A clustering-based flight flow management strategy recommendation method and electronic equipment

Abstract

The present invention relates to the technical field of civil aviation air traffic management, and in particular to a flight flow management strategy recommendation method and electronic equipment based on clustering. The present invention constructs a complete air traffic static network model and an air traffic dynamic flow model; and then performs cluster analysis on the flow management strategy through machine learning. Finally, the similarity between the flow situation of the flight flow data to be analyzed and the historical flow situation is analyzed, and a flow control strategy with high similarity is recommended to provide decision support for the traffic management of the controller, thereby realizing intelligent traffic management strategy recommendation and formulation. The time for formulating strategies is greatly reduced, and the workload of the controller is also greatly reduced; at the same time, the accuracy of the control strategy is also enhanced, ensuring that the control time and control range are implemented in a reasonable range, avoiding unnecessary delays caused by inaccurate control strategies and the situation where flights wait in the air, and effectively improving the control efficiency of the controller.

Description

Technical Field

[0001] The invention relates to the technical field of civil aviation air traffic management, and in particular to a clustering-based flight flow management strategy recommendation method and electronic equipment. Background Art

[0002] The aviation industry plays an important role in promoting the world's economic and social development. As a complex system that ensures air traffic flight safety and maintains air traffic order, the air traffic control system faces huge challenges. Traffic management uses big data to centrally analyze all operating data of air traffic control departments, airports, and airlines to provide support for air traffic control decisions. Studying traffic management measures and establishing a traffic management system are effective measures taken by various countries for traffic management.

[0003] Over the past few decades, the United States, Europe and other countries have successively built traffic management systems. The United States has built an air traffic control system command center (ATCSCC) to monitor air traffic control operations and coordinate solutions to daily air traffic flow problems. Europe has built a central flow management unit (CFMU) to enhance the coordination and solution capabilities of traffic problems across Europe. China is also establishing a national traffic management system and a regional flight flow system. Compared with the traffic management methods and systems in Europe and the United States, China's flight flow management methods and procedures are different from those in Europe and the United States, mainly reflected in: 1. China's flight flow management is mainly proposed by regional traffic management and approved by national traffic management; while the European and American traffic management centers are unified in formulating and issuing traffic management measures; 2. Due to different management methods and information, China's traffic management measures and parameters are different from those in Europe and the United States.

[0004] Therefore, the current formulation of traffic management measures at home and abroad mainly relies on the experience of traffic management personnel, and there is a lack of intelligent means. With the development of artificial intelligence and big data technology, intelligent methods are used to improve the efficiency and rationality of traffic management, and avoid unnecessary delays caused by manually formulated traffic management measures that take too long or cover too wide a range, or control time is insufficient, causing flights to wait in the air, and the safety hazards of excessive workload of controllers.

[0005] Therefore, in order to improve the intelligence and rationality of flight flow management strategy formulation, the present invention proposes a flight flow management strategy recommendation method based on clustering, which considers the similarity of flight flow management scenarios, clusters historical flight flow trends and scenarios, and recommends flight flow management strategies based on recommendation algorithms. The clustering-based recommendation algorithm can meet both control rules and the control habits of controllers, providing decision support for flight flow management strategy formulation. Summary of the invention

[0006] The purpose of the present invention is to overcome the safety hazards existing in the prior art, such as unnecessary delays caused by too long control time or too wide a scope due to manually formulating traffic management measures, or flights waiting in the air due to insufficient control time and excessive workload of controllers, and to provide a flight traffic management strategy recommendation method and electronic equipment based on clustering.

[0007] In order to achieve the above-mentioned object of the invention, the present invention provides the following technical solutions:

[0008] A clustering-based flight flow management strategy recommendation method comprises the following steps:

[0009] S1: Obtain flight plans, surveillance data, flow control data and control rules to build a static network model of air traffic;

[0010] S2: establishing an air traffic dynamic flow model for airports and key waypoints based on the air traffic static network model, and generating flight flow data and flow and capacity change trends of each piece of the flight flow data;

[0011] S3: According to the flow and capacity change trends of each piece of the flight flow data, the flight flow data with capacity and flow imbalance is obtained, and the controller formulates a flow management strategy to make the flight flow data with capacity and flow imbalance reach a state of capacity and flow balance;

[0012] S4: preprocessing and feature extraction of the traffic management strategy, and then generating a traffic management strategy feature set according to the extracted features;

