Aircraft flight intercourse planning method, system, equipment and medium

Abstract

The invention relates to the technical field of flight planning, and provides an aircraft flight intercourse planning method, system and device and a medium, the method comprises the steps that flight information of multiple cities is utilized to construct a flight information graph, and flight information of each city a in the flight information graph comprises the city a, coordinates of the city a and a city b with direct flights between the city a and the city b; the average duration of the flights from the city a to the city b; constructing a heuristic function, a loss function and an expected loss function of an A * path planning algorithm based on information in the flight information graph, performing path planning by using the A * path planning algorithm, and outputting a planned path from the starting node to the target node; based on the planned path, flight matching is carried out, and a matching result is obtained and output. According to the method, the data in the flight timetable can be fully utilized, so that the planned flight intercourse is more scientific.

Description

Technical Field

[0001] The present invention relates to the technical field of flight planning, and in particular to a method, system, device and medium for planning flight connections. Background Art

[0002] In the aviation industry, recommending suitable connections based on the departure and destination cities input by the user is a core business scenario. Generally speaking, a common solution to this problem is: first generate a sequence of cities that can be directly reached by airplanes (i.e., perform path planning) based on whether there are direct airplanes between cities (or airports, cities are used as examples below), and then match appropriate flights for each city sequence. The present invention mainly optimizes the first step of the problem. This problem is generally modeled as the shortest path problem on a graph, which can be solved by a path search algorithm in graph theory. The following are classic methods and solutions in the field:

[0003] The input of this problem is the user's query request, which includes the user's departure and destination cities. The final output is a sequence of directly accessible cities for the recommendation system to make specific flight matching recommendations and return to the user.

[0004] The generated sequence of cities that can be reached directly needs to have as few transfers as possible and reduce the expected duration of the total connection as much as possible. In order to know the direct flights between cities and estimate the flight time between cities, the current flight schedule needs to be input into the system.

[0005] Generally speaking, the flight schedule should be converted into a graph first. In this graph, each city is a node, and the direct flight between two cities is an edge. You can use an adjacency matrix or adjacency list to store graph data. An adjacency matrix is ​​a two-dimensional array that indicates whether there is a direct flight between any two cities. An adjacency list is an array, each element of which is a list of other cities that have direct flights to the city. Then use a graph search algorithm to find a direct path from one city to another. Common algorithms include: 1. Depth-first search: recursively traverse the graph until the target node is found; 2. Breadth-first search: use a queue to traverse the graph layer by layer until the target node is found; 3. Dijkstra algorithm: suitable for weighted graphs, can find the shortest path, but in this problem, if the weights of direct flights are the same, it can be used to find a direct sequence; 4. A* search algorithm: combines Dijkstra algorithm and heuristic search to find the target faster. Since the flight schedule may change, the system needs to be able to process real-time data and update the data structure of the graph in a timely manner. After completing the route planning, various recommendation algorithms can be used to match specific flights for each segment, and the batch of connections with the highest recommendation scores will be returned to the user.

[0006] In the classic solution, when building a map, only the direct reachability of airplanes between cities is considered. This makes the weights of the edges in the map consistent. Therefore, the flight time between cities cannot be considered in path planning, which will affect the upper limit of the system's subsequent flight matching effect and the overall recommendation effect of the connection generation system. This problem can be alleviated if the distance between cities is used as the edge weight in the graph when building the map, but the distance between cities is not simply proportional to the flight time of the flight, which will lead to certain deviations.

[0007] The above methods do not extract and efficiently use the data in the flight schedule, resulting in poor planned flight connections. Summary of the invention

[0008] To solve the above problems, the present disclosure provides an aircraft flight connection planning method, system, device and medium, which adopts A* algorithm for path planning based on multiple flight factors to obtain better flight connections.

