival capacity is used to predict the flight arrival taxiing delay and the queuing network model is used to predict the flight terminal area delay time. Among them, the departure delay prediction model is expressed as follows: when the aircraft is at the parking position, the key parameter limiting the aircraft push is the departure capacity size, which is derived from the random capacity scenario generated by the departure capacity probability map. When the aircraft is pushed away and reaches the runway head after taxiing, the key limiting the aircraft takeoff is the runway head service time interval; the departure delay prediction model is a simulation system for each departing aircraft from the calculation of the off-block time to the actual take-off time. The arrival delay prediction model is expressed as follows: when the aircraft enters the terminal area at the earliest, the terminal area delay of the aircraft is predicted by simplifying the queuing network model. After the aircraft lands, the key parameter restricting the aircraft from entering the runway is the arrival capacity, which is derived from the random capacity scenario generated by the arrival capacity probability graph. The arrival delay prediction model is a simulation system that starts from the planned landing time for each arriving aircraft, moves forward in time to obtain the terminal area delay, and moves backward to the end of the actual block time.

[0016] Specifically, in S4, it also includes: before establishing the airport flow control delay prediction model, using the airport traffic management data to perform an airport flow control delay correlation analysis; and establishing the flow control delay prediction model based on the results of the airport flow control delay correlation analysis.

[0017] The airport flow control delay correlation analysis includes: obtaining the quantitative value of the correlation between the size of flow control delay and the size of traffic in different city pairs in different time periods, wherein, if the quantitative value of the correlation is higher than a threshold, the flow control delay prediction model is set as a positive correlation function about the traffic size; if the quantitative value of the correlation is less than or equal to the threshold, the flow control delay prediction model is obtained by statistically analyzing the size of flow control delay and the possibility of flow control delay, wherein the possibility of flow control delay includes: the ratio of the number of flights subject to flow control to the total number of flights for a specific city pair in a specific time period. For example: using airport traffic management data, an airport flow control delay prediction model is established through airport flow control delay correlation analysis. Wherein, based on the results of the airport flow control delay correlation analysis, the airport traffic and the flow control delay model selection direction are determined. Specifically, the airport flow control delay correlation analysis is a correlation test between the size of flow control delays and the size of traffic for different city pairs in different time periods. If the correlation is high, the flow control delay prediction model is a positive correlation function of the traffic size. Otherwise, the flow control delay prediction model is obtained by statistically analyzing the size of flow control delays and the possibility of flow control delays. The possibility of flow control delays is expressed as the ratio of the number of flights subject to flow control to the total number of flights for a specific city pair in a specific time period. For example: For a correlation test between the size of flow control delays and the size of traffic for different city pairs in different time periods, if the correlation is higher than a threshold, the flow control delay prediction model is a positive correlation function of the traffic size. Otherwise, the flow control delay prediction model is obtained by statistically analyzing the average size of flow control delays and the possibility of flow control delays. The threshold is generally 0.5. The possibility of flow control delays is expressed as the ratio of the number of flights subject to flow control to the total number of flights for a specific city pair in a specific time period. Specifically, the flow control delay size is expressed as the sum of the flow control delays in a specific city within a specific time period; the average flow control delay size is expressed as the flow control delay size and the number of flights affected by the flow control.

[0018] Specifically, in S5, it includes: using the arrival native delay prediction model to generate the arrival delay prediction value of the airport flight plan, and then converting the arrival delay prediction value of the airport flight plan into the departure propagation delay prediction value; generating the flight flow control delay through the flow control delay prediction model, and then using the departure native delay prediction model to generate the departure delay prediction value of the flight plan. For example:. Wherein, the flight plan at least includes: time, aircraft registration number, planned off-block time, planned on-block time, departure airport, and destination airport. The airport flight delay propagation judgment formula includes: aircraft registration number, flight number, aircraft type, actual block time and planned on-block time of the preceding arrival flight, planned off-block time of the succeeding departure flight, and the shortest aircraft transit time.

