Civil airport passenger loading application dynamic planning and distribution method and system based on multi-service data

Abstract

The invention belongs to the technical field of civil aviation flight passenger application data processing, and discloses a civil airport passenger application dynamic planning and distribution method and system based on multi-business data. The method comprises the steps of data acquisition and cleaning processing, feature selection and model training, after data acquisition and cleaning processing, indexes or attributes capable of effectively reflecting inherent characteristics of historical data are selected from the historical data, and a prediction model is constructed by using a deep learning algorithm. The obtained indexes or attributes reflecting the inherent characteristics of the historical data are used as input information to carry out prediction model training; and calculating and issuing data, and verifying and optimizing output. According to the invention, through analysis and calculation based on historical operation data and in combination with related data such as air traffic control, a priority guarantee rule in a passenger loading application process is learned, complex rule maintenance configuration is not needed, dynamic planning processing of future flight passenger loading applications is realized, and airport guarantee efficiency and rationality of resource allocation are improved.

Description

Technical Field

[0001] The invention belongs to the technical field of civil aviation flight boarding application data processing, and in particular relates to a civil airport boarding application dynamic planning allocation method and system for multi-business data. Background Art

[0002] With the development of the aviation transportation industry, airports are facing more and more challenges, especially during peak periods such as summer travel, air traffic control, and large-scale flight delays. Flight delays and stranded passengers often face a series of disastrous consequences. How to efficiently handle passenger boarding applications and ensure that flights can board passengers in time and take off on time has become an important part of airport operations.

[0003] The traditional way of processing passenger applications relies on manual labor, lacks flexibility and intelligence, and often has the following problems: Reliance on manual experience: Traditionally, passenger pick-up applications are usually processed by dispatchers based on their own experience and intuition. This approach may result in unscientific and unreasonable task allocation, and may easily lead to deviations or omissions.

[0004] Inefficient dispatching: Dispatching efficiency is greatly affected by the need to repeatedly confirm progress, especially during busy hours. Dispatchers may need to frequently use walkie-talkies or other communication tools to coordinate work, which not only wastes time but also increases communication costs.

[0005] Difficulty in fully understanding on-site conditions: Due to the ever-changing on-site conditions, it is difficult for dispatchers to fully understand the on-site operation conditions in real time, which may lead to insufficient information when making dispatch decisions, thus affecting the quality of dispatch.

[0006] Lack of flexibility: Traditional methods have limited room for adjustment when faced with emergencies and are difficult to adapt to changes quickly. For example, when flight delays or cancellations occur, resources need to be re-coordinated. Traditional methods often seem unable to cope with such situations.

[0007] With the continuous advancement of information technology, big data technology and artificial intelligence technologies such as deep learning have been widely used in various fields, providing new ways and solutions to solve the problems existing in traditional methods. In the field of civil aviation, big data technology can help collect and process a large amount of flight data, airport operation data, weather information and route information, providing decision makers with more comprehensive and accurate information support. Deep learning technology can automatically extract useful features from massive data through training models, and make predictions and decisions based on them, thereby improving the operating efficiency and management level of the civil aviation system. Summary of the invention

[0008] In order to overcome the problems existing in the related art, the disclosed embodiments of the present invention provide a method and system for dynamically planning and allocating civil airport boarding applications based on multi-business data, and in particular, a method for dynamically planning and processing civil aviation flight boarding applications based on big data and deep learning. The purpose of the present invention is to provide a method for dynamically planning and processing airport boarding applications based on multi-business data, aiming to improve the efficiency and quality of boarding application processing through automated means, and meet the growing needs brought about by the gradual development of the civil aviation industry.

[0009] The technical solution is as follows: A method for dynamically planning and allocating multi-business data for civil airport boarding applications, comprising: S1, data collection and cleaning processing: collect historical data from multiple data sources, perform data cleaning and data transformation; S2, feature selection and model training: select indicators and attributes that reflect the intrinsic characteristics of historical data from historical data, use deep learning algorithms to build prediction models, and use the obtained indicators and attributes as input information to train the prediction model; S3, data calculation and release: The prediction model is trained through historical data, and the prediction model is applied to the production scenario of flight boarding application. Real-time dynamic flight information and air traffic control (ATC) information are collected and input into the prediction model. The prediction model combines historical data analysis, processes boarding applications, predicts the optimal boarding time and boarding order, and ensures the flight regularity rate; S4, output verification and optimization: Use statistical methods to automatically screen out result items that do not conform to the normal pattern, mark them, continuously update the data resource pool, and eliminate interference items.

[0010] Furthermore, historical data from various data sources are collected for data cleaning and data transformation, including: Collect historical data from different data sources, which are obtained through database extraction, API interface calls, and web crawling; Data cleaning, including: removing redundant information, consistency checking, outlier detection and correction; Data transformation, including: creating new features by performing mathematical operations on existing attributes; adjusting the value range through standardization and normalization.

[0011] Furthermore, historical data includes: flight number, aircraft type, aircraft number, route, area, mission, VIP, aircraft position, boarding gate, support completion time, planned landing time, actual landing time, planned take-off time, actual take-off time, boarding start time, boarding deadline, door opening and closing time, boarding time, COBT, CTOT, weather information, route restrictions, flow control, and delay information.