[0009] The following is the content of the present invention:

[0010] A method for planning an aircraft flight connection, characterized by comprising:

[0011] Constructing a flight information graph using flight information of multiple cities, wherein the flight information of each city a in the flight information graph includes: city a, coordinates of city a, city b with direct flights to city a, and average flight duration from city a to city b;

[0012] Use the data of the airline schedule to update the data in the flight information map; obtain the departure city and the target city, with the departure city as the starting node and the target city as the target node;

[0013] Based on the coordinates of the city a node and the target node, the distance between the city a node and the target node is calculated and divided by the flight speed of the aircraft as the heuristic function h(a); a loss function g(b) is constructed, wherein the loss function g(b)=g1+g2(b), wherein g1 is a basic loss added for each city passed; g2(b) is the average duration of a flight directly from city a to city b; an expected loss function f(n) is constructed, wherein f(n)=h(a)+g(b);

[0014] Based on the heuristic function h(a), the loss function g(b) and the expected loss function f(n), the A* path planning algorithm is used to perform path planning, and the planned path from the starting node to the target node is output;

[0015] Based on the planned path, flights are matched, matching results are obtained and output.

[0016] Further,

[0017] The flight information of city a also includes:

[0018] The number of flights from city a and the number of flights directly from city a to city b.

[0019] Further,

[0020] Use the data dictionary to record the flight information of city a;

[0021] The flight information of city a in the data dictionary includes a key and a corresponding key value; wherein the key is the name of city a; the key value is a first tuple; the first digit of the first tuple records the number of flights to city a, the second digit records the coordinates of city a, and the third digit is a first data dictionary of a sub-level;

[0022] The key of the first data dictionary is the name of city b that city a can reach directly, and the key value of the first data dictionary is a second tuple; the first bit of the second tuple records the number of flights directly from city a to city b, and the second bit records the average duration of flights directly from city a to city b.

[0023] Further,

[0024] The method of updating the data in the flight information chart using the data of the aviation timetable comprises:

[0025] Initialize the data dictionary of the flight information graph;

[0026] Read the flight record data of the airline timetable, extract the departure city a and the destination city b involved in the flight, and calculate the flight time t between the cities;

[0027] Search the data dictionary with city a as the key to find city a;

[0028] Search for city b in the first data dictionary in the key value corresponding to city a;

[0029] After finding city b, add 1 to the number of flights to city a in the record of city a; add 1 to the number of flights directly from city a to city b in the first data dictionary of city a, and modify the average duration of flights directly from city a to city b.

[0030] Further,

[0031] The method of using the A* path planning algorithm to plan flight connections and outputting the planned path includes:

[0032] Take the starting node as the current node n1 and calculate the heuristic function h(n1) of the current node n1;

[0033] Cities with direct flights to the departure city are considered as neighboring cities of the departure city, and the loss function g(n2) and heuristic function h(n1) of all neighboring cities are calculated; based on the loss function g(n2) and heuristic function h(n1), the expected loss function f(n) of all neighboring cities is calculated;

[0034] Select the neighbor city n2 with the lowest expected loss function f(n), and record the current node n1 as the parent node of the neighbor city n2;

[0035] Take the neighbor city n2 as the new current node, and repeatedly select neighbor nodes of the new current node until the target node is selected;

[0036] Starting from the target node, trace back to the start node through the parent node link, and construct and output the shortest path from the start node to the target node.

[0037] Further,

[0038] The matching of flights based on the planned path includes:

[0039] For each route, retrieve the flight information that meets the time constraints and has remaining tickets from the flight schedule;

[0040] Use the recommendation algorithm to score and sort each flight;

[0041] Select flights based on the scoring and sorting results to generate multiple connections that meet the user's personalized characteristics.

[0042] Further,

[0043] The recommendation algorithm is a collaborative filtering algorithm or a recommendation algorithm based on a deep neural network.

[0044] An aircraft flight connection planning system, characterized by comprising:

[0045] An information construction module is used to construct a flight information graph using flight information of multiple cities, wherein the flight information of each city a in the flight information graph includes: city a, coordinates of city a, city b with direct flights to city a, and average flight duration from city a to city b;

[0046] The data update module is used to update the data in the flight information map using the data of the aviation schedule; obtain the departure city and the target city, with the departure city as the starting node and the target city as the target node;

[0047] The path planning module is used to calculate the distance between the city a node and the target node based on the coordinates of the city a node and the target node and divide it by the flight speed of the aircraft as the heuristic function h(a); construct a loss function g(b), wherein the loss function g(b)=g1+g2(b), wherein g1 is a basic loss added for each city passed; g2(b) is the average duration of a flight directly from city a to city b; construct an expected loss function f(n), wherein f(n)=h(a)+g(b);

[0048] Based on the heuristic function h(a), the loss function g(b) and the expected loss function f(n), the A* path planning algorithm is used to perform path planning, and the planned path from the starting node to the target node is output;

[0049] The flight matching module is used to match flights based on the planned routes, obtain matching results and output them.

