Method and device for flight seat re-allocation optimization based on decision tree pruning strategy

Abstract

The application discloses a flight seat re-allocation optimization method and device based on a decision tree pruning strategy, belongs to the technical field of civil aviation seat management, and comprises the following steps: obtaining historical data related to seat re-allocation decision; extracting key features from the historical data according to business logic, and calculating information gain values of the key features; selecting a key feature with the largest information gain value as a root node of a decision tree, and adjusting the decision tree; and performing a seat re-allocation process on flight seats according to the adjusted decision tree. The adjusted decision tree reduces system resource consumption of seat selection business, is beneficial to smooth operation of the system, reduces seat protection concurrent conflict during flight season change, and better guarantees smooth development of flight and order business.

Description

Technical Field

[0001] This application belongs to the field of civil aviation seat management technology, and specifically relates to a method and apparatus for optimizing flight seat reallocation based on decision tree pruning strategy. Background Technology

[0002] Currently, civil aviation reservation and departure systems use flight segment / seat maps to store seat occupancy information, and also record passenger / flight segment seat reservation information in passenger orders, which corresponds to the seat occupancy information in the seat map.

[0003] The order rebook function is mainly used in the passenger protection process when passengers reschedule, upgrade, or change their flights. After the flight segment data of the order is updated, seat protection will be triggered, and seats will be reassigned to passengers / flight segments within the order who have rebooked.

[0004] As passengers' demands for personalized services continue to rise, the Rebook function has evolved to include the ability to modify individual PNR orders on flights (hereinafter referred to as the Rebook PNR function). The seat reassignment calculation for PNRs involves data such as orders, main maps, seat maps, and seat class rules for relevant flights / cabins, placing high demands on system resources. During seasonal flight changes, seat protection transactions have repeatedly become congested, impacting seat selection. Therefore, it is necessary to optimize the seat reassignment strategy to reduce its system resource requirements.

[0005] Statistics show that the average onboard seat reservation (ASR) rate is less than 60% of the load factor, especially on less popular domestic routes, where the ASR rate generally does not exceed 50% of the load factor. On popular international routes, the ASR rate generally does not exceed 70% of the seat utilization rate for that segment. Currently, the method of implementing seat protection after the PNR segment is updated in all rebook PNR scenarios does not consider whether the passengers / segments of the rebooked PNR have reserved an ASR. This existing technology has the following shortcomings: Resource waste issue: 70% of RebookPNR transactions trigger invalid seat protection; System bottleneck: During the seasonal flight changeover, the number of concurrent transactions exceeds the system's capacity, causing delays. Business contradiction: The seat reservation rate (≤60%) does not match the protection execution rate (100%). Summary of the Invention

[0006] To address the aforementioned issues, this application employs a pruning strategy-based approach to analyze and optimize the seat reallocation mechanism triggered by Rebook PNR. First, relevant data on seat reallocation decisions is collected and features are extracted. Then, information gain is calculated for each feature, and based on the information gain maximization principle (ID3 algorithm): Feature selection is based on whether the rebooked passenger / flight segment within the order has a reserved onboard seat. Before reallocating seats for the current order, it analyzes whether the flight segment / passenger of the current order has a reserved onboard seat. Seat reallocation is only performed on orders with reserved onboard seats; orders without reserved onboard seats do not require reassignment. This information gain maximization-based seat reallocation optimization reduces the number of seat protection transactions within the system by approximately 28%, reducing system resource consumption for seat selection and contributing to stable system operation. It also reduces seat protection concurrency conflicts during seasonal flight changes, better ensuring the smooth operation of flight and order processes.

[0007] This application provides a method for optimizing flight seat reassignment based on a decision tree pruning strategy, including: Obtain historical data related to seat reassignment decisions; Extract key features from historical data based on business logic, and calculate the information gain value of each key feature; Select the key feature with the largest information gain value as the root node of the decision tree, and adjust the decision tree accordingly; The process of reallocating flight seats is based on the adjusted decision tree.

[0008] Optionally, the historical data includes one or more of the following: Flight data, flight segment data, PNR order data, or seat protection transaction data.

[0009] Optionally, the key feature with the largest information gain value is: Does the passenger / flight segment in the order have a reserved seat on board?