[0012] In step S2, feature selection includes: based on civil aviation field knowledge and exploratory data analysis, excluding obviously irrelevant or noisy data fields, calculating the Pearson coefficients between variables and between each variable and the target variable, and finding the existence of strong correlations; combining machine learning algorithms to iteratively remove unimportant features until the ideal scale is reached; the calculation formula of the Pearson coefficient is: In the formula, is the Pearson coefficient, are the values ​​of the two variables in the sample, is the mean value of the two variables in the sample, is the sample size; The Pearson coefficient calculation formula is used to calculate the linear correlation between two variables, and its value is between -1 and 1; When it is close to 1 or -1, it indicates that there is a strong linear relationship between the two variables; when When it is close to 0, it means that there is almost no linear relationship between the two variables; positive values ​​indicate positive correlation, and negative values ​​indicate negative correlation.

[0013] In step S2, a prediction model is constructed using a deep learning algorithm, and the obtained indicators and attributes are used as input information to perform prediction model training, including: (a) Dataset division: The dataset is divided into training set, validation set and test set in a ratio of 8:1:1; (b) Constructing the Transformer model: The Transformer model is selected as the underlying architecture of the deep learning algorithm. The features of historical flights (flight number, take-off and landing airports, take-off and landing times, COBT, weather conditions, and traffic information) are converted into embedding vectors for use by the Transformer encoder. A 6-layer encoder is set up with 8 heads per layer to form a deep Transformer model network. The self-attention mechanism is used to capture the dependencies in the historical flight data to capture complex features. (c) Reinforcement learning strategy optimization: Using Transformer as the basic architecture, reinforcement learning elements RL are introduced for strategy optimization; (d) Output verification and optimization: After the prediction model training is completed, the output results of the model are verified.

[0014] In step (c), we use Transformer as the basic architecture and introduce reinforcement learning elements for policy optimization, including: Define the state space, including: the number of passengers on the current flight, the waiting time of passengers, and the flight delays; the action space consists of the decision to allow the flight to start boarding passengers, and the reward function rewards or punishes by evaluating the waiting time of passengers and the flight release regularity index; Design a deep Q network as an intelligent agent to learn the best boarding decision strategy and maximize the cumulative reward by optimizing the boarding process. Combined with the experience replay mechanism, collect the experience tuples related to flight boarding and store them in the experience replay buffer for training the deep Q network. The RL module works together with the Transformer module, which processes flight boarding application data. Through feature extraction, standardization, and time series data processing, it models the boarding sequences of different flights, captures long-distance dependencies between input sequences, and performs big data prediction.

[0015] In step (d), the output results of the model are verified, including: checking the fit of the model on the historical data through accuracy, mean absolute error, and mean square error.

[0016] In step S3, the calculated result includes the pick-up application result information, pick-up time, pick-up location, and pick-up sequence; the result set is output to the airport management system, air traffic control system, and airline system for user query and reference.

[0017] In step S4, after the prediction model is launched, a trial period is set up, during which a manual review mechanism is retained to compare the differences between the two. Statistical methods are used to automatically screen out result items that do not conform to the normal pattern and mark them for review and confirmation. The data resource pool is continuously updated, the internal knowledge graph content is continuously improved, and interference items are removed.

[0018] Another object of the present invention is to provide a system for dynamically planning and allocating civil airport boarding applications with multiple business data, the system implementing the method for dynamically planning and allocating civil airport boarding applications with multiple business data, the system comprising: Data collection and cleaning processing module, used for data collection and preliminary cleaning and processing of data; Feature selection and feature extraction module, which is used to select the most influential subset of historical data for the prediction target, and extract new feature representations that can better reflect the internal laws by transforming or mapping historical data; The data modeling and model training module is responsible for training the prediction model based on the extracted feature values. It uses deep learning algorithms to build prediction models based on the flight boarding application business scenario. It also makes plans for future flight boarding applications through analysis and training of historical data. The decision support or predictive analysis module applies the trained prediction model to the actual production process. By comprehensively analyzing factors such as flight history records, passenger flow changes, and weather conditions, it predicts possible delays or congestion in advance and generates the optimal scheduling plan accordingly.

[0019] Combining all the above technical solutions, the beneficial effects of the present invention are as follows: the present invention is oriented to the field of civil aviation airport operation management technology, and through analysis and calculation based on historical operation data, combined with relevant data such as air traffic control, the priority protection rules in the passenger boarding application process are learned, and no complex rule maintenance configuration is required, so as to realize dynamic planning processing of future flight passenger boarding applications, and improve airport protection efficiency and rationality of resource allocation. It is specifically reflected in the following aspects: Speed ​​up passenger clearance: By accurately predicting the boarding order of flights during peak hours and adjusting the number of open security checkpoints based on real-time data, passengers' waiting time in queues is effectively reduced.

[0020] Optimize flight scheduling process: Use advanced prediction algorithms to identify factors that may affect flight punctuality (such as weather changes, aircraft maintenance needs, flow control reasons, etc.) in advance, and formulate flexible scheduling plans accordingly to ensure flights take off and land on time.

[0021] Improve baggage handling capacity: Based on in-depth analysis of past data, it is possible to more accurately estimate the amount of baggage that needs to be handled for each flight, and then rationally allocate relevant equipment and personnel to avoid baggage backlogs.

[0022] Enhanced emergency response capabilities: When faced with emergencies (such as severe weather causing a large number of flight cancellations or delays), the system can quickly generate emergency plans and coordinate resources from all parties for efficient handling to minimize the scope of negative impacts.

[0023] Optimal resource allocation: By analyzing historical data, we can better understand the usage of various airport facilities in different time periods, so as to achieve refined management of key resources such as parking spaces and boarding gates, and avoid idle resources or overcrowding.