[0050] Compared with the prior art, the present invention has the following advantages:

[0051] The present invention can comprehensively integrate flight data by constructing a flight information map containing information such as cities, coordinates, direct flights, and average flight duration, thus overcoming the problem that traditional map construction cannot consider the upper limit of the matching effect affected by flight time;

[0052] The distance is calculated using the city coordinates and combined with the aircraft speed to construct the heuristic function h(a), which can provide the A* algorithm with a cost estimate from the current node to the target node to guide the search direction and reduce blindness;

[0053] The core purpose of constructing the loss function g(b) is to comprehensively measure the path cost in flight connection planning and guide the algorithm to plan a better solution. From the perspective of the number of transfers, the basic loss increases for each city passed through. This is because the more transfers there are, the greater the uncertainty and inconvenience faced by passengers. This setting encourages the algorithm to give priority to routes with fewer transfers, thereby improving the quality of connections and passenger experience. In terms of flight time, the average duration of flights from city to city is included. Since flight time is a key factor in connection planning and is directly related to the total duration of passengers' journeys, and the average duration of flights between different cities varies significantly, including it can make the algorithm pay more attention to selecting flight combinations with short flight times, reducing the total travel time, so that the path planning can achieve a better balance between the number of transfers and flight time, and improve the scientificity and rationality of the overall flight connection planning;

[0054] The purpose of constructing the expected loss function f(n) is to comprehensively evaluate the quality of the path;

[0055] Guided by these functions, the A* algorithm can efficiently find the best solution and improve the efficiency and quality of path planning. Finally, it can accurately match flights based on the planned path to meet user time and personalized needs, improve user experience, and thus solve the problems in the background technology.

[0056] Other features and advantages of the present disclosure will be described in the following description, and partly become apparent from the description, or be understood by implementing the present disclosure. The purpose and other advantages of the present disclosure can be realized and obtained by the structures pointed out in the description, claims and drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0057] In order to more clearly illustrate the embodiments of the present disclosure or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present disclosure. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0058] Figure 1 A schematic diagram of the method of the present invention is shown;

[0059] Figure 2 A flow chart of the flight information graph maintenance process is shown;

[0060] Figure 3 A flow chart of the joint trip recommendation process is shown. DETAILED DESCRIPTION

[0061] In order to make the purpose, technical solution and advantages of the embodiments of the present disclosure clearer, the technical solution in the embodiments of the present disclosure will be clearly and completely described below in conjunction with the drawings in the embodiments of the present disclosure. Obviously, the described embodiments are part of the embodiments of the present disclosure, not all of the embodiments. Based on the embodiments in the present disclosure, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present disclosure.

[0062] This patented technology extracts and organizes the data in the flight schedule that is valuable for route planning in advance, and uses it to construct a flight information graph for route planning. The solution of the present invention decouples the route planning process from the flight schedule while utilizing the flight schedule data, which can improve the route planning effect while ensuring that the algorithm time consumption is acceptable, thereby improving the overall flight connection recommendation effect.

[0063] In general, the scheme of the present invention needs to maintain a flight information graph, in which the number of flights from each city to other cities and the average flight duration are recorded.

[0064] The flight information graph is recorded in a data dictionary, where the key is the list of cities a, and the corresponding key value is a tuple, where the first digit records the number of flights related to the city; the second digit records the longitude and latitude coordinates of the city; the third digit is a sub-level data dictionary, where the key of this sub-level data dictionary is the list of cities b that city a can reach, and the corresponding key value is a tuple, which records the number and average duration of flights from a to b. These data will be used to calculate the heuristic function and loss function in the A* itinerary planning algorithm.