[0019] Furthermore, in S5, it also includes: comparing the difference between the stopover time of a specific flight at an airport and the minimum stopover time with the arrival delay of the specific flight. If the former is larger, it is determined that the arrival delay of the specific flight will not be propagated to subsequent flights, otherwise the subsequent flights of the specific flight will generate delay propagation. The size of the propagated delay is equal to the difference between the difference and the arrival delay of the flight. For example: the airport delay propagation judgment formula includes: comparing the difference between the stopover time of a specific flight at an airport and the minimum stopover time with the arrival delay of the flight. If the former is larger, the arrival delay of the flight will not be propagated to subsequent flights, otherwise the subsequent flights of the flight will generate delay propagation, and the size is equal to the difference between the difference and the arrival delay of the flight. Specifically, the flight arrival delay is expressed as AD i =AIBT i –SIBT i , of which AIBT i is the actual block time of the flight, SIBT i is the flight's scheduled block time; the difference between the stopover time of the specific flight at the airport and the minimum stopover time is RDCT i =SOBT i+1 –SIBT i –MTTT i , among which SOBT i+1 is the scheduled off-block time of the subsequent flight, MTTT i is the shortest stopover time of the preceding flight; the delay propagation size is DP i =AD i -RDCT i , if DP i <0, there is no delay propagation, otherwise the target off-block time (TOBT) of the subsequent flight i+1 =DP i +SOBT i+1 .

[0020] Specifically, in S6, the flight plan to be processed includes: the date of the flight to be processed, the departure airport and destination airport of the flight to be processed, the flight plan off-block time of the flight to be processed, and the flight plan on-block time of the flight to be processed. The arrival and departure delay prediction results of the flight plan to be processed include: the average delay of arrival flights, the average flow control delay of departure flights, the average delay of departure flights, and the average total delay of departure flights.

[0021] The average delay of incoming flights is the ratio of total incoming flights to the total number of incoming flights; the average flow control delay of outgoing flights is the ratio of total flow control delays to the total number of outgoing flights; the average delay of outgoing flights is the ratio of total outgoing flight delays to the total number of outgoing flights; the average total delay of outgoing flights is the sum of the average flow control delays of outgoing flights and the average delay of outgoing flights.

[0022] Specifically, in S1, the clustering results of the airport's arriving and departing flights include: the flight departure flow clustering results and the flight arrival flow clustering results; the flight departure flow clustering results are: the clustering center results of the airport's actual flight takeoff flow profile using the Kmeans clustering method; the flight arrival flow clustering results are: the clustering center results of the airport's actual flight block flow profile using the Kmeans clustering method. For example: the airport's arriving and departing flight clustering results include: the flight departure flow clustering results, and the flight arrival flow clustering results. The clustering method is Kmeans clustering, which is a data partitioning algorithm based on Euclidean distance, expressed as: dividing the data set X into k clusters, where k is a positive integer of moderate size. For a specific cluster, the clustering between the observations in the cluster should be smaller than the distance between the value and the observations in other clusters. For each cluster, the center point of all observations in the cluster. k=1,2…K, also called The centroid of cluster k, represented by in is the qth observation in cluster k, and n is the number of observations in cluster k. The Kmeans algorithm performs a Euclidean distance test on each observation x and each centroid, and assigns x to the cluster corresponding to the nearest centroid. If the sum of the distances between each observation and the cluster to which it belongs is the smallest, the algorithm iteration ends.

[0023] In this embodiment, the clustering object is the flow profile of airport flights. Specifically, the departure flow profile is expressed as follows: the actual departure time of daily departure flights is taken, and the number of departure flights is counted in every 15-minute time window to obtain the flow profile data set X = {x1, x2…, x n}, where x i represents the number of departing flights in the i-th time window; the inbound flow profile is expressed as: taking the actual block time of the daily inbound flights, counting the number of inbound flights in every 15-minute time window, and obtaining the flow profile data set. The clustering result is expressed as: in the same coordinate system, with the time window as the x-axis and the flow size in the unit time window as the y-axis, the continuous broken line formed by different cluster centers.

[0024] The clustering results contain results with cluster numbers ranging from 3 to 16, and the clustering results need to be evaluated and screened. The clustering result evaluation indicators include: silhouette coefficient, variance ratio criterion, and pruning method. For example: the silhouette coefficient is the similarity between a single observation and other observations in the cluster to which it is assigned, expressed as: s(x i ) is the silhouette coefficient of each observation. The silhouette coefficient of an observation x is expressed as: Where a(x i ) is the average distance from observation x to all other data points in its cluster, b(x i ) is the average distance between the observation x and the observations in the adjacent clusters; Where d(x j ,x i ) is the distance between observations i and j, and |Ci| is the number of observations in the cluster. Based on the silhouette coefficient s(x i ), the average value of the silhouette coefficient SI close to 1 indicates that the clusters are compact and well separated, close to 0 indicates overlapping clusters, and close to -1 indicates too many or too few clusters.