[0013] The characteristics of the traffic management policy include a sequence number and / or an effective date and / or a publishing unit and / or a controlled unit and / or a start time and / or an end time and / or a duration and / or a flow control point and / or a flow control value and / or a flow control reason;

[0014] S5: performing cluster analysis on the traffic management strategy feature set according to the features of the traffic management strategy feature set, and outputting the traffic situation of the traffic management strategy corresponding to each traffic management strategy feature set;

[0015] The flow situation includes: short duration, small flow control value; long duration, small flow control value; long duration, large flow control value; short duration, large flow control value;

[0016] S6: Using causal recommendation technology, according to the similarity between the flow situation of the flight flow data to be analyzed and the historical flow situation, the recommended flow management strategy is output and the similarity is sorted. The present invention constructs a complete static network model of air traffic and a dynamic flow model of air traffic based on flight plans, monitoring data, flow control data and control rules, so as to comprehensively reflect the flow change situation; then cluster analysis of the flow management strategy is performed through machine learning to establish a typical scenario library. Finally, the similarity between the flow situation of the flight flow data to be analyzed and the historical flow situation is analyzed, and a flow control strategy with high similarity is recommended to provide decision support for the traffic management of the controller, thereby realizing intelligent traffic management strategy recommendation and formulation. It greatly reduces the time for formulating strategies and greatly reduces the workload of the controller; at the same time, it also enhances the accuracy of the control strategy, ensures that the control time and control range are implemented in a reasonable range, avoids unnecessary delays caused by inaccurate control strategies and the situation of flights waiting in the air, and effectively improves the control efficiency of the controller.

[0017] As a preferred solution of the present invention, the cluster analysis in S5 includes K-means clustering and / or hierarchical clustering and / or density-based clustering.

[0018] As a preferred embodiment of the present invention, S1 comprises:

[0019] S1-1: Obtaining a flight plan, monitoring data, flow control data, and control rules; the flight plan includes the departure airport, route, and landing airport of each aircraft;

[0020] S1-2: Establish an airport network layer G1 (V according to the departure airport, route and landing airport d ,V a ,E1), where V d is the departure airport collection, V a It is the set of landing airports, the route string composed of waypoints between E1 take-off and landing airports, which comes from the route field of the flight plan;

[0021] S1-3: Establish the association relationship between the airport approach point and the airport according to the monitoring data, and establish the airport approach network layer G2 (V a ,V e ,E2); where, V a It is to meet at the airport, V e is a set of approach points, where the approach point is the point where area control is handed over to approach control in the control command, and E2 is the line between the airport and the approach point;

[0022] S1-4: Determine the flow control points of adjacent control intervals and the network layer between the flow control points and the airport approach point according to the flow control data and the control rules, G3(V e,V c ,E3); where V e is the set of entry points, V c is a set of flow control points, wherein the flow control point (air traffic flow management control point) is a point where the operation flow is transferred and restricted between the control of two adjacent areas, and E3 is a line between the flow control point and the approach point;

[0023] S1-5: Determine the network layer between the flow control point and the departure airport and the route according to the flow control data, G4(V c ,V d ,E4); where V c is the set of air traffic flow management control points, V d is the set of departure airports, and E4 is the line between the flow control point and the departure airport;

[0024] S1-6: Construct an air traffic static network model G(V according to the network layer established in the above steps a ,V e ,V c ,V d ,E1,E2,E3,E4).

[0025] As a preferred embodiment of the present invention, S2 comprises:

[0026] S2-1: Establishing an air traffic dynamic flow model of airports and key waypoints based on the air traffic static network model, and generating flight flow data; the flight flow data includes flow data of airports, approach points and flow control points, and capacity values ​​of airports and flow control points;

[0027] S2-2: Generate the flow and capacity change trend of each flow control point i in each time slice t according to the flight flow data; the flow and capacity change trend includes: flow Traffic change trend The proportion of traffic at flow control point i Capacity value Capacity change trend Capacity-current ratio

[0028] As a preferred embodiment of the present invention, the flow rate variation trend Includes: rise, hold, fall;

[0029] When t=1, the flow change trend is set to maintain;

[0030] When t>1, the flow change is calculated as:

[0031]

[0032] in, is the flow rate at time slice t-1;

[0033] The capacity change trend Includes: rise, hold, fall;

[0034] When t=1, the capacity change trend is set to maintain;

[0035] When t>1, the capacity change trend is calculated as:

[0036]

[0037] in, is the capacity at time slice t-1.