[0065] This patented method is divided into two major stages, namely, the flight information graph maintenance stage and the connection generation stage;

[0066] The first stage requires extracting information from the flight schedule into a flight information map, and when the flight schedule changes, the flight information map needs to be adjusted;

[0067] The second stage is the stage of actually interacting with users to make connecting flight recommendations. It is necessary to perform A* path planning on the flight information graph and match appropriate flights for each route.

[0068] like Figure 1 A schematic diagram of the method according to the present invention is shown, and the specific implementation details of the present invention include:

[0069] 1. Step 1: System preparation and initialization:

[0070] Step 101: Prepare the flight schedule data to ensure the validity of the data of the city and the time of the flight in each record in the flight schedule;

[0071] Step 102: Initialize the flight information graph data dictionary and create records for the cities involved in the flight schedule, where the key is the city name or number a, and the key value is a tuple, where the first digit records the number of related flights to the city and is initialized to 0, the second digit records the longitude and latitude coordinate pair of the city, and the third digit is a sub-level data dictionary, which records the list of cities b that city a can directly reach, and this data dictionary is initialized to empty.

[0072] 2. Step 2: Initialize the flight information map according to the flight schedule:

[0073] Step 201: read an unprocessed flight record, extract the departure city a and the destination city b involved in the flight, and calculate the flight time t between the cities;

[0074] Step 202: Retrieve and initialize in the flight information graph. Search in the data dictionary flight information graph with city a as the key. If not found, initialize with city a as the key value in the manner of step 102. If found, determine whether the sub-level data dictionary in the key value corresponding to city a contains records in city b. If so, go to step 203. Otherwise, initialize records in city b in the sub-level data dictionary, where the key is city b and its corresponding key value is a tuple, where the first bit records the number of flights that can be directly reached, and the second bit records the average duration of flights from a to b.

[0075] Step 203: Update the records from city a to city b in the flight information diagram. Specifically, increase the number of related flights in the first record of city a by 1; increase the number of direct flights in the first record of city b in the sub-level data dictionary of city a by 1, and modify the average duration in the second record.

[0076] 3. Step 3: Modify the flight information chart according to the changes in the flight schedule:

[0077] Step 301: Read the historical flight information map from the external storage file;

[0078] Step 302: read an unprocessed flight record, extract the departure city a and the destination city b involved in the flight, calculate the flight time t between the cities, and mark whether it is an added flight or a deleted flight;

[0079] Step 303: Modify the flight information map in the same manner as steps 202 to 203;

[0080] Step 304: When the number of flights involved in a city is 0, the record is deleted in the data dictionary. When the number of direct flights from city a to city b is 0 after a flight is deleted, the record of city b is deleted in the sub-level data dictionary under city a.

[0081] 4. Step 4: Store the real-time flight information chart to an external storage file for backup and mark the update time.

[0082] 5. Step 5: Receive the user request from the user interface, extract the departure city and destination city, and record the user's departure time.

[0083] 6. Step 6: Define the heuristic function, loss function and expected loss function:

[0084] Step 601: define a heuristic function h(n1) for estimating the shortest path cost from the current node n1 to the target node. The input of the function is the coordinates of the starting city and the end city. Then the actual distance between the two coordinates is calculated and divided by the general speed of the aircraft and output;

[0085] Step 602: define a loss function g(n2) for calculating the current path loss (hereinafter referred to as g value), where n2 is a historical city, and a basic loss g1 is added for each city passed through to reduce the number of transfers in the result, and the average flight time between cities is queried in the flight information map as the loss g2(n2), g(n2)=g1+g2(n2);

[0086] Step 603: define an expected loss function f(n)=h(n1)+g(n2), where n is the union of n1 and n2, to estimate the expected loss function of the current path (hereinafter referred to as f value), which represents the expected connection time.

[0087] 7. Step 7: Initialization of path planning:

[0088] Step 701: Read the flight information map from the external storage file, and verify whether it is the latest flight information map according to the update time;

[0089] Step 702: Initialize the algorithm open list to store the nodes to be expanded. At the initial moment, add the starting node to the open list and set the f value of the starting node;

[0090] Step 703: Initialize the algorithm closing list, which is used to store nodes that have been completely processed to avoid repeated processing and is initialized to empty.