[0024] Promoting multi-party collaboration: By establishing a unified data sharing platform, information exchange and cooperation between airlines, air traffic control departments and ground service providers are strengthened, forming an efficient and collaborative working mechanism.

[0025] Continuous improvement mechanism: As more real-time feedback data accumulates, the system will have the ability to self-learn and continuously adjust and improve the algorithm logic to ensure that a high level of service quality standards is maintained in the long term. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] The accompanying drawings herein are incorporated in and constitute a part of the specification, illustrate embodiments consistent with the present disclosure, and together with the description, serve to explain the principles of the present disclosure; Figure 1 It is a flow chart of a method for dynamically planning and allocating civil airport boarding applications for multi-service data provided by an embodiment of the present invention; Figure 2 It is a schematic diagram of a method for dynamically planning and allocating civil airport boarding applications for multi-service data provided by an embodiment of the present invention; Figure 3 It is a schematic diagram of a civil airport boarding application dynamic planning allocation system for multi-service data provided by an embodiment of the present invention; Figure 4 is a schematic diagram of a Transformer model provided by an embodiment of the present invention; In the figure: 1. Data collection and cleaning processing module; 2. Feature selection and feature extraction module; 3. Data modeling and model training module; 4. Decision support or predictive analysis module. DETAILED DESCRIPTION

[0027] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific embodiments of the present invention are described in detail below in conjunction with the accompanying drawings. In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention can be implemented in many other ways different from those described herein, and those skilled in the art can make similar improvements without violating the connotation of the present invention, so the present invention is not limited by the specific implementation disclosed below.

[0028] The innovation of the present invention is that the present invention comprehensively considers various business information in the passenger boarding application data of civil airports, such as flight information, resource usage, flow control, delays, etc., filling the gaps in multi-source data fusion, real-time processing, intelligent decision-making, real-time monitoring and feedback in the domestic civil aviation field. Through dynamic planning and optimization allocation, the passenger boarding process is flexibly adjusted to maximize resource utilization and improve efficiency. In addition, by combining big data and deep learning technology, intelligent decisions are made based on historical data and real-time conditions, and personalized solutions are provided, which not only improves passenger satisfaction, but also improves airport operation assessment indicators. In addition, the system has a real-time monitoring function, which can track the passenger boarding process at any time and respond quickly and adjust dynamically to ensure smooth passenger boarding. In summary, this method brings a new solution to improve passenger boarding efficiency and passenger satisfaction, and promotes civil airport operations to move towards intelligence, real-time and personalization.

[0029] Example 1: The method for dynamically planning and allocating multi-business data for civil airport boarding applications provided in the embodiment of the present invention is based on the calculation and analysis of a large amount of historical data. Through deep learning, a dynamic planning processing method for flight boarding applications is proposed, including: data collection and cleaning, feature selection and model training, data calculation and release, and output verification and optimization.

[0030] like Figure 1 As shown, specifically including: S1, data collection and cleaning processing: collect historical data from multiple data sources, perform data cleaning and data transformation; In order to ensure accurate dynamic planning and automatic processing of passenger boarding applications, a large amount of historical data analysis and calculation is required. The data comes from a wide range of sources, including but not limited to flight number, aircraft model, aircraft number, route, area, mission, VIP, aircraft position, boarding gate, guarantee completion time, planned landing time, actual landing time, planned take-off time, actual take-off time, boarding start time, boarding deadline, door opening and closing time, boarding time, COBT, CTOT, weather information, route restrictions, flow control, delay information, etc. These data come from multiple data sources such as the air traffic control system, the national flow system, and the airport command and dispatch system. After the initial collection, the crucial cleaning and processing stage follows, which includes but is not limited to data cleaning (such as removing redundant information, filling missing values, format conversion), consistency checking (ensuring no duplicates, uniqueness constraints), outlier detection and correction (outlier detection, outlier correction) and data transformation. Data cleaning and processing includes the following tasks: Data cleaning: The process of improving data quality by identifying and correcting erroneous, incomplete or incorrectly formatted records. This process usually includes removing redundant information, filling missing values, and unifying formats. Using the MurmurHash hash function, the threshold is set to 0.8, and each record is mapped to a unique hash value through the hash function. The hash values ​​of two records are compared. If the hash values ​​are greater than the threshold of 0.8, the two records are considered similar or duplicated, and the duplicate records are marked or deleted; for missing values, the appropriate filling strategy is selected according to the data type, such as the mean and median to fill numerical data, and the mode to fill categorical data; at the same time, the date and time fields are converted to a unified standard date and time format (such as yyyy-MM-dd HH:mm format) to ensure data format consistency. Consistency check: Determine the fields that need to be unique, and set unique constraints for related fields in the database table or data set. Outlier detection: Use the box plot method to identify data points that deviate from the norm, and the formula is IQR=Q_3-Q_1. Quartiles and IQR are calculated with the help of Apache Commons Math library, where (Q_1) and (Q_3) are the first quartile and the third quartile respectively. Any value less than (Q_1-1.5*IQR) or greater than (Q_3+1.5*IQR) is considered an outlier. The outlier can be replaced with the mean, median, or boundary value according to the business situation, or the outlier record can be deleted directly. Data transformation: Create new features by performing mathematical operations on existing attributes, or adjust the numerical range through standardization and normalization to improve the performance of subsequent predictive models. For example, multiply the flight time by the number of passengers to obtain the interaction term, and create this new feature to represent the relationship between the flight time and the number of passengers. And through the normalization method , normalize the route restriction information in order to balance the weight of route restrictions in the calculation process, where, is the normalized data, This step aims to improve the overall data quality, eliminate potential noise interference and irregularities, and lay a solid foundation for subsequent variable screening and model training.