[0010] Optionally, the process of reallocating flight seats based on the adjusted decision tree includes: Read PNR order information from historical data to determine whether the PNR has booked an onboard seat for the From segment; If the PNR has not booked an onboard seat for the From segment, the process will end directly without triggering seat protection. If the PNR has booked an onboard seat for the From segment, then read the data for the From segment and the To segment and perform flight-level and segment-level verification. If the verification passes, the flight seats will be reassigned according to the seat class rules, and the PNR information and seat map will be updated.

[0011] Based on the same inventive concept, this application also provides an apparatus for optimizing flight seat reassignment based on a decision tree pruning strategy, comprising: The data collection unit is used to acquire historical data related to seat reassignment decisions; The data analysis unit is used to extract key features from historical data based on business logic and calculate the information gain value of each key feature. The decision tree pruning unit is used to select the key feature with the largest information gain value as the root node of the decision tree and adjust the decision tree. The seat reassignment unit is used to perform the seat reassignment process for flights based on the adjusted decision tree.

[0012] Optionally, the historical data includes one or more of the following: Flight data, flight segment data, PNR order data, or seat protection transaction data.

[0013] Optionally, the key feature with the largest information gain value is: Does the passenger / flight segment in the order have a reserved seat on board?

[0014] Optionally, the seat reassignment unit specifically includes: The PNR data read / write component is used to read PNR order information from historical data; The PNR-level verification component is used to determine whether the PNR has reserved an onboard seat for the From segment; Flight-level verification component, used to perform flight-level verification on From and To segments; The flight segment and verification component are used to perform segment-level verification on the To flight segment. The seat reassignment component is used to reassign seats on a flight according to seat class rules; The PNR change information synchronization component is used to update PNR information and seating charts.

[0015] Furthermore, this application also provides a computing device, comprising: at least one processor and a memory; The memory is used to store one or more programs; When the one or more programs are executed by the one or more processors, a point cloud intelligent cropping and dynamic optimization method based on edge computing, as described above, is implemented.

[0016] In another aspect, this application also provides a computer-readable storage medium having a computer program stored thereon, which, when executed, implements the point cloud intelligent cropping and dynamic optimization method based on edge computing as described above.

[0017] Compared with the prior art, this application has the following advantages: This application provides a method and apparatus for optimizing flight seat reassignment based on a decision tree pruning strategy. The method includes: acquiring historical data related to seat reassignment decisions; extracting key features from the historical data according to business logic and calculating the information gain value of each key feature; selecting the key feature with the largest information gain value as the root node of the decision tree and adjusting the decision tree; and performing a seat reassignment process based on the adjusted decision tree. By optimizing seat reassignment based on the principle of maximizing information gain, the system reduces the volume of seat protection transactions by approximately 28%, reducing system resource consumption for seat selection and facilitating stable system operation. It also reduces concurrent conflicts in seat protection during seasonal flight changes, better ensuring the smooth operation of flight and order processing.

[0018] Other features and advantages of this application will be set forth in the description which follows, and will be apparent in part from the description, or may be learned by practicing the application. The objectives and other advantages of this application may be realized and obtained by means of the structures pointed out in the description, claims and drawings. Attached Figure Description

[0019] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0020] Figure 1 This paper presents a schematic diagram of the method for optimizing flight seat reassignment based on decision tree pruning strategy provided in this application. Figure 2 This illustrates the original seat reassignment decision tree provided in this application; Figure 3 The decision tree for the adjusted seat reassignment provided in this application is shown; Figure 4 The diagram shows the seat protection transaction volume from February 1, 2024 to March 5, 2024, as provided in this application; Figure 5 The diagram showing the seat protection transaction volume from February 1, 2025 to March 5, 2025, provided in this application is illustrated. Figure 6 The diagram shows the onboard seat reassignment optimization components based on pruning strategies provided in this application; Figure 7 A flowchart of the seat protection system based on pruning strategy provided in this application is shown. Detailed Implementation

[0021] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0022] Example 1 This application provides a method for optimizing flight seat reassignment based on a decision tree pruning strategy, such as... Figure 1 ,include: Step 1: Obtain historical data related to seat reassignment decisions; Step 2: Extract key features from historical data based on business logic, and calculate the information gain value of each key feature; Step 3: Select the key feature with the largest information gain value as the root node of the decision tree and adjust the decision tree; Step 4: Perform seat reallocation process for flights based on the adjusted decision tree.