[0091] 8. Step 8: Path search loop expands the current node:

[0092] Step 801: When the open list is empty, it means that there is no path between the departure city and the destination city that satisfies the requirement that both adjacent cities can be reached directly by plane, and the algorithm ends and returns that the connection planning failed; in other words, if it is assumed that there must be a connection between the input departure city and the destination city, then the situation where the open node is empty will not occur;

[0093] Step 802: When the open list is not empty, select the node with the minimum f value from the open list as the current node; if the current node is the target node, the algorithm ends and returns the path; otherwise, remove the current node from the open list and add it to the closed list;

[0094] Step 803: Expand the current node. For each neighbor m of the current node n, if neighbor m is in the closed list, skip this neighbor; otherwise, calculate the cost from the start node to neighbor m, that is, the cost g(m) to reach neighbor m through the current node; if neighbor m is not in the open list, or the new cost to reach neighbor m through the current node is less than the previous cost, update the g value of neighbor m and add it to the open list. At the same time, record that the current node n is the parent node of neighbor m;

[0095] Step 804: For each updated neighbor m, recalculate its f value and adjust its position in the open list according to the f value.

[0096] 9. Step 9: Output the planned path: If the planning is successful, start from the target node, trace back to the start node through the parent node link, build and return the shortest path from the start node to the target node, and output it to the recommendation algorithm for flight matching.

[0097] 10. Step 10: Match flights for each route:

[0098] Step 1001: for each route segment, retrieve flight information that meets the time constraints and has remaining tickets from the flight schedule;

[0099] Step 1002: Use a recommendation algorithm to score and sort each flight. The recommendation algorithm used in this step can be of various types, such as a traditional collaborative filtering method or the latest deep neural network-based recommendation algorithm. The former is easy to implement, and the latter can provide personalized services. These two recommendation algorithms are widely used in various industries.

[0100] Step 1003: Select flights based on the scoring and sorting results, and generate several connecting flights that meet the user's personalized characteristics.

[0101] 11. Step 11: Output the joint trip planning results to the user: After receiving the joint trip planning recommendation from the backend, the joint trip results are output to the user interface for the user to view and purchase.

[0102] The following is a specific embodiment of the flight information graph maintenance process:

[0103] like Figure 2 This is a flowchart of the flight information map maintenance process. The function of this part is to initialize or change the flight information map to be consistent with the latest flight schedule when the system is initialized or the flight schedule changes. Specifically, if it is system initialization, the flight information must be initialized first. If the module is called due to flight schedule changes, the historical flight information map should be read from the external storage file.

[0104] Then read a flight that needs to be processed, extract the city pairs involved, and calculate the estimated flight time between the city pairs based on the time marked in the timetable. The time loss involved in this step is reading flight information and preprocessing, and the time complexity is O(1).

[0105] After obtaining the city pair and time, search the data dictionary for the corresponding city pair. If not, create a new city pair with the number of flights initialized to 0 and the average flight time to 0 (this data will be overwritten immediately). Then update the average flight time and number of flights. If the data dictionary is implemented using a hash table, the time complexity of this step can be considered to be O(1).

[0106] After processing each item one by one, if the number of flights between any city pair is 0, the city can be deleted (this situation does not happen often in actual scenarios). After all processing is completed, the new flight information graph is stored in an external storage file. In general, if the number of flight changes is n, since this operation is only called when the flight schedule changes, that is, each flight record is only processed once when it changes, rather than being called every time the user queries, the time complexity of this method is O(n), and the time required by the system can be ignored except for the necessary time to read the airline schedule.

[0107] like Figure 3 Recommended process flow chart for joint trip:

[0108] The algorithm generally uses the A* algorithm framework and relies on the flight information graph mentioned above. Specifically, after receiving the departure and destination city pairs input by the user, since only the cases requiring transfers are processed here, if the departure and destination city pairs can be reached directly, and there is an edge between the departure city and the destination city in the corresponding flight information graph, this edge will be blocked (direct flights will be recommended separately).