[0031] S2, feature selection and model training: select indicators and attributes that reflect the intrinsic characteristics of historical data from historical data, use deep learning algorithms to build prediction models, and use the obtained indicators and attributes as input information to train the prediction model; The indicators or attributes that reflect the intrinsic characteristics of historical data are used as input information to train the prediction model.

[0032] For example, in the dynamic planning process of flight boarding application, the indicators that may be involved include flight number, aircraft model, route, area, VIP, boarding gate, guarantee completion time, planned landing time, actual landing time, planned take-off time, actual take-off time, boarding start time, boarding deadline, cabin door opening and closing time, boarding time, COBT, CTOT, weather information, route restrictions, flow control, delay information and other indicators.

[0033] After completing the selection and extraction of features, the prediction model is further constructed. This system uses deep learning technology as the core algorithm, especially the Transformer model and reinforcement learning (RL). The Transformer model performs well when processing long sequences, can be calculated in parallel, and it uses the self-attention mechanism to better capture global dependencies. Reinforcement learning (RL) is not just about making predictions based on existing data, but about continuous optimization based on these prediction results. Combined with technologies such as deep Q network (DQN) and policy gradient, effective scheduling can be achieved in a dynamic environment. As a result, it has the ability to process dynamic planning for future flight boarding applications and can conduct in-depth analysis to support a more scientific and reasonable decision-making process.

[0034] S3, data calculation and release: The prediction model is trained through historical data, and the prediction model is applied to the production scenario of flight boarding application. Real-time dynamic flight information and air traffic control (ATC) information are collected and input into the prediction model. The prediction model combines historical data analysis, processes boarding applications, predicts the optimal boarding time and boarding order, and ensures the flight regularity rate; Exemplarily, after receiving data from the airport operation management system, air traffic control system, national traffic system, and airline system, the system predicts and processes flight boarding applications based on the prediction model to ensure the accuracy and real-time nature of the opening and sequence of boarding applications.

[0035] The calculation results may include the boarding application result information, boarding time, boarding location, boarding sequence, etc. The result set can be output to the airport management system, air traffic control system, airline system and other related platforms for user query and reference.

[0036] S4, output verification and optimization: Use statistical methods to automatically screen out results that do not conform to the normal pattern, mark them, continuously update the data resource pool, and eliminate interference items; For example, after the prediction model is first launched, a period of time is set as a trial operation period, during which the traditional manual review mechanism is retained to compare the differences between the two. And statistical methods (such as Principle), automatically screen out result items that do not conform to the normal pattern, mark them for subsequent review and confirmation, continuously update the data resource pool, continuously improve the internal knowledge graph content, and eliminate interference items.

[0037] For example, Figure 2 The invention discloses a method and principle for dynamically planning and allocating multi-service data for boarding applications at civil airports.

[0038] Dataset: The whole process starts with a dataset containing all relevant data, which can be obtained through database extraction, API interface calls, web crawling, etc. In this example, the dataset contains various features of historical flights, such as flight number, departure and arrival airports, departure and arrival times, COBT (estimated departure time), weather conditions, traffic information, etc. Feature Engineering: The process of converting raw data into features that can be used for modeling. In this stage, the raw data is cleaned, integrated, processed, and converted into a format suitable for model input. In particular, the features of historical flights are converted into embedding vectors, which is an input format required by the Transformer model. Data Partitioning: The dataset is divided into three subsets: training set, validation set, and test set. The training set is used to train the model, the validation set is used to evaluate the performance of the model during the training process and prevent overfitting, and the test set is used to evaluate the final performance of the model after the training is completed. Model Training: The Transformer model is selected as the underlying architecture. It consists of 6 layers of encoders, each with 8 heads, forming a deep Transformer model network. The Transformer model is a deep learning model that is particularly good at processing sequence data. Through its unique self-attention mechanism, it can capture the dependencies in the input historical flight data without considering the sequence order, which is very effective for processing complex features. During the training process of this model, the Transformer model captures the dependencies in the historical flight data through the self-attention mechanism, converts the features of historical flights (such as flight number, take-off and landing airports, take-off and landing times, COBT, weather conditions, traffic information, etc.) into embedding vectors, and adds position encoding to each embedding vector to retain the position information in the sequence. Then, through multi-layer encoder processing, each layer contains a multi-head self-attention mechanism and a feedforward neural network, and finally generates a prediction output for the flight boarding application time and sequence. Next, during the model training process, the generated prediction output is compared with the true label in the actual historical flight data, and a loss function such as mean square error (MSE) or mean absolute error (MAE) is used to quantify the difference between the prediction output and the true label. The loss value calculated by the loss function is used to evaluate the performance of the model in the current iteration. The smaller the loss value, the closer the model's prediction is to the true value. Through continuous iterative training, it is ensured that the model can be continuously optimized to improve its accuracy and reliability in actual predictions. Model evaluation: During the training process, the model is periodically evaluated on the validation set, which helps monitor the learning progress of the model and adjust the parameters or structure of the model based on the performance on the validation set. Model prediction: When the model is trained and performs well on the validation set, the test set can be used for the final evaluation. This step is to test whether the model is ready to deal with new data that has not been seen.Prediction results: Based on the performance of the model on the test set, the final prediction results can be obtained, including the boarding application result information, boarding time, boarding location, boarding sequence, etc. The result set can be used to guide the dynamic planning and allocation of boarding applications at civil airports, thereby achieving more efficient resource management and services and improving airport operation efficiency. Example 2, such as. Figure 3 As shown, the civil airport boarding application dynamic planning allocation system of multi-service data provided by the embodiment of the present invention includes: Data collection and cleaning module 1 is mainly responsible for data collection and preliminary cleaning of data to ensure the accuracy of subsequent prediction model training. The main steps are: (1) Data collection: Collect historical data (raw data) from different data sources, which may include database extraction, API interface calls, web crawling, etc.