[0023] Step 1: Obtain historical data related to seat reassignment decisions Optionally, the historical data includes one or more of the following: Flight data, flight segment data, PNR order data, or seat protection transaction data.

[0024] Step 2: Extract key features from historical data based on business logic, and calculate the information gain value of each key feature. This solution innovatively introduces the ID3 algorithm into decision optimization in the aviation field, by calculating the information gain value of each feature:

[0025] in, Let H(S) represent the information gain of dataset S based on feature A, H(S) represent the entropy of dataset S, S represent the rebook PNR scene set, and A represent the feature to be evaluated. This indicates that feature A in dataset S takes values ​​of 10 ... a subset of the dataset, Representing a subset of datasets entropy, Representing a subset of datasets The proportion of the total number of samples in the original dataset S. Represents the set of all possible values ​​for feature A. Based on collected historical data, assuming the seat protection is triggered by a Rebook PNR, the following are the influencing factors and their probabilities at the flight / segment / order level:

[0026] Since flights consist of segments, and orders are based on segment bookings, we use a decision tree to represent the current decision-making process regarding whether to reallocate seats, following a descending order of granularity. Figure 2 : Root node: Flight verification (IG=0.08) Second layer: Flight segment verification (IG=0.05) Third layer: Whether a seat is reserved for this flight segment within the PNR (information gain 0.42) Leaf node: Execute seat protection judgment depth 1*0.08+2*0.05+3*0.87==2.79 Calculations show that the information gain of the feature "whether passengers / flight segments within the order have booked onboard seats" is 0.42, which is significantly higher than that of flight-level features (0.08) and flight segment-level features (0.05).

[0027] We use the features that contribute the most to the model to divide the nodes. There are many evaluation criteria for contribution, and we use information gain: Information gain is the difference between the entropy value before data division and the average entropy value after node division. The larger the difference, the better, that is, the larger the information gain, the better.

[0028] Step 3: Select the key feature with the largest information gain value as the root node of the decision tree, and adjust the decision tree. Optionally, the key feature with the largest information gain value is: Does the passenger / flight segment in the order have a reserved seat on board?

[0029] For seat protection triggered by Rebook PNR, the system determines whether seat reassignment is necessary in the order of flight level / segment level / order level. If seat reassignment is not necessary, the system will prune the seat protection and not activate it. For scenarios where Rebook PNR triggers seat protection, the information gain of flight-level and segment-level features is <10%, while the information gain of order-level features is >40%. The feature with the greatest information gain is whether the passenger / segment of the order has reserved a seat, which has a far greater impact on decision-making than flight / segment-level features. Statistics show that the average onboard seat reservation rate for orders is less than 60% of the load factor. Not reassigning seats to passengers / segments without reservations can directly reduce the overall seat protection transactions in the system by about 30%, and reduce the reading, writing, and calculation of corresponding flight / seat map / seat class rule data. Select the feature with the greatest information gain: whether there are seat reservations for passengers / flight segments within the order, as the root node, and adjust the decision tree as follows: Figure 3 : Root node: Whether a seat is booked for this flight segment within the PNR (Information gain 0.42) Second layer: Flight verification (IG=0.08) Third layer: Flight segment verification (IG=0.05) Leaf node: Execute seat protection judgment depth 1*0.42+2*0.08+3*0.5==2.08 Step 4: Reassigning flight seats based on the adjusted decision tree. Optionally, the seat reassignment process based on the adjusted decision tree includes: Read PNR order information from historical data to determine whether the PNR has booked an onboard seat for the From segment; If the PNR has not booked an onboard seat for the From segment, the process will end directly without triggering seat protection. If the PNR has booked an onboard seat for the From segment, then read the data for the From segment and the To segment and perform flight-level and segment-level verification. If the verification passes, the flight seats will be reassigned according to the seat class rules, and the PNR information and seat map will be updated.