[0025] The variance ratio criterion is the distance ratio between the inter-cluster error variance and the intra-cluster error variance, expressed as: Among them, the inter-cluster error variance is expressed as: where n i is the number of data points in cluster i, c i is the centroid of cluster i, c is the overall centroid of all data points, and the intra-cluster error variance is expressed as: x is all observations in cluster i, c i It is the centroid of cluster i. The higher the CHI value, the better the clustering effect. The pruning method is used to indicate the removal of unimportant weights in the model, so as to reduce the model parameters and calculation amount without affecting the model accuracy. Specifically, if the statistical single cluster accounts for less than a certain threshold in the clustering result of the inbound and outbound flights of the airport, the clustering result will not be adopted. The pruning method is to remove unimportant weights in the model, so as to reduce the model parameters and calculation amount while ensuring that the model accuracy is not affected. Specifically, the proportion of each cluster in the clustering result is statistically calculated. If a cluster with a proportion less than the threshold appears, it means that the pruning involves a cluster with smaller features and less influence, and the size of k will no longer be increased. The threshold is generally 5%-10%.

[0026] Specifically, S2 includes: using the clustering results of the airport's arrival and departure flights, and obtaining the airport's arrival and departure capacity probability map through clustering result evaluation indicators and confidence interval judgment methods; the confidence interval judgment method includes: within a given time window, comparing the cluster center results with the confidence interval boundaries obtained using the airport's arrival and departure flight clustering results. If the cluster center results do not exceed the confidence interval boundaries, then determining that the cluster center is merged with the airport's arrival and departure flight clustering results. Specifically, taking the historical data set as the overall sample, using the clustering results for classification, calculating the sample center, standard deviation, and error limit for the independent samples in each cluster, and obtaining the upper and lower limits of the confidence interval. The upper and lower limits of the confidence interval are compared with the cluster center in the same coordinate system. If the cluster center traffic size exceeds the upper and lower limits of the confidence interval within a specific time window, then the cluster and the cluster corresponding to the confidence interval are separated within the specific time window, otherwise the two are merged. Among them, the sample center is represented as |S q | is the number of observations in the independent sample q, is the jth observation in the independent sample q; the sample standard deviation is The sample error limit is where z_critical(0.995) is the Z critical value for 99% confidence.

[0027] The airport flight plan strategic delay prediction method based on delay decomposition provided by the embodiment of the present invention uses airport flight operation data to obtain the clustering results of airport arrival and departure flights; uses the airport arrival and departure flight clustering results to obtain the airport arrival and departure capacity probability map through clustering result evaluation index and confidence interval judgment method; according to the airport operation mode, by analyzing the airport operation queuing network model, using the airport arrival and departure capacity probability map, establish the arrival and departure native delay prediction model; using airport traffic management data, through airport flow control delay correlation analysis, establish the airport flow control delay prediction model; according to the arrival and departure native delay prediction model and the airport flow control delay prediction model, use the airport flight delay propagation judgment formula to establish the airport flight plan delay prediction model; according to the airport flight plan delay prediction model, use the airport flight plan to simulate and obtain the airport flight plan arrival and departure delay prediction value. Thus, timely and effective prediction of the airport flight plan arrival and departure delay is achieved. BRIEF DESCRIPTION OF THE DRAWINGS

[0028] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings required for use in the embodiments will be briefly introduced below. 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 paying creative work.

[0029] Figure 1 A schematic diagram of a method flow chart provided by an embodiment of the present invention;

[0030] Figure 2 A schematic diagram of a probability diagram of arrival and departure capacity provided by an embodiment of the present invention;

[0031] Figure 3 A schematic diagram of a departure native delay queuing network model provided by an embodiment of the present invention;

[0032] Figure 4 A schematic diagram of a network model for a native inbound delay queue provided by an embodiment of the present invention;

[0033] Figure 5 A schematic diagram of an airport flight delay propagation determination formula provided by an embodiment of the present invention;