[0039] (2) Data cleaning: The process of improving data quality by identifying and correcting erroneous, incomplete or incorrectly formatted records. This process usually includes removing redundant information, filling missing values, and unifying formats. Using the MurmurHash hash function, the threshold is set to 0.8, and each record is mapped to a unique hash value through the hash function. The hash values ​​of two records are compared. If the hash values ​​are greater than the threshold of 0.8, the two records are considered similar or duplicate, and the duplicate records are marked or deleted. For missing values, an appropriate filling strategy is selected according to the data type, such as filling numerical data with the mean or median, and filling categorical data with the mode. At the same time, the date and time fields are converted to a unified standard date and time format (such as yyyy-MM-dd HH:mm format) to ensure data format consistency.

[0040] (3) Consistency check: Determine the fields that need to remain unique and set uniqueness constraints for related fields in the database table or data set.

[0041] (4) Outlier detection: The box plot method is used to identify data points that deviate from the normal state. The formula is IQR=Q_3-Q_1. The Apache Commons Math library is used to calculate the quartiles and IQR, where (Q_1) and (Q_3) are the first quartile and the third quartile, respectively. Any value less than (Q_1-1.5*IQR) or greater than (Q_3+1.5*IQR) is considered an outlier. The outlier can be replaced with the mean, median or boundary value according to the business situation, or the outlier record can be deleted directly.

[0042] (5) Data transformation: Create new features by performing mathematical operations on existing attributes, or adjust the value range through standardization and other methods to improve the performance of subsequent prediction models. For example, multiply the flight time by the number of passengers to obtain an interaction term, and create this new feature to represent the relationship between the flight time and the number of passengers. And through the normalization method , normalize the route restriction information in order to balance the weight of route restrictions in the calculation process, where, is the normalized data, is the original data.

[0043] Feature selection and feature extraction module 2: It is mainly responsible for selecting the most influential subsets of historical data for the prediction target to reduce the dimension, improve the performance of the prediction model and enhance the interpretability. And by transforming or mapping the historical data, it generates a new feature representation that can better reflect the internal rules, thereby improving the learning ability of the algorithm and achieving the dimensionality reduction effect, thereby improving the accuracy and efficiency of the overall prediction model. The main steps are: 1) Feature selection: Based on knowledge in the civil aviation field and exploratory data analysis, data fields that are obviously irrelevant or noisy are excluded. The Pearson Correlation Coefficient can also be calculated between variables and between them and the target variable to find the existence of strong correlations. Combined with the random forest method, multiple decision trees are constructed, with the number of trees set to 200, the maximum depth set to 10, and the threshold set to 30%. The configuration parameters are modified iteratively to avoid overfitting and underfitting problems. The features are sorted according to the feature importance score, and the top-ranked features are selected. The least important features are removed until the ideal scale is reached.

[0044] For example, the Pearson coefficient is a method for measuring the degree of linear correlation between two variables. Based on civil aviation knowledge and exploratory data analysis, it excludes obviously irrelevant or noisy data fields, calculates the Pearson coefficients between variables and between each variable and the target variable, and finds out the existence of strong correlations; it repeatedly removes unimportant features in combination with machine learning algorithms until the ideal scale is reached; the calculation formula of the Pearson coefficient is: In the formula, is the Pearson coefficient, are the values ​​of the two variables in the sample, is the mean value of the two variables in the sample, is the sample size; The Pearson coefficient calculation formula is used to calculate the linear correlation between two variables, and its value is between -1 and 1; When it is close to 1 or -1, it indicates that there is a strong linear relationship between the two variables; when When it is close to 0, it means that there is almost no linear relationship between the two variables; positive values ​​indicate positive correlation, and negative values ​​indicate negative correlation.

[0045] 2) Feature extraction: After feature selection, the feature values ​​that are closely related to the prediction model are extracted for use in prediction model training. In terms of dynamic planning processing of flight boarding applications, relevant features may include but are not limited to time information: flight departure time, flight arrival time, boarding time, COBT, CTOT, door opening and closing time, and guarantee completion time. Flight information: airline, flight number, region, mission, VIP, aircraft number, aircraft model, destination, flight duration, and passenger capacity. Resource information: terminal, aircraft position, boarding gate, baggage carousel, security check channel, runway. Route information: route restrictions, flow control information, and weather conditions.

[0046] Data modeling and model training module 3: Mainly responsible for training the prediction model for the extracted feature values, combining the flight boarding application business scenario, and using the deep learning algorithm to build a prediction model. The prediction model mainly includes the Transformer module and the RL module. The two are combined to form a hybrid model that combines the sequence modeling capabilities of the Transformer model and the decision-making optimization capabilities of the RL module. Different modules collaborate with each other to complete the prediction task, so that the model can better handle complex prediction problems and have higher performance and generalization capabilities. It aims to plan for future flight boarding applications through analysis and training of historical data. The main steps of this module are: (a) Dataset partitioning: The dataset is divided into training set, validation set, and test set in a ratio of 8:1:1 to ensure the generalization ability of the prediction model.