[0109] Then, based on the distance between cities, the data recorded in the flight information diagram and other information, the heuristic estimation function of the loss between the current visited city and the destination is calculated respectively, and the loss function of the traversed route is calculated. Based on the two, the next hop city is selected. If the destination is reached, the city string is output, otherwise the search continues. After path planning, various recommendation algorithms can be used to select suitable flights for matching. Since this part is not the focus of this patent, it will not be expanded. Any recommendation algorithm can be used in actual use.

[0110] In summary, the present invention designs an efficient path planning method, which dynamically preprocesses the flight schedule. The method pre-efficiently extracts parts valuable for path planning from the large table of the aviation schedule, and stores them in the flight information graph, so as to avoid retrieving and processing the original large table of the flight schedule every time a user query is processed, and uses the A* framework to search in the flight information graph, so as to make the query process more efficient, effectively improve the path planning speed, reduce the query cost of the flight schedule, reduce the response time of the system, improve the user experience, and the system can respond to the dynamic changes of the flight schedule.

[0111] Based on the method of the present invention, the embodiment of the present disclosure also provides a system corresponding to the above method, which includes:

[0112] An information construction module is used to construct a flight information graph using flight information of multiple cities, wherein the flight information of each city a in the flight information graph includes: city a, coordinates of city a, city b with direct flights to city a, and average flight duration from city a to city b;

[0113] The data update module is used to update the data in the flight information map using the data of the aviation schedule; obtain the departure city and the target city, with the departure city as the starting node and the target city as the target node;

[0114] The path planning module is used to calculate the distance between the city a node and the target node based on the coordinates of the city a node and the target node and divide it by the flight speed of the aircraft as the heuristic function h(a); based on the average duration of flights directly from city a to city b, a loss function g(b) is constructed, wherein the loss function g(b)=g1+g2(b), wherein g1 is a basic loss added for each city passed; g2(b) is the average duration of flights directly from city a to city b; based on the heuristic function h(a) and the loss function g(b), an expected loss function f(n) is constructed, wherein f(n)=h(a)+g(b);

[0115] Based on the heuristic function h(a), the loss function g(b) and the expected loss function f(n), the A* path planning algorithm is used to perform path planning, and the planned path from the starting node to the target node is output;

[0116] The flight matching module is used to match flights based on the planned routes, obtain matching results and output them.

[0117] Based on the same inventive concept as the above disclosed content, an embodiment of the present disclosure also provides a device corresponding to the above method, which includes at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to execute the above method.

[0118] It should be noted that the electrical connection between the above-mentioned units does not necessarily mean the connection between lines. An indirect connection method can be applied to the embodiments of the present disclosure as long as the purpose of the present disclosure is achieved.

[0119] Based on the same inventive concept, the present disclosure also provides a computer storage medium on which executable instructions are stored, and the instructions are executed by a processor to execute the above method.

[0120] Although the present disclosure has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present disclosure.

Claims

1. A method for planning flight connections, characterized in that: include: Constructing a flight information graph using flight information of multiple cities, wherein the flight information of each city a in the flight information graph includes: city a, coordinates of city a, city b with direct flights to city a, and average flight duration from city a to city b; Use the data of the airline schedule to update the data in the flight information map; obtain the departure city and the target city, with the departure city as the starting node and the target city as the target node; Based on the coordinates of the city a node and the target node, the distance between the city a node and the target node is calculated and divided by the flight speed of the aircraft as the heuristic function h(a); a loss function g(b) is constructed, wherein the loss function g(b)=g1+g2(b), wherein g1 is a basic loss added for each city passed; g2(b) is the average duration of a flight directly from city a to city b; an expected loss function f(n) is constructed, wherein f(n)=h(a)+g(b); Based on the heuristic function h(a), the loss function g(b) and the expected loss function f(n), the A* path planning algorithm is used to perform path planning, and the planned path from the starting node to the target node is output; Based on the planned path, flights are matched, matching results are obtained and output.