[0034] Figure 6 A schematic diagram of an airport flight schedule delay prediction system provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0035] In order to enable those skilled in the art to better understand the technical solution of the present invention, the present invention is further described in detail below in conjunction with the accompanying drawings and specific embodiments. The embodiments of the present invention will be described in detail below, and examples of the embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements with the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and cannot be interpreted as limiting the present invention. It can be understood by those skilled in the art that, unless specifically stated, the singular forms "one", "one", "said" and "the" used herein may also include plural forms. It should be further understood that the term "including" used in the specification of the present invention refers to the presence of the features, integers, steps, operations, elements and / or components, but does not exclude the presence or addition of one or more other features, integers, steps, operations, elements, components and / or their groups. It should be understood that when we say that an element is "connected" or "coupled" to another element, it can be directly connected or coupled to other elements, or there may also be intermediate elements. In addition, the "connection" or "coupling" used here may include wireless connection or coupling. The term "and / or" used herein includes any unit and all combinations of one or more associated listed items. It will be understood by those skilled in the art that, unless otherwise defined, all terms (including technical terms and scientific terms) used herein have the same meaning as generally understood by those of ordinary skill in the art to which the present invention belongs. It should also be understood that terms such as those defined in general dictionaries should be understood to have meanings consistent with the meanings in the context of the prior art, and will not be interpreted with idealized or overly formal meanings unless defined as herein.

[0036] The embodiment of the present invention provides a method for predicting airport flight schedule strategic delays based on delay splitting. The main design idea is: the method for predicting airport flight schedule strategic delays based on delay splitting is used to obtain the airport capacity probability map by using the airport flight schedule and historical flight operation data, and obtain the airport flight arrival and departure delay prediction value according to the airport flight schedule, airport operation mode modeling and airport capacity probability map. This embodiment is suitable for evaluating the rationality and effectiveness of airport flight schedule / planning in the strategic stage, such as Figure 1 As shown, the main process of this method includes:

[0037] The embodiment of the present invention provides a method for predicting airport flight schedule strategic delays based on delay decomposition. Figure 1 As shown, including:

[0038] Using the airport flight operation data, the clustering results of the airport's inbound and outbound flights are obtained, wherein the flight operation data includes: flight calculated off-block time data and flight actual landing time data;

[0039] Using the clustering results of the airport's arrival and departure flights, a probability map of the airport's arrival and departure capacity is obtained;

[0040] Using the arrival and departure capacity probability map, a model for predicting the arrival and departure delays of the airport is established;

[0041] Using the airport traffic management data, an airport flow control delay prediction model is established, wherein the airport traffic management data includes: flight departure airport and destination airport, flight target off-block time data and flight calculated off-block time data;

[0042] Establishing an airport flight schedule delay prediction model based on the original arrival and departure delay prediction model and the airport flow control delay prediction model;

[0043] According to the airport flight plan delay prediction model, the airport flight plan to be processed is predicted and processed, and the airport flight plan arrival and departure delay prediction value is obtained.

[0044] Specifically, before building an airport flight schedule delay prediction model, it is necessary to first conduct an airport capacity assessment and statistically analyze the airport arrival and departure flow profile based on historical flight operation data; secondly, use the Kmeans clustering method to perform cluster analysis on the arrival and departure flow profile, and evaluate and screen the clustering results based on the silhouette coefficient, variance ratio criterion and pruning method; finally, use the confidence interval method to obtain the key time nodes, and combine the screened cluster center results to generate the airport arrival and departure capacity probability map.

[0045] The airport queuing network model is analyzed to obtain the airport arrival and departure queuing network model, and then the airport arrival and departure native delay simulation model is constructed. At the same time, the historical flight flow control data is statistically analyzed to analyze the correlation between the flow control delay size and the flow size. According to the correlation test results, a flow control delay prediction model is constructed. Finally, based on the airport flight delay propagation judgment formula, the arrival and departure native delay simulation model and the flow control delay prediction model are coupled to construct an airport flight schedule delay prediction model based on delay decomposition and simulation.

[0046] Airport capacity assessment, including: statistical flow profiles, clustering of flow data, evaluation of clustering results, confidence interval methods and capacity probability plots.

[0047] Statistical traffic profile, including: departure traffic profile, arrival traffic profile. The departure traffic profile is to take the actual departure time of the daily departure flights, count the number of departure flights in every 15-minute time window, and obtain the traffic profile data set X = {x1, x2…, x n}, where x i represents the number of departure flights in the i-th time window; the inbound flow profile is expressed as follows: taking the actual block time of the daily inbound flights, counting the number of inbound flights in every 15-minute time window, and obtaining the flow profile data set. The data set statistical method is the same as the departure flow profile.