[0047] (b) Build a Transformer model (such as Figure 4 ): The present invention selects the Transformer model as the underlying architecture of the deep learning algorithm, and converts the characteristics of historical flights such as flight number, take-off and landing airports, take-off and landing times, COBT, weather conditions, traffic information, etc. into embedded vectors for use by the Transformer encoder. At the same time, a 6-layer encoder is set with 8 heads in each layer to form a deep Transformer model network, which captures the dependencies in the historical flight data through the self-attention mechanism to capture complex features.

[0048] Figure 4In the figure, multiple tasks (T1, T2...TN) and corresponding inputs (E1, E2...EN) are shown. Each business area (such as flight scheduling, passenger service, ground operation, etc.) is regarded as a separate task (T1, T2...TN), and each task has its own unique input (E1, E2...EN) (such as flight schedule, passenger information, weather information, airport resource usage, flow control information, etc.). These inputs are first fed into the shared representation layer (Trm), which plays a core role. It converts the inputs of each task into a common form and extracts common features or high-level concepts across domains. These features may include flight frequency patterns, passenger behavior preferences, weather impact, flow control impact, etc. After the shared representation layer, each task has its own dedicated representation layer (Trm), which further processes the output from the shared layer to make it more suitable for its specific needs. For example, in the flight schedule management task, the model may need to pay attention to the connection time and idle time period between flights; while in the flow control processing task, the model may be more concerned with the heading and distance of the outbound flight. In this way, the Transformer-based multi-task learning framework can not only effectively integrate and utilize data from different business areas, but also promote knowledge sharing and complementarity between different tasks, thereby improving the overall prediction accuracy and decision quality. In addition, this framework also has good scalability. As new business needs emerge, new tasks can be easily added without major modifications to the entire system.

[0049] (c) Reinforcement learning strategy optimization: The prediction model uses Transformer as the basic architecture, and introduces reinforcement learning elements for strategy optimization. The reinforcement learning (RL) module is designed to optimize the prediction of flight boarding application time and order. First, the state space is defined, including the number of passengers on the current flight, the waiting time of passengers, and flight delays. The action space consists of the decision to allow a flight to start boarding, that is, whether to choose whether to allow the flight to start boarding at a specific time. The reward function can reward or punish by evaluating indicators such as passenger waiting time and flight release rate, so that the prediction model can make better boarding decisions and update the strategy through the deep Q network (DQN). Through multiple rounds of iterations and continuous optimization to improve the overall performance, the RL module will eventually work together with the Transformer part to enable the system to flexibly respond to different flight scenarios and provide accurate predictions of boarding time and order.

[0050] In step (c), the reinforcement learning strategy is optimized, including: the reinforcement learning RL module is used to optimize the prediction of the time and sequence of flight boarding applications. First, the state space is defined, including the number of passengers on the current flight, the waiting time of passengers, and the delay of the flight; the action space is composed of the decision of allowing a certain flight to start boarding, and whether to let the flight start boarding at a specific time; the reward function rewards or punishes by evaluating the waiting time of passengers and the normal flight release rate index, such as successfully processing a boarding application +2 points, passenger waiting time less than 1 hour +1 point, flight release and takeoff are normal +3 points; failed to process a boarding application -2 points, not processing the boarding application beyond the scheduled time -1 point, violating business rules -2 points, flight takeoff delay -3 points; adjust the reward function through user feedback, compare the effects of different reward functions, and select the best solution to encourage the model to make better boarding decisions and improve the operational efficiency of the whole day; A deep Q network (DQN) is designed as an intelligent agent to learn the best boarding decision strategy. By optimizing the boarding process, the cumulative reward is maximized, such as reducing passenger waiting time and improving the normal flight release rate. Combined with the experience replay mechanism, we collect experience tuples related to flight boarding and store them in the experience replay buffer. These experiences will be used to train the deep Q network to learn the best boarding decision strategy. At the same time, the target network is introduced to stabilize the training process and reduce the volatility during training. By using the mean square error (MSE) or Huber loss function, the gap between the predicted Q value and the target Q value is minimized. The model is continuously optimized and the training data is updated to continuously improve the model performance, improve flight boarding efficiency and flight takeoff release rate. Continuous optimization through multiple rounds of iterations.

[0051] Finally, the RL module works in collaboration with the Transformer module. The Transformer module processes the flight boarding application data, models the boarding sequences of different flights through feature extraction, standardization and time series data processing, captures the long-distance dependencies between the input sequences, and performs big data prediction. At the same time, the RL module learns the best boarding decision strategy based on the deep Q network (DQN). Through interaction with the environment, it continuously explores and optimizes the boarding behavior to maximize the cumulative reward, such as reducing the waiting time for boarding and improving the normal rate of flight takeoff release. It combines experience replay, target network and loss function and other technologies to stabilize the training process and gradually optimize the decision strategy. In the collaborative work, the Transformer module can provide predicted boarding needs and trends, and provide the basis and background information for the RL module to make decisions. The RL module can dynamically adjust the boarding strategy according to the prediction results output by the Transformer module to realize the intelligent decision-making process. The collaborative work between the two can realize the organic combination of data prediction and decision optimization, improve the performance and efficiency of the overall system, make the system more intelligent and efficient, and achieve better service quality and passenger experience. This method of comprehensive utilization of different technical means can effectively cope with complex passenger boarding scenarios and improve the overall performance and competitiveness of the system.