[0038] As a preferred solution of the present invention, S3 includes:

[0039] S31: According to the flow and capacity change trend of each piece of the flight flow data, the capacity-flow ratio The matching degree between flow and capacity;

[0040] S32: When the matching degree of the flight flow data is less than a preset threshold, the flight flow data is capacity-flow imbalance, and the controller generates a flow management strategy for the flight flow data according to the flow and capacity change trend; otherwise, the flight flow data is capacity-flow balance;

[0041] S33: Output the flight flow data of capacity-flow imbalance and the corresponding flow management strategy.

[0042] As a preferred embodiment of the present invention, S4 comprises:

[0043] S4-1: Preprocessing the traffic management policy; the preprocessing includes deleting duplicate, erroneous, and traffic management policies with incomplete key information;

[0044] S4-2: Establish a traffic management policy feature set, and classify the pre-processed traffic management policies into corresponding traffic management policy feature sets according to the features of the traffic management policies.

[0045] As a preferred embodiment of the present invention, S6 comprises:

[0046] S6-1: inputting flight flow data to be analyzed, and using causal recommendation technology to obtain the flow situation of the flight flow data to be analyzed;

[0047] S6-2: Based on the similarity between the traffic situation and the historical traffic situation, the recommended traffic management strategy is output and sorted by similarity.

[0048] As a preferred embodiment of the present invention, the S6 further includes S6-3:

[0049] S6-3: Determine whether the number of currently affected flights is greater than 0; when the number of affected flights is 0, stop calculating;

[0050] The calculation formula for the number of affected flights is:

[0051]

[0052] in, is the number of affected flights at time slice t, is the flow rate at time slice t, is the flow rate at time slice t-1, is the capacity of flow control point i at time slice t, and T is the duration of the flight flow management strategy.

[0053] An electronic device comprises at least one processor and a memory communicatively connected to the at least one processor; the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute any of the methods described above.

[0054] Compared with the prior art, the present invention has the following beneficial effects:

[0055] The present invention constructs a complete static network model of air traffic and a dynamic flow model of air traffic based on flight plans, monitoring data, flow control data and control rules, so as to comprehensively reflect the flow change trend; then cluster analysis is performed on the flow management strategy through machine learning to establish a typical scenario library. Finally, the similarity between the flow situation of the flight flow data to be analyzed and the historical flow situation is analyzed, and a flow control strategy with high similarity is recommended to provide decision support for the traffic management of the controller, thereby realizing intelligent traffic management strategy recommendation and formulation. It greatly reduces the time for formulating strategies and greatly reduces the workload of the controller; at the same time, it also enhances the accuracy of the control strategy, ensures that the control time and control range are implemented in a reasonable range, avoids unnecessary delays caused by inaccurate control strategies and the situation where flights wait in the air, and effectively improves the control efficiency of the controller. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] Figure 1 It is a schematic diagram of the flight flow management process in a specific implementation manner of the present invention.

[0057] Figure 2 This is a flow chart of a clustering-based flight flow management strategy recommendation method described in Example 1 of the present invention.

[0058] Figure 3This is a structural diagram of an air traffic static network model in a clustering-based flight flow management strategy recommendation method described in Example 2 of the present invention.

[0059] Figure 4 This is a traffic distribution diagram of airport approach traffic hours in a clustering-based flight traffic management strategy recommendation method described in Example 2 of the present invention.

[0060] Figure 5 This is a statistical chart of the start time of flow control measures in a clustering-based flight flow management strategy recommendation method described in Example 2 of the present invention.

[0061] Figure 6 This is a flow control point feature representation diagram in a clustering-based flight flow management strategy recommendation method described in Example 2 of the present invention.

[0062] Figure 7 A schematic diagram of the evaluation scores of different clustering models in a clustering-based flight flow management strategy recommendation method described in Example 2 of the present invention.

[0063] Figure 8 This is a distribution diagram of flow control value and flow control duration of a traffic management strategy in a clustering-based flight traffic management strategy recommendation method described in Example 2 of the present invention.

[0064] Fig. 9 An electronic device described in Example 5 of the present invention utilizes the clustering-based flight flow management strategy recommendation method described in Example 2. DETAILED DESCRIPTION

[0065] The present invention is further described in detail below in conjunction with test examples and specific implementation methods. However, this should not be understood as the scope of the above subject matter of the present invention being limited to the following embodiments, and all technologies realized based on the content of the present invention belong to the scope of the present invention.