[0048] After obtaining the inbound and outbound traffic profile, the Kmeans clustering method is used to cluster the inbound and outbound traffic. Kmeans clustering is a data partitioning algorithm based on Euclidean distance, which is expressed as: divide the data set X into k clusters, where k is a positive integer of moderate size. For a specific cluster, the clustering distance between the observations in the cluster should be smaller than the distance between the observations in the cluster and the observations in other clusters. For each cluster, the center point of all observations in the cluster is k=1,2…K, also called The centroid of cluster k, the centroid is expressed as:

[0049]

[0050] In the formula is the qth observation in cluster k, and n is the number of observations in cluster k. The Kmeans algorithm performs a Euclidean distance test on each observation x and each centroid, and assigns x to the cluster corresponding to the nearest centroid. If the sum of the distances between each observation and the cluster to which it belongs is the smallest, the algorithm iteration ends.

[0051] Take K=3-16 as the initial number of clusters, draw all cluster centers as a line graph, and count the proportion of each cluster in the total observations. The line graph is a legend with the time window as the x-axis and the flow size as the y-axis, containing K cluster center lines.

[0052] All clustering results are evaluated and screened using the silhouette coefficient, variance ratio criterion, and pruning method. Specifically:

[0053] Silhouette coefficient: A measure of the degree of separation between clusters by comparing the similarity of each observation to the other observations in its assigned cluster. The silhouette coefficient of an observation x is:

[0054]

[0055] Where a(x i ) is the average distance from observation x to all other data points in its cluster, calculated as:

[0056]

[0057] b(x i ) is the average distance between the observation value x and the observation values ​​in the adjacent clusters, and the calculation formula is:

[0058]

[0059] In formula (3-4), d(x j ,x i ) is the distance between observations i and j, and |Ci| is the number of observations in the cluster.

[0060] Based on the silhouette coefficient s(x i ), the average value of the silhouette coefficient is:

[0061]

[0062] SI close to 1 indicates that the clusters are compact and well separated, close to 0 indicates overlapping clusters, and close to -1 indicates too many or too few clusters.

[0063] Variance ratio criterion: It is the distance ratio between the inter-cluster error variance and the intra-cluster error variance. The calculation formula is:

[0064]

[0065] In the formula, n is the total number of observations, and the formula for calculating the inter-cluster error variance is:

[0066]

[0067] Where n i is the number of observations in cluster i, c i is the centroid of cluster i, c is the overall centroid of all observations; the intra-cluster error variance calculation formula is:

[0068]

[0069] In the formula, x is all the observations in cluster i, c i It is the centroid of cluster i. The higher the CHI value, the better the clustering effect.

[0070] The pruning method is to remove unimportant weights in the model and reduce the number of model parameters and calculations while ensuring that the model accuracy is not affected. Specifically, the proportion of each cluster in the clustering result is counted. If a cluster with a proportion less than the threshold appears, it means that the pruning involves clusters with smaller features and less influence, and the size of k will no longer be increased. The threshold is generally 5%-10%.

[0071] According to the above cluster evaluation results, the best clustering results are obtained, and the key time nodes are obtained using the confidence interval method. The confidence interval judgment method includes: within a given time window, the cluster center result is compared with the confidence interval boundary obtained using the clustering result of the airport's incoming and outgoing flights. If the cluster center result does not exceed the confidence interval boundary, it is determined that the cluster center is merged with the clustering result of the airport's incoming and outgoing flights. Specifically, the historical data set is used as the overall sample, and the clustering results are used for classification. The sample center, standard deviation and error limit are calculated for the independent samples in each cluster to obtain the upper and lower limits of the confidence interval. The upper and lower limits of the confidence interval are compared with the cluster center in the same coordinate system. If the cluster center flow size exceeds the upper and lower limits of the confidence interval within a specific time window, the cluster and the cluster corresponding to the confidence interval are separated within the specific time window, otherwise the two are merged.

[0072] The sample center is represented by

[0073]

[0074] In the formula, |S q | is the number of observations in the independent sample q, is the jth observation in the independent sample q; the sample standard deviation is:

[0075]

[0076] The sample error calculation formula is:

[0077]

[0078] Where z_critical(0.995) is the Z critical value with a confidence level of 99%.