2. The method for planning an aircraft flight connection according to claim 1, characterized in that: The flight information of city a also includes: The number of flights to city a and the number of flights directly from city a to city b.

3. The method for planning an aircraft flight connection according to claim 2, characterized in that: Use the data dictionary to record the flight information of city a; The flight information of city a in the data dictionary includes a key and a corresponding key value; wherein the key is the name of city a; the key value is a first tuple; the first digit of the first tuple records the number of flights to city a, the second digit records the coordinates of city a, and the third digit is a first data dictionary of a sub-level; The key of the first data dictionary is the name of city b that city a can reach directly, and the key value of the first data dictionary is a second tuple; the first bit of the second tuple records the number of flights directly from city a to city b, and the second bit records the average duration of flights directly from city a to city b.

4. The method for planning an aircraft flight connection according to claim 3, characterized in that: The method of updating the data in the flight information chart using the data of the aviation schedule comprises: Initialize the data dictionary of the flight information graph; Read the flight record data of the airline timetable, extract the departure city a and the destination city b involved in the flight, and calculate the flight time t between the cities; Search the data dictionary with city a as the key to find city a; Search for city b in the first data dictionary in the key value corresponding to city a; After finding city b, add 1 to the number of flights to city a in the record of city a; add 1 to the number of flights directly from city a to city b in the first data dictionary of city a, and modify the average duration of flights directly from city a to city b.

5. The method for planning an aircraft flight connection according to claim 1, characterized in that: The method of using the A* path planning algorithm to plan flight connections and outputting the planned path includes: Take the starting node as the current node n1 and calculate the heuristic function h(n1) of the current node n1; Cities with direct flights to the departure city are considered as neighboring cities of the departure city, and the loss function g(n2) and heuristic function h(n1) of all neighboring cities are calculated; based on the loss function g(n2) and heuristic function h(n1), the expected loss function f(n) of all neighboring cities is calculated; Select the neighbor city n2 with the lowest expected loss function f(n), and record the current node n1 as the parent node of the neighbor city n2; Take the neighbor city n2 as the new current node, and repeatedly select neighbor nodes of the new current node until the target node is selected; Starting from the target node, trace back to the start node through the parent node link, and construct and output the shortest path from the start node to the target node.

6. The method for planning an aircraft flight connection according to claim 1, characterized in that: The matching of flights based on the planned path includes: For each route, retrieve the flight information that meets the time constraints and has remaining tickets from the flight schedule; Use the recommendation algorithm to score and sort each flight; Select flights based on the scoring and sorting results to generate multiple connections that meet the user's personalized characteristics.

7. The method for planning an aircraft flight connection according to claim 6, characterized in that: The recommendation algorithm is a collaborative filtering algorithm or a recommendation algorithm based on a deep neural network.

8. An aircraft flight connection planning system, characterized in that: include: An information construction module is used to construct a flight information graph using flight information of multiple cities, wherein the flight information of each city a in the flight information graph includes: city a, coordinates of city a, city b with direct flights to city a, and average flight duration from city a to city b; The data update module is used to update the data in the flight information map using the data of the aviation schedule; obtain the departure city and the target city, with the departure city as the starting node and the target city as the target node; The path planning module is used to calculate the distance between the city a node and the target node based on the coordinates of the city a node and the target node and divide it by the flight speed of the aircraft as the heuristic function h(a); construct a loss function g(b), wherein the loss function g(b)=g1+g2(b), wherein g1 is a basic loss added for each city passed; g2(b) is the average duration of a flight directly from city a to city b; construct an expected loss function f(n), wherein f(n)=h(a)+g(b); Based on the heuristic function h(a), the loss function g(b) and the expected loss function f(n), the A* path planning algorithm is used to perform path planning, and the planned path from the starting node to the target node is output; The flight matching module is used to match flights based on the planned routes, obtain matching results and output them.

9. An aircraft flight connection planning device, comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor so that the at least one processor can execute the aircraft flight connection planning method described in any one of claims 1 to 7.

10. A computer storage medium having executable instructions stored thereon, wherein when the instructions are executed by a processor, the processor is enabled to implement the method according to any one of claims 1 to 7.

Images