[0066] like Figure 1 As shown in the figure, the implementation process of the air traffic flow management strategy includes: predicting the traffic flow and capacity of airports and key flights, monitoring the capacity and flow matching on this basis, and analyzing the capacity and flow imbalance situation such as the degree of capacity and flow imbalance, time distribution, and flow distribution when monitoring the capacity and flow imbalance in the future. Formulate, publish and implement flight flow management strategies. During the implementation of flight flow management, continuously monitor the capacity, flow and capacity and flow balance status, and analyze and take necessary measures in a timely manner when capacity and flow imbalance is found. After the flow management measures are implemented, conduct a post-analysis of the flow management measures.

[0067] However, the formulation of traffic management measures at home and abroad currently mainly relies on the experience of traffic management personnel, and there is a lack of intelligent means. With the development of artificial intelligence and big data technology, intelligent methods are used to improve the efficiency and rationality of traffic management, to avoid unnecessary delays caused by too long control time or too wide range of manually formulated traffic management measures, or insufficient control time causing flights to wait in the air and excessive workload of controllers, which poses a safety hazard. Therefore, in order to improve the intelligence and rationality of flight traffic management strategy formulation, the present invention provides a flight traffic management strategy recommendation method and electronic equipment based on clustering.

[0068] Example 1

[0069] like Figure 2 As shown, a clustering-based flight flow management strategy recommendation method includes the following steps:

[0070] S1: Obtain flight plans, surveillance data, flow control data, and control rules to build a static network model of air traffic.

[0071] S2: Establishing a dynamic air traffic flow model for airports and key waypoints based on the air traffic static network model, and generating flight traffic data and the flow and capacity change trends of each piece of the flight traffic data.

[0072] S3: According to the flow and capacity change trends of each piece of the flight flow data, the flight flow data with capacity and flow imbalance is obtained, and the controller formulates a flow management strategy to make the flight flow data with capacity and flow imbalance reach a state of capacity and flow balance;

[0073] S4: Preprocessing and feature extraction are performed on the traffic management strategy, and then a traffic management strategy feature set is generated according to the extracted features.

[0074] The characteristics of the traffic management strategy include serial number and / or effective date and / or issuing unit and / or controlled unit and / or start time and / or end time and / or duration and / or flow control point and / or flow control value and / or flow control reason.

[0075] S5: Perform cluster analysis on the traffic management policy feature set according to the features of the traffic management policy feature set, and output the traffic situation of the traffic management policy corresponding to each traffic management policy feature set; the cluster analysis includes K-means clustering and / or hierarchical clustering and / or density-based clustering.

[0076] The flow situation includes: short duration, small flow control value; long duration, small flow control value; long duration, large flow control value; short duration, large flow control value;

[0077] S6: Using causal recommendation technology, according to the similarity between the traffic situation of the flight traffic data to be analyzed and the historical traffic situation, the recommended traffic management strategy is output and sorted by similarity.

[0078] Example 2

[0079] This embodiment is an implementation method of the method described in Example 1, and specifically includes the following steps:

[0080] S1: Obtain flight plans, surveillance data, flow control data and control rules to build a static network model of air traffic;

[0081] S1-1: Obtain flight plan, monitoring data, flow control data and control rules; the flight plan of the flight includes: flight number, departure airport, landing airport, expected departure time, route and specific route information. The flight taking off or landing from the airport of interest can be determined based on the departure airport, landing airport and route fields.

[0082] S1-2: Establish an airport network layer G1 (V according to the departure airport, route and landing airport d ,V a ,E1),V d is the set of departure airports; where V a It is the set of landing airports, the route string composed of the waypoints between the take-off and landing airports of E1. For example, the route string of the flight taking off from Beijing Airport ZBAA and landing at Chengdu Airport ZUUU is {VYK, GODONATERI, WJC, SUBUL, ...};

[0083] S1-3: The landing airport has multiple approach points. Since there is no association between the landing airport and the approach point in the flight plan, it is necessary to establish an association between the airport approach point and the airport based on the monitoring data, and establish the airport approach network layer G2 (V a ,V e ,E2); where V a It is to meet at the airport, V e It is a set of approach points, where the area control is transferred to the approach control in the control command, and E2 is the line between the airport and the approach point; the route in the flight plan is mainly the route passed during the level flight, and lacks the route information of the aircraft from the airport to the cruising process. By analyzing the historical monitoring data, the relationship between the airport approach point and the airport can be determined.