[0079] When the best clustering result is obtained, the capacity probability graph method is used to merge similar scenes and determine the separation scenes. The algorithm is as follows:

[0080] Using the confidence interval method to calculate the cluster center, the confidence interval set is obtained as follows:

[0081]

[0082] In the formula, For the 99% confidence interval of scene i in time window t, take cluster center j:

[0083] {x j (t)|t=1,2…,T} For confidence interval i and cluster center j, in the time period [t x ,t y ], if it satisfies the formula:

[0084]

[0085] Then cluster i and cluster j are in the time period [t x , t y ] is a fusion scene, otherwise the key node is specified in the time period [t x , t y ], define this key node as t ij .

[0086] Let set g represent the set of all scenes, if it satisfies the formula:

[0087]

[0088] Then for scene i, its separation time from other scenes is formula 15:

[0089] t ij (g)= min(t ij ), j∈g, j≠i (15)

[0090] Let G be the set of all groups, and the probability graph is constructed as follows: t ij = 1, then for unique i∈g, scene i is an independent scene, otherwise The earliest branching time of scene i is t ij (g).

[0091] The capacity probability diagram is obtained by using the confidence interval method and the capacity probability diagram method, as shown in the attached figure. Figure 2As shown in the figure, the capacity probability graph is represented as a set of capacity probability curves that merge and separate with time. The usage method is as follows: after a specific time window is selected, a capacity scenario in the time window is randomly selected. If an independent scenario is selected, the selection probability is the proportion of the cluster corresponding to the scenario to the total observation value, and the capacity size is the traffic profile size of the corresponding cluster center in the current time window; if a fusion scenario is selected, the selection probability is the difference between 1 and the remaining independent scenarios, and the capacity size is the weighted value of the corresponding traffic size values ​​of each independent scenario in the fusion scenario in a specific time window, where the weight is expressed as the quotient of the selection probability of each independent scenario in the fusion scenario and the sum of the selection probabilities.

[0092] Construct the native delay prediction model for arrival and departure, including: native delay prediction model for departure and native delay prediction model for arrival.

[0093] The departure delay prediction model is a simulation model of the queuing network of parking stands and runway heads. The simulation process is as follows: Figure 3 As shown, the initial arrival sequence and the initial capacity sequence are obtained by using the flight plan data and the departure capacity probability diagram, wherein the initial arrival sequence includes: the flight calculates the off-block time.

[0094] Iterate the flights. When the aircraft is in the parking state, if the number of currently operating flights is less than the departure capacity, the current aircraft can withdraw the wheel block and start taxiing. When the aircraft reaches the runway head, the waiting time is the difference between the actual take-off time of the last aircraft that has not taken off and the time when the current aircraft arrives at the runway head. When the iteration completes all aircraft on the day or the iteration time exceeds 24:00, the iteration ends, and the departure delay is the average departure delay of all flights on the day.

[0095] The original arrival delay prediction model is a simulation model of the terminal area and runway queue network. The simulation process is as follows: Figure 4 As shown, the initial arrival sequence and initial capacity sequence are obtained by using the flight schedule data and the arrival capacity probability map, where the initial arrival sequence includes: the flight scheduled landing time.

[0096] Iterate the flights. When the aircraft is in the terminal area, use the steady-state queuing model to calculate the aircraft's air queue delay. When the aircraft lands, its taxiing delay time is the difference between the entry time of the last aircraft that has not yet entered the taxiing system and the actual landing time of the current aircraft. When the iteration is completed for all aircraft on the day or the iteration time exceeds 24:00, the iteration ends, and the arrival delay is the average arrival delay of all flights on the day.

[0097] The airport flow control delay correlation analysis is a correlation test between the size of flow control delays and the size of traffic in different city pairs in different time periods. If the correlation is higher than a threshold, the flow control delay prediction model is a positive correlation function of the traffic size. Otherwise, the flow control delay prediction model is obtained by statistically analyzing the average size of flow control delays and the possibility of flow control delays. The threshold is generally 0.5, and the possibility of flow control delays is expressed as the ratio of the number of flights subject to flow control to the total number of flights for a specific city pair in a specific time period. Specifically, the size of flow control delays is expressed as the sum of flow control delays for a specific city pair in a specific time period; the average size of flow control delays is expressed as the size of flow control delays and the number of flights affected by flow control.