[0052] (d) Output verification and optimization: After the prediction model training is completed, the output results of the model need to be verified to ensure the accuracy and reliability of the prediction. For example, the prediction accuracy, mean absolute error, mean square error, etc., to check the fit of the model on historical data. This helps to identify problems, such as whether the model is too complex (overfitting) or fails to capture effective trends (underfitting). Based on the evaluation results, the present invention may adjust hyperparameters or select different features to improve the prediction effect. Through continuous testing and verification, the present invention ensures that the prediction model can accurately predict the boarding situation of flights in actual applications, thereby helping airlines make better decisions and operational plans.

[0053] The mean absolute error :Mean absolute error is an important indicator to measure the performance of a prediction model, especially for regression problems. It evaluates the quality of the model by calculating the average of the absolute values ​​of the difference between the predicted value and the actual value in all observations. Its formula is as follows: In the formula, is the number of samples, To predict flight boarding time, is the actual boarding time of the flight. In the process of output result verification, according to the historical data model prediction, multiple sets of data are substituted into the formula to calculate the final average absolute error value; Mean square error The square average of the difference between the predicted value and the true value reflects the overall volatility of the model's prediction results. The formula is as follows: Accuracy: It indicates the proportion of correctly predicted samples to the total samples, which directly reflects the model's ability to classify samples. The formula is as follows, where TP indicates the number of samples correctly predicted as positive examples, FP indicates the number of samples incorrectly predicted as positive examples, TN indicates the number of samples correctly predicted as negative examples, and FN indicates the number of samples incorrectly predicted as negative examples. Substitute multiple sets of data into the formula to calculate the final accuracy.

[0054] Decision support or prediction analysis module 4: The trained passenger application prediction model is used in the actual production process. This module can predict possible delays or congestion in advance by comprehensively analyzing flight history records, passenger flow changes, weather conditions and other factors, and generate the optimal scheduling plan accordingly. At the same time, the system can also dynamically adjust the release priority according to real-time updated information to ensure that each aircraft can take off or land smoothly according to the best time window. With this advanced tool, airport managers can greatly improve operational efficiency, shorten passenger waiting time, and significantly improve the overall service quality and user experience.

[0055] It can be seen from the above embodiments that in the decision analysis of boarding application: In recent years, with the continuous increase in the number of flights and the number of travelers, the main problems faced by airports include inefficient boarding processes, extended waiting time for passengers, and unreasonable resource allocation. These challenges not only increase operating costs, but may also lead to the risk of decreased passenger satisfaction and increased safety hazards. By implementing a dynamic planning processing method for boarding applications, combined with historical data analysis and air traffic control-related release restriction instructions, it is possible to push information such as the best boarding time, boarding order, and security resources to airports, airlines and other systems, assisting airport scheduling operations, improving boarding efficiency, reasonably allocating resources, and improving the overall experience of passengers.

[0056] In terms of delay warning: we use historical data and real-time information such as control and flow control to conduct comprehensive analysis to determine whether the flight can meet the normal boarding and take-off requirements under the release restrictions, provide early warning of possible delays, and provide managers with scientific decision-making support to help them achieve more reasonable flight scheduling and resource allocation, thereby optimizing overall operational efficiency and service levels, improving passengers' travel experience and controlling potential economic losses and safety risks.

[0057] The present invention can effectively process boarding applications, reduce boarding time, optimize resource utilization, reduce labor costs and energy consumption, thereby improving airport operating efficiency and maximizing economic benefits. At the same time, it can improve passenger satisfaction, enhance customer loyalty, and enhance the airport's positive reputation and brand value. In addition, the technical solution of the present invention can not only be applied in the field of passenger boarding optimization at civil airports, but can also be extended to other fields, such as public transportation places such as railway stations and ferry terminals, and can even be applied to personnel management at large-scale event sites and other fields, expanding commercial application potential and market prospects.

[0058] The present invention solves the following problems: ① Traditional passenger boarding optimization algorithms often adopt static planning or heuristic methods, which are difficult to effectively respond to the real-time changing passenger boarding application requirements and conditions. The present invention introduces a reinforcement learning module as the core of the optimization algorithm. Through interactive learning with the environment, the system can continuously adjust the decision-making strategy according to the reward signal, filling the application gap of RL reinforcement learning in the field of passenger boarding optimization. ② Traditional passenger boarding optimization algorithms often require manual design of features and rules when processing data, which is difficult to adapt to the complex and changeable passenger boarding environment. The present invention adopts the Transformer module and the deep learning model based on the self-attention mechanism, which can better understand and process the input data, improve the system's data learning ability and inference accuracy, and fill the application gap of the deep learning model in the field of passenger boarding optimization. ③ Most traditional boarding optimization algorithms adopt static planning methods and lack the ability to dynamically adjust to real-time changes. The present invention combines RL reinforcement learning and Transformer modules to achieve intelligent decision-making and dynamic adjustment. The system can autonomously learn and optimize according to real-time conditions, filling the technical gap of combining dynamic planning with intelligent decision-making in the field of passenger boarding optimization.

[0059] The above description is only a preferred specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any modifications, equivalent substitutions and improvements made by any technician familiar with the technical field within the technical scope disclosed by the present invention and within the spirit and principles of the present invention should be covered within the protection scope of the present invention.