[0084] S1-4: Determine the flow control points of adjacent control intervals and the network layer between the flow control points and the airport approach point according to the flow control data and the control rules, G3(V e ,V c ,E3); where V eis the set of entry points, V c It is a set of flow control points. The flow control point (i.e., air traffic flow management control point) is the point where the operating flow is transferred and restricted between two adjacent regional controls. The flow control point is an air traffic flow management control point. E3 is the line between the flow control point and the approach point. Flight flow control controls the traffic volume at the handover point of adjacent control areas, analyzes historical flow control information and unstructured control handover protocols (i.e., control rules), and can determine the flow control points of adjacent control intervals and the connection relationship between the flow control points and the airport approach points.

[0085] S1-5: Determine the network layer between the flow control point and the departure airport and the route according to the flow control data, G4(V c ,V d ,E4); where V c is the set of air traffic flow management control points, V d is the set of departure airports, and E4 is the line between the flow control point and the departure airport;

[0086] S1-6: Construct an air traffic static network model G (Va, Ve, Vc, Vd, E1, E2, E3, E4) according to the network layer established in the above steps. The structure of the air traffic static network model is as follows: Figure 3 As shown, the black dots represent airports, the gray dots connected to the airports are approach points, and the triangles represent flow control points.

[0087] The relationship between the above points is represented in the form of a matrix, and the flow control points and the entry points are numbered. The corresponding relationship matrix between the entry points and the flow control points is shown as follows:

[0088]

[0089] Among them, ec mn is the approach point m and the flow control point n. When there is a route between the flow control point and the approach point, it is 1; when there is no route between the flow control point and the approach point, it is 0.

[0090] S2: Establishing a dynamic air traffic flow model for airports and key waypoints based on the air traffic static network model, and generating flight traffic data and the flow and capacity change trends of each piece of the flight traffic data.

[0091] S2-1: Based on the static air traffic network model, a dynamic air traffic flow model for airports and key waypoints is established, and flight traffic data is generated; the flight traffic data includes the traffic data of airports, approach points and flow control points, as well as the capacity values ​​of airports and flow control points. S2-2: Based on the flight traffic data, the flow and capacity change trends of each flow control point i at each time slice t are generated; the flow and capacity change trends include: flow Traffic change trend The proportion of traffic at flow control point i Capacity value Capacity change trend Capacity-current ratio The proportion of the flow at the flow control point is the ratio of the flow value at the flow control point to the total flow value of the airport.

[0092] The flight flow values ​​of each time slice of the flow control point are shown in the following table:

[0093] Flow Control Point Time Slice 0800 Time Slice 0900 Time slice 1000 Time Slice 1100 HFE 88 96 75 82 ZHO 62 84 80 74 OBLIK 56 53 66 61 PAVIU 66 79 73 61

[0094] The flow rate change trend Includes: rise, hold, fall;

[0095] When t=1, the flow change trend is set to maintain; when t>1, the flow change calculation is:

[0096]

[0097] in, is the flow rate at time slice t-1;

[0098] The capacity change trend Includes: rise, hold, fall;

[0099] When t=1, the capacity change trend is set to maintain; when t>1, the capacity change trend is calculated as:

[0100]

[0101] in, is the capacity at time slice t-1.

[0102] S3: According to the flow and capacity change trends of each piece of the flight flow data, the flight flow data with capacity and flow imbalance is obtained, and the controller formulates a flow management strategy to make the flight flow data with capacity and flow imbalance reach a state of capacity and flow balance.

[0103] S31: According to the flow and capacity change trend of each piece of the flight flow data, the capacity-flow ratio The matching degree between flow and capacity;

[0104] S32: When the matching degree of the flight flow data is less than a preset threshold, the flight flow data is capacity-flow imbalance, and the controller generates a flow management strategy for the flight flow data according to the flow and capacity change trend; otherwise, the flight flow data is capacity-flow balance;

[0105] S33: Output the flight flow data of capacity-flow imbalance and the corresponding flow management strategy.

[0106] S4: Preprocessing and feature extraction are performed on the traffic management policy, and a traffic management policy feature set is generated according to the extracted features.

[0107] S4-1: Preprocess the traffic management policy; the preprocessing includes deleting duplicate (deleting redundant data when two or more traffic management policy data are exactly the same), erroneous and incomplete key information (key information of the flow control policy such as flow control points, start time, end time, etc., and traffic management policies that are obviously wrong or missing).

[0108] Preferably, the preprocessing further includes decomposing a traffic management policy containing two or more flow control points in a traffic management policy into two or more traffic management policies, so that each traffic management policy contains only one flow control point.