[0098] Airport delay propagation judgment formula, as shown in the attached Figure 5 As shown, it includes: comparing the difference between the stopover time of a specific flight at the airport and the minimum stopover time with the arrival delay of the flight. If the former is greater, the arrival delay of the flight will not be propagated to the subsequent flights. Otherwise, the subsequent flights of the flight will generate delay propagation, and its size is equal to the difference between the difference and the arrival delay of the flight. Flight arrival delay, denoted as AD i =AIBT i –SIBT i , of which AIBT i is the actual block time of the flight, SIBT i is the flight's scheduled block time; the difference between the stopover time of the specific flight at the airport and the minimum stopover time is

[0099] RDCT i =SOBT i+1 –SIBT i – MTTT i (16)

[0100] In the formula, SOBT i+1 is the scheduled off-block time of the subsequent flight, MTTT i is the shortest stopover time of the preceding flight; the delay propagation size is DP i =AD i -RDCT i , if DP i <0, there is no delay propagation, otherwise the target off-block time (TOBT) of the subsequent flight i+1 =DP i +SOBT i+1 .

[0101] Based on the original arrival and departure delay prediction model and the airport flow control delay prediction model, the airport flight delay propagation judgment formula is used to establish the airport flight schedule delay prediction model, as shown in the attached figure. Figure 6As shown in the figure, the arrival flight plan is simulated using the arrival native delay prediction model, which includes non-originating flights, and the average delay of arrival flights is obtained by simulation; the delay propagation judgment formula is used to generate delay propagation for the departure flight plan, where the departure flights include originating flights and non-originating flights, there is no delay propagation for the originating flights, and the delay propagation for non-originating flights is obtained according to the flight registration number and the delay propagation judgment formula; the flow control delay prediction model is used to generate flow control delay for the departure flight plan; the departure flight plan is combined with the delay propagation and flow control delay to generate the time series for calculating the block withdrawal of the departure flights; finally, the departure native delay prediction model is used to simulate the time series for calculating the block withdrawal of the departure flights to obtain the average delay of the departure flights.

[0102] The airport flight schedule delay prediction method provided by the embodiment of the present invention evaluates the airport capacity, and based on the historical flight operation data, the airport arrival and departure flow profile is statistically analyzed; the arrival and departure flow profile is clustered by using the Kmeans clustering method; the clustering results are evaluated and screened according to the silhouette coefficient, the variance ratio criterion and the pruning method; the key time nodes are obtained by using the confidence interval method, and the airport arrival and departure capacity probability map is generated in combination with the selected cluster center results; the airport queuing network model is analyzed to obtain the airport arrival and departure queuing network model, and the airport arrival and departure native delay simulation model is constructed; the historical flight flow control data is statistically analyzed, the correlation between the flow control delay size and the flow size is analyzed, and the flow control delay prediction model is constructed according to the correlation test result; based on the airport flight delay propagation judgment formula, the arrival and departure native delay simulation model and the flow control delay prediction model are coupled to construct an airport flight schedule delay prediction model based on delay decomposition and simulation.

[0103] The airport flight plan strategic delay prediction method based on delay decomposition and the airport flight plan delay prediction method based on delay decomposition and simulation modeling provided by the embodiment of the present invention use the airport flight operation data to obtain the airport arrival and departure flight clustering results; use the airport arrival and departure flight clustering results to obtain the airport arrival and departure capacity probability map through the clustering result evaluation index and the confidence interval judgment method; according to the airport operation mode, by analyzing the airport operation queuing network model, using the airport arrival and departure capacity probability map, establish the arrival and departure native delay prediction model; use the airport flow management data, through the airport flow control delay correlation analysis, establish the airport flow control delay prediction model; according to the arrival and departure native delay prediction model and the airport flow control delay prediction model, use the airport flight delay propagation judgment formula to establish the airport flight plan delay prediction model; according to the airport flight plan delay prediction model, use the airport flight plan to simulate and obtain the airport flight plan arrival and departure delay prediction value. Thus, the delay prediction of the airport flight plan is realized.

[0104] Each embodiment in this specification is described in a progressive manner, and the same or similar parts between the embodiments can be referred to each other, and each embodiment focuses on the differences from other embodiments. In particular, for the device embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiment. The above is only a specific implementation method of the present invention, but the protection scope of the present invention is not limited thereto. Any changes or replacements that can be easily thought of 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. Therefore, the protection scope of the present invention should be based on the protection scope of the claims.