[0030] The advantages of this invention compared to existing technologies are that it comprehensively fills the gaps in current technologies: This invention is the first of its kind in the aviation information industry, applying a pruning strategy to optimize the seat reassignment decision tree for the first time, reducing the amount of flight seat protection transactions in the field of flight seat management. It reduces the system resource consumption of seat protection triggered by Rebook PNR by more than 40%, and reduces the number of seat protection transactions within the system by about 30% overall; for the Rebook PNR scenario, it is the first to extract the features affecting seat reassignment decisions based on business logic and business data statistics. Instead of using the original flight-segment-order order to determine whether seat reassignment is needed, it selects the features of the order based on the principle of maximizing information gain, thereby reducing seat reassignment by more than 30% at the order-level verification. The concurrency of seat protection during seasonal flight changes was reduced. At the end of February 2025, although the number of flights increased by 2% compared to 2025, the number of seat protection transactions decreased by 16%. This is because the proportion of RebookPNR decreased during seasonal flight changes, and flight changes became the main reason for triggering seat protection. There was no congestion of seat protection transactions during seasonal flight changes as in the past. The optimization effect after March was obvious, which improved the satisfaction of passengers and airlines and has great practical value.

[0031] Since the implementation of the pruning-based onboard seat reallocation optimization on February 20, 2025, comparisons have been made with seat protection data from the same period in 2024 (such as...). Figure 4 AR1 represents the original flight seat protection transaction volume, and AR2 represents the seat protection transaction volume under this scheme. The average daily number of seat protection transactions decreased by 28% in 2025 (e.g., Figure 5 Furthermore, considering that the average daily number of PNRs in the system increased by 2% in 2025 compared to 2024 (based on flight and order statistics within the system), the average daily number of seat protection transactions in the system decreased by 28.5%, demonstrating a significant optimization effect.

[0032] Example 2 Based on the same inventive concept, this application also provides an apparatus for optimizing flight seat reassignment based on a decision tree pruning strategy, comprising: The data collection unit is used to acquire historical data related to seat reassignment decisions; The data analysis unit is used to extract key features from historical data based on business logic and calculate the information gain value of each key feature. The decision tree pruning unit is used to select the key feature with the largest information gain value as the root node of the decision tree and adjust the decision tree. The seat reassignment unit is used to perform the seat reassignment process for flights based on the adjusted decision tree.

[0033] The seat reallocation unit specifically includes: The PNR data read / write component is used to read PNR order information from historical data; The PNR-level verification component is used to determine whether the PNR has reserved an onboard seat for the From segment; Flight-level verification component, used to perform flight-level verification on From and To segments; The flight segment and verification component are used to perform segment-level verification on the To flight segment. The seat reassignment component is used to reassign seats on a flight according to seat class rules; The PNR change information synchronization component is used to update PNR information and seat charts, and to notify the airline's main control / SEAT of any changes to the PNR information.

[0034] like Figure 6 This system is divided into three stages: data collection, data reading and verification, and seat protection. 1. The data collection phase includes data collection / analysis components: collecting and analyzing flight / segment / PNR / seat protection transaction-related data within the system; extracting features by combining existing business logic; 2. The data reading and verification stage includes: Flight data read / write component: Reads and writes flight and flight segment data based on the current flight number / date; Flight-level verification component: Verifies whether the flight is managed by the flight information system, whether there are booked PNR passenger information on the flight, and whether the flight has the in-flight seat selection service enabled; Seat business data read / write component: Reads and writes main map, seat class rules, seat map, system configuration and other data corresponding to flight segment and aircraft type layout; Flight segment verification component: Verifies whether there are confirmed PNR passenger information on the flight segment and whether the in-flight seat selection service is enabled on the current flight segment. PNR data read / write component: For a specific PNR of this flight, read and write order information such as flight segment and corresponding onboard seat (ASR) reservation information; PNR-level verification component: Determines whether the PNR contains seat information for the corresponding seat on board for the current flight segment, whether the flight segment reservation is valid, and whether the seat exists on the seating chart for the flight segment / class of service, etc. 3. The seat protection phase includes a seat reassignment component and a PNR change information synchronization component.