[0109] S4-2: Establish a traffic management policy feature set, and classify the pre-processed traffic management policies into corresponding traffic management policy feature sets according to the features of the traffic management policies.

[0110] The characteristics of the traffic management strategy include serial number and / or effective date and / or issuing unit and / or controlled unit and / or start time and / or end time and / or duration and / or flow control point and / or flow control value and / or flow control reason.

[0111] The main reasons for flow control include: flight schedule, weather, activities, etc. Figure 4 As shown in Figure 1, the airport arrival traffic exceeds the capacity in several time periods every day. In order to ensure flight safety, the number of flow control measures issued during these time periods is the largest, such as Figure 5 As shown in the figure, the flow control caused by flight schedules is mainly concentrated in the evening and night. Therefore, the flight schedules are divided into evening, night and other time periods according to the time period.

[0112] Flow control caused by weather reasons is classified into thunderstorm, low visibility, etc. according to weather characteristics. One-hot is used to encode the flow control reasons. The flow control reason encoding is shown in the following table:

[0113] Flow control reasons coding Evening flight schedule 100000 Night flight schedule 010000 Other flight schedules 001000 Thunderstorm 000100 Low visibility weather 000010 Activity 000001

[0114] The flow control duration and flow control value are segmented and coded, such as duration of 1 hour, 2 hours, etc. The flow control value is coded according to the number of flights passing through. The flow control value and coding are shown in the following table:

[0115] Flow Control Value Flow control value encoding (flow / hour) 2 in 10 minutes 12 2 flights per 15 minutes 8 2 flights in 20 minutes or 1 flight in 10 minutes 6 …… …… 20 minutes / flight 3 30 minutes / flight 2 60 minutes / flight 1

[0116] Based on the above dynamic traffic model and traffic management measures feature representation, the flow control point features are represented, such as Figure 6 The characteristics of each flow control point are represented from three dimensions: time, flow control point, and feature. The characteristics include: flow rate, flow change trend, flow ratio, capacity, overcapacity, capacity change trend, flow control information, etc.

[0117] S5: Perform cluster analysis on the traffic management policy feature set according to the features of the traffic management policy feature set, and output the traffic situation of the traffic management policy corresponding to each traffic management policy feature set; the cluster analysis includes K-means clustering and / or hierarchical clustering and / or density-based clustering.

[0118] The controller receives information about weather and activities, and evaluates the impact of weather and activities on capacity. The controller formulates flow control strategies based on the degree of flow exceeding capacity and flow change trends. When the airport flow exceeds the airport capacity for a certain period of time, the controller formulates a flow strategy due to flight schedules. Seven clustering models, including K-means clustering, DBSCAN clustering, Agglomerative Clustering hierarchical clustering, BIRCH clustering, spectral clustering, and Gaussian mixture model, are used to perform cluster analysis on flow control measures. The silhouette coefficient, Calinski Harabasz score, and Davies Bouldin index are used to evaluate the effect of the flow control clustering model. The larger the Silhouette coefficient and Calinski Harabasz score, the better the clustering effect, and the smaller the Davies Bouldin index value, the better the clustering effect. From the evaluation scores, it can be seen that the K-means clustering effect is better than other clustering models, such as Figure 7 shown.

[0119] For each type of flow control reason, the traffic management strategy is divided into the following four categories based on density-based cluster analysis according to flow control duration, flow control value, etc.: short duration, small flow control value; long duration, small flow control value; long duration, large flow control value; short duration, large flow control value. Statistics on the flow control value and flow control duration distribution of the traffic management strategy are as follows: Figure 8 shown.

[0120] S6: Using causal recommendation technology, according to the similarity between the traffic situation of the flight traffic data to be analyzed and the historical traffic situation, the recommended traffic management strategy is output and sorted by similarity.

[0121] S6-1: inputting flight flow data to be analyzed, and using causal recommendation technology to obtain the flow situation of the flight flow data to be analyzed;

[0122] S6-2: Based on the similarity between the traffic situation and the historical traffic situation, the recommended traffic management strategy is output and sorted by similarity.

[0123] Changes in traffic trends, reduced capacity, and traffic values ​​exceeding capacity values ​​are the reasons for traffic measures. Reasonable traffic strategies are adopted to limit flight traffic and ensure flight safety. Causal recommendation technology is used to recommend traffic management strategies based on the similarity between flow control causes, flow control start time, end time, flow control duration, flow control trends, etc. and historical traffic trends and flow control measures, and sort them according to the similarity.