Claims

1. A delay prediction method for airport flight planning strategy based on delay decomposition, characterized in that: include: S1. Obtaining clustering results of inbound and outbound flights from the airport using the airport flight operation data, wherein the flight operation data includes: flight calculated off-block time data and flight actual landing time data; S2. Using the clustering results of the airport's arrival and departure flights, obtain an airport arrival and departure capacity probability map; S3. Using the arrival and departure capacity probability map, establish an airport arrival and departure native delay prediction model; S4. Establishing an airport flow control delay prediction model using the airport traffic management data, wherein the airport traffic management data includes: flight departure airport and destination airport, flight target off-block time data and flight calculated off-block time data; S5. Establishing an airport flight schedule delay prediction model based on the original arrival and departure delay prediction model and the airport flow control delay prediction model; S6. According to the airport flight plan delay prediction model, the airport flight plan to be processed is predicted and processed, and the arrival and departure delay prediction result of the airport flight plan to be processed is obtained.

2. The method according to claim 1, characterized in that In S3, it also includes: Before establishing the airport's native arrival and departure delay prediction model using the arrival and departure capacity probability map, a queuing network model corresponding to the airport operation mode is loaded, wherein the airport operation mode includes: runway service mode, airport departure operation mode and airport arrival operation mode.

3. The method according to claim 2, characterized in that The native arrival and departure delay prediction model includes: a departure native delay prediction model and an arrival native delay prediction model; In the departure delay prediction model, the airport departure capacity is used to predict the flight stand waiting time and the queuing network model is used to predict the flight runway head waiting time; In the arrival delay prediction model, the airport arrival capacity is used to predict the flight arrival taxiing delay and the queuing network model is used to predict the flight terminal area delay time.

4. The method according to claim 1, characterized in that: In S4, it also includes: Before establishing the airport flow control delay prediction model, the airport flow control delay correlation analysis was conducted using the airport flow management data; According to the results of the airport flow control delay correlation analysis, the flow control delay prediction model is established.

5. The method according to claim 4, characterized in that The airport flow control delay correlation analysis includes: Obtaining quantitative values ​​of the correlation between the flow control delay size and the flow size of different city pairs in different time periods, wherein if the quantitative value of the correlation is higher than a threshold, the flow control delay prediction model is set as a positive correlation function with respect to the flow size; If the quantitative value of the correlation is less than or equal to the threshold, the flow control delay prediction model is obtained by statistically analyzing the size of the flow control delay and the possibility of the flow control delay, wherein the possibility of the flow control delay includes: the ratio of the number of flights subject to flow control to the total number of flights for a specific city pair within a specific time period.

6. The method according to claim 1, characterized in that In S5, it includes: generating an arrival delay prediction value of the airport flight plan using the arrival native delay prediction model, and then converting the arrival delay prediction value of the airport flight plan into a departure propagation delay prediction value; The flight flow control delay is generated by the flow control delay prediction model, and then the departure delay prediction value of the flight plan is generated by using the departure native delay prediction model.

7. The method according to claim 6, characterized in that In S5, also included: The difference between the stopover time of a specific flight at the airport and the minimum stopover time is compared with the arrival delay of the specific flight. If the former is greater, it is determined that the arrival delay of the specific flight will not be propagated to subsequent flights. Otherwise, the subsequent flights of the specific flight will be delayed.

8. The method according to claim 1, characterized in that In S6, the flight plan of the airport to be processed includes: the date of the flight of the airport to be processed, the departure airport and the destination airport of the flight of the airport to be processed, the planned off-block time of the flight of the airport to be processed and the planned on-block time of the flight of the airport to be processed. The predicted results of the planned arrival and departure delays of the airport flights to be processed include: the average delay of arrival flights, the average flow control delay of departure flights, the average delay of departure flights, and the average total delay of departure flights.

9. The method according to claim 8, characterized in that The average inbound flight delay is the ratio of the total inbound delays to the total number of inbound flights; The average flow control delay of the departing flights is the ratio of the total flow control delay to the total number of departing flights; The average departure flight delay is the ratio of the total departure flight delays to the total number of departure flights; The average total delay of the departing flights is the sum of the average flow control delay of the departing flights and the average delay of the departing flights.

10. The method according to claim 8, characterized in that In S1, the airport arrival and departure flight clustering results include: flight departure flow clustering results and flight arrival flow clustering results; The flight departure flow clustering result is: the clustering center result of the actual take-off flow profile of the airport flights using the Kmeans clustering method; The flight arrival flow clustering result is: the clustering center result of the actual block flow profile of airport flights using the Kmeans clustering method.

Images