[0035] like Figure 7 This is a flowchart of the seat protection system based on the Seat Reassignment Process in this device. 1. The user rebooks the PNR. After the PNR's flight segment is changed, the PNR information is read. For example, if the PNR does not have a reserved seat for the corresponding flight segment, proceed to step 7. 2. Read the From segment and To segment data and perform flight-level verification. For example, if the From segment is less than or equal to the To segment and has already been initialized, cancel the original seat reservation and proceed to step 7. 3. Perform segment-level verification. For example, if the status of segment To is UN (invalid), proceed to step 7. 4. Read the From segment / seat chart data / seat class rule data to calculate the class and attributes of the original seat number in the PNR; 5. Implement seat protection, read seat map data / seat class rules and other seat selection business-related data, reassign seats and modify PNR and seat map; 6. Synchronize PNR and seat change information with the airline's main control and peripheral systems; 7. End and return; Example 3 Based on the same inventive concept, this application also provides an electronic device. The electronic device of this application includes at least one processor and at least one storage medium electrically connected to the processor. The storage medium is electrically connected to the processor, wherein the storage medium stores instructions executable by the at least one processor, which, when executed by the at least one processor, enables the at least one processor to perform the method described above.

[0036] Example 4 Based on the same inventive concept, this application also provides a storage medium storing instructions executable by at least one processor, the instructions being executed by at least one processor to enable at least one processor to perform the method described above.

[0037] Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application.

Claims

1. A method for optimizing flight seat reassignment based on decision tree pruning strategy, characterized in that, include: Obtain historical data related to seat reassignment decisions; Extract key features from historical data based on business logic, and calculate the information gain value of each key feature; Select the key feature with the largest information gain value as the root node of the decision tree, and adjust the decision tree accordingly; The process of reallocating flight seats is based on the adjusted decision tree.

2. The method according to claim 1, characterized in that, The historical data includes one or more of the following: Flight data, flight segment data, PNR order data, or seat protection transaction data.

3. The method according to claim 1, characterized in that, The key characteristic of the largest information gain value is: Does the passenger / flight segment in the order have a reserved seat on board? 4. The method according to claim 1, characterized in that, The process of reallocating flight seats based on the adjusted decision tree includes: Read PNR order information from historical data to determine whether the PNR has booked an onboard seat for the From segment; If the PNR has not booked an onboard seat for the From segment, the process will end directly without triggering seat protection. If the PNR has booked an onboard seat for the From segment, then read the data for the From segment and the To segment and perform flight-level and segment-level verification. If the verification passes, the flight seats will be reassigned according to the seat class rules, and the PNR information and seat map will be updated.

5. A device for optimizing flight seat reassignment based on a decision tree pruning strategy, characterized in that, include: The data collection unit is used to acquire historical data related to seat reassignment decisions; The data analysis unit is used to extract key features from historical data based on business logic and calculate the information gain value of each key feature. The decision tree pruning unit is used to select the key feature with the largest information gain value as the root node of the decision tree and adjust the decision tree. The seat reassignment unit is used to perform the seat reassignment process for flights based on the adjusted decision tree.

6. The apparatus according to claim 5, characterized in that, The historical data includes one or more of the following: Flight data, flight segment data, PNR order data, or seat protection transaction data.

7. The apparatus according to claim 5, characterized in that, The key characteristic of the largest information gain value is: Does the passenger / flight segment in the order have a reserved seat on board? 8. The apparatus according to claim 5, characterized in that, The seat reallocation unit specifically includes: The PNR data read / write component is used to read PNR order information from historical data; The PNR-level verification component is used to determine whether the PNR has reserved an onboard seat for the From segment; Flight-level verification component, used to perform flight-level verification on From and To segments; The flight segment and verification component are used to perform segment-level verification on the To flight segment. The seat reassignment component is used to reassign seats on a flight according to seat class rules; The PNR change information synchronization component is used to update PNR information and seating charts.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, implements the steps of the method for optimizing flight seat reassignment based on a decision tree pruning strategy as described in any one of claims 1-4.

10. An electronic device, characterized in that, It includes a processor, a communication interface, a memory, and a communication bus, wherein the processor, the communication interface, and the memory communicate with each other through the communication bus; Memory, used to store computer programs; When a processor executes a program stored in memory, it implements the steps of the method for optimizing flight seat reassignment based on a decision tree pruning strategy as described in any one of claims 1-4.

Images