[0124] Example 3

[0125] The difference between this embodiment and embodiment 2 is that: S6 further includes S6-3:

[0126] S6-3: Determine whether the current number of affected flights is greater than 0; when the number of affected flights is 0, stop calculating.

[0127] When the controller receives information about weather and activities, he / she determines whether the airport capacity has decreased based on the scope, duration, and degree of impact of the weather and activities, and determines the new dynamic capacity; calculates the flow rate and capacity-flow ratio within the duration of the weather and activities, and does not issue flow control measures when the flow rate is less than or equal to the capacity. When the flow rate exceeds the capacity due to flight schedules, the flow control value and duration are determined based on the capacity-flow ratio, flow development trend, etc. When the flow rate is greater than the capacity, calculate the number of affected flights in each time slice, the number of affected flights = the number of flows - the number of capacities + the number of backlogs (the number of affected flights in the previous time slice), and stops calculating when the number of affected flights is not greater than 0.

[0128] The calculation formula for the number of affected flights is:

[0129]

[0130] in, is the number of affected flights at time slice t, is the flow rate at time slice t, is the flow rate at time slice t-1, is the capacity of flow control point i at time slice t, and T is the duration of the flight flow management strategy.

[0131] Example 4

[0132] This embodiment is a practical application example of using the method described in Example 2 to process actual flight traffic data.

[0133] The historical flight traffic data are the traffic values ​​of each flow control point from 22:00 on July 25, 2019 to 2:00 on July 26, 2019; the flight traffic data to be analyzed are the traffic values ​​of each flow control point from 22:00 on August 6, 2019 to 2:00 on August 7, 2019. The details are shown in the following table.

[0134]

[0135] Since the traffic information predicted by each traffic control point from 22:00 on July 25 to 2:00 on July 26 and from 22:00 on August 6 to 2:00 on August 7 is similar, as shown in the following table, the traffic control strategy for August 6 can be recommended to be the traffic control strategy for July 25.

[0136]

[0137] Example 5

[0138] like Fig. 9 As shown, an electronic device includes at least one processor and a memory connected to the at least one processor in communication; the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the clustering-based flight flow management strategy recommendation method described in the above embodiment. The input and output interface may include a display, a keyboard, a mouse, and a USB interface for inputting and outputting data; the power supply is used to provide power to the electronic device.

[0139] Those skilled in the art can understand that: all or part of the steps of implementing the above-mentioned method embodiment can be completed by hardware related to program instructions, and the aforementioned program can be stored in a computer-readable storage medium. When the program is executed, it executes the steps of the above-mentioned method embodiment; and the aforementioned storage medium includes: mobile storage devices, read-only memories (ROM), disks or optical disks, etc. Various media that can store program codes.

[0140] When the above-mentioned integrated unit of the present invention is implemented in the form of a software functional unit and sold or used as an independent product, it can also be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the embodiment of the present invention can be essentially or partly reflected in the form of a software product that contributes to the prior art. The computer software product is stored in a storage medium, including several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the methods described in each embodiment of the present invention. The aforementioned storage medium includes: various media that can store program codes, such as mobile storage devices, ROMs, magnetic disks, or optical disks.

[0141] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A clustering-based flight flow management strategy recommendation method, characterized in that: The following steps are involved: S1: Obtain flight plans, surveillance data, flow control data and control rules to build a static network model of air traffic; S2: establishing a dynamic air traffic flow model for airports and key waypoints based on the air traffic static network model, and generating flight traffic data and the flow and capacity change trend of each flight traffic data; S3: According to the flow and capacity change trends of each piece of the flight flow data, the flight flow data with capacity and flow imbalance is obtained, and the controller formulates a flow management strategy to make the flight flow data with capacity and flow imbalance reach a state of capacity and flow balance; S4: preprocessing and feature extraction of the traffic management strategy, and then generating a traffic management strategy feature set according to the extracted features; The characteristics of the traffic management policy include a sequence number and / or an effective date and / or a publishing unit and / or a controlled unit and / or a start time and / or an end time and / or a duration and / or a flow control point and / or a flow control value and / or a flow control reason; S5: performing cluster analysis on the traffic management strategy feature set according to the features of the traffic management strategy feature set, and outputting the traffic situation of the traffic management strategy corresponding to each traffic management strategy feature set; The flow situation includes: short duration, small flow control value; long duration, small flow control value; long duration, large flow control value; short duration, large flow control value; S6: Using causal recommendation technology, based on the similarity between the traffic situation of the flight traffic data to be analyzed and the historical traffic situation, the recommended traffic management strategy is output and sorted by similarity; The S1 includes: S1-1: Obtaining a flight plan, monitoring data, flow control data, and control rules; the flight plan includes the departure airport, route, and landing airport of each aircraft; S1-2: Establish an airport network layer G1 (V according to the departure airport, route and landing airport d ,V a ,E1); where V d is the departure airport collection, V a It is the set of landing airports, and the route string composed of waypoints between E1 take-off and landing airports; S1-3: Establish the association relationship between the airport approach point and the airport according to the monitoring data, and establish the airport approach network layer G2 (V a ,V e ,E2); where V a It is to meet at the airport, V e is a set of approach points, where the approach point is the point where area control is handed over to approach control in the control command, and E2 is the line between the airport and the approach point; S1-4: Determine the flow control points of adjacent control intervals and the network layer between the flow control points and the airport approach point according to the flow control data and the control rules, G3(V e ,V c ,E3); where V e is the set of entry points, V c is a set of flow control points, where the flow control points are points for transferring and limiting the operating flow between the control of two adjacent areas, and E3 is a line between the flow control points and the entry points; S1-5: Determine the network layer between the flow control point and the departure airport and the route according to the flow control data, G4(V c ,V d ,E4); where V c is the set of air traffic flow management control points, V d is the set of departure airports, and E4 is the line between the flow control point and the departure airport; S1-6: Construct an air traffic static network model G(V according to the network layer established in the above steps a ,V e ,V c ,V d ,E1,E2,E3,E4); The S2 includes: S2-1: Establishing a dynamic air traffic flow model for airports and key waypoints based on the air traffic static network model, and generating flight flow data; the flight flow data includes flow data of airports, approach points and flow control points, and capacity values ​​of airports and flow control points; S2-2: Generate the flow and capacity change trend of each flow control point i in each time slice t according to the flight flow data; the flow and capacity change trend includes: flow Traffic change trend The proportion of traffic at flow control point i Capacity value Capacity change trend Capacity-current ratio The S3 includes: S31: According to the flow and capacity change trend of each piece of the flight flow data, the capacity-flow ratio The matching degree between flow and capacity; S32: When the matching degree of the flight flow data is less than a preset threshold, the flight flow data is capacity-flow imbalance, and the controller generates a flow management strategy for the flight flow data according to the flow and capacity change trend; otherwise, the flight flow data is capacity-flow balance; S33: Output the flight flow data of capacity-flow imbalance and the corresponding flow management strategy.

2. The clustering-based flight flow management strategy recommendation method according to claim 1, characterized in that: The clustering analysis of S5 includes K-means clustering, hierarchical clustering or density-based clustering.

3. The clustering-based flight flow management strategy recommendation method according to claim 1, characterized in that: The flow rate change trend Includes: rise, hold, fall; When t=1, the flow change trend is set to maintain; When t>1, the flow change is calculated as: in, For time slice t-1 flow rate at that time; The capacity change trend Includes: rise, hold, fall; When t=1, the capacity change trend is set to maintain; When t>1, the capacity change trend is calculated as: in, For time slice t-1 The capacity when.

4. The clustering-based flight flow management strategy recommendation method according to claim 1, characterized in that: The S4 includes: S4-1: Preprocessing the traffic management policy; the preprocessing includes deleting duplicate, erroneous, and traffic management policies with incomplete key information; S4-2: Establish a traffic management policy feature set, and classify the pre-processed traffic management policies into corresponding traffic management policy feature sets according to the features of the traffic management policies.

5. The clustering-based flight flow management strategy recommendation method according to claim 1, characterized in that: The S6 includes: S6-1: inputting flight flow data to be analyzed, and using causal recommendation technology to obtain the flow situation of the flight flow data to be analyzed; S6-2: Based on the similarity between the traffic situation and the historical traffic situation, the recommended traffic management strategy is output and sorted by similarity.

6. The clustering-based flight flow management strategy recommendation method according to claim 5, characterized in that: The S6 also includes S6-3: S6-3: Determine whether the number of currently affected flights is greater than 0; when the number of affected flights is 0, stop calculating; The calculation formula for the number of affected flights is: in, For time slice t The number of flights affected by For time slice t Hourly flow, For time slice t-1 Hourly flow, For time slice t The capacity of flow control point i, T Duration of the flight flow management strategy.

7. An electronic device, characterized in that: It includes at least one processor and a memory communicatively connected to the at least one processor; the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the method described in any one of claims 1 to 6.

Images