Claims

1. A method for dynamically planning and allocating multi-service data for civil airport boarding applications, characterized in that: The method includes: S1, data collection and cleaning processing: collect historical data from multiple data sources, perform data cleaning and data transformation; S2, feature selection and model training: select indicators and attributes that reflect the intrinsic characteristics of historical data from historical data, use deep learning algorithms to build prediction models, and use the obtained indicators and attributes as input information to train the prediction model; S3, data calculation and release: The prediction model is trained through historical data, and the prediction model is applied to the production scenario of flight boarding application. Real-time dynamic flight information and air traffic control (ATC) information are collected and input into the prediction model. The prediction model combines historical data analysis, processes boarding applications, predicts the optimal boarding time and boarding order, and ensures the flight regularity rate; S4, output verification and optimization: Use statistical methods to automatically screen out result items that do not conform to the normal pattern, mark them, continuously update the data resource pool, and eliminate interference items.

2. The method for dynamically planning and allocating multi-service data for civil airport boarding applications according to claim 1, characterized in that: Collect historical data from multiple data sources, perform data cleaning and data transformation, including: Collect historical data from different data sources, which are obtained through database extraction, API interface calls, and web crawling; Data cleaning, including: removing redundant information, consistency checking, outlier detection and correction; Data transformation, including: creating new features by performing mathematical operations on existing attributes; adjusting the value range through standardization and normalization.

3. The method for dynamically planning and allocating multi-service data for civil airport boarding applications according to claim 2, characterized in that: Historical data, including: flight number, aircraft type, aircraft number, route, area, mission, VIP, aircraft position, boarding gate, support completion time, planned landing time, actual landing time, planned take-off time, actual take-off time, boarding start time, boarding deadline, door opening and closing time, boarding time, COBT, CTOT, weather information, route restrictions, flow control, and delay information.

4. The method for dynamically planning and allocating multi-service data for civil airport boarding applications according to claim 1, characterized in that: In step S2, feature selection includes: based on civil aviation field knowledge and exploratory data analysis, excluding obviously irrelevant or noisy data fields, calculating the Pearson coefficient between variables and between each variable and the target variable, and finding the existence of strong correlations; combining machine learning algorithms to iteratively remove unimportant features until the ideal scale is reached; the calculation formula of the Pearson coefficient is: In the formula, is the Pearson coefficient, are the values ​​of the two variables in the sample, is the mean value of the two variables in the sample, is the sample size; The Pearson coefficient calculation formula is used to calculate the linear correlation between two variables, and its value is between -1 and 1; When it is close to 1 or -1, it indicates that there is a strong linear relationship between the two variables; when When it is close to 0, it means that there is almost no linear relationship between the two variables; positive values ​​indicate positive correlation, and negative values ​​indicate negative correlation.

5. The method for dynamically planning and allocating multi-service data for civil airport boarding applications according to claim 1, characterized in that: In step S2, a prediction model is constructed using a deep learning algorithm, and the obtained indicators and attributes are used as input information to perform prediction model training, including: (a) Dataset division: The dataset is divided into training set, validation set and test set in a ratio of 8:1:1; (b) Constructing the Transformer model: The Transformer model is selected as the underlying architecture of the deep learning algorithm. The features of historical flights (flight number, take-off and landing airports, take-off and landing times, COBT, weather conditions, and traffic information) are converted into embedding vectors for use by the Transformer encoder. A 6-layer encoder is set up with 8 heads per layer to form a deep Transformer model network. The self-attention mechanism is used to capture the dependencies in the historical flight data to capture complex features. (c) Reinforcement learning strategy optimization: Using Transformer as the basic architecture, reinforcement learning elements are introduced for strategy optimization; (d) Output verification and optimization: After the prediction model training is completed, the output results of the model are verified.

6. The method for dynamically planning and allocating multi-service data for civil airport boarding applications according to claim 5, characterized in that: In step (c), we use Transformer as the basic architecture and introduce reinforcement learning elements for policy optimization, including: Define the state space, including: the number of passengers on the current flight, the waiting time of passengers, and the flight delays; the action space consists of the decision to allow the flight to start boarding passengers, and the reward function rewards or punishes by evaluating the waiting time of passengers and the flight release regularity index; Design a deep Q network as an intelligent agent to learn the best boarding decision strategy and maximize the cumulative reward by optimizing the boarding process. Combined with the experience replay mechanism, collect the experience tuples related to flight boarding and store them in the experience replay buffer for training the deep Q network. The RL module works together with the Transformer module, which processes flight boarding application data. Through feature extraction, standardization, and time series data processing, it models the boarding sequences of different flights, captures long-distance dependencies between input sequences, and performs big data prediction.

7. The method for dynamically planning and allocating multi-service data for civil airport boarding applications according to claim 5, characterized in that: In step (d), the output results of the model are verified, including: checking the fit of the model on the historical data through accuracy, mean absolute error, and mean square error.

8. The method for dynamically planning and allocating multi-service data for civil airport boarding applications according to claim 1, characterized in that: In step S3, the calculated result includes the pick-up application result information, pick-up time, pick-up location, and pick-up sequence; the result set is output to the airport management system, air traffic control system, and airline system for user query and reference.

9. The method for dynamically planning and allocating multi-service data for civil airport boarding applications according to claim 1, characterized in that: In step S4, after the prediction model is launched, a trial period is set up, during which a manual review mechanism is retained to compare the differences between the two. Statistical methods are used to automatically screen out result items that do not conform to the normal pattern and mark them for review and confirmation. The data resource pool is continuously updated, the internal knowledge graph content is continuously improved, and interference items are removed.

10. A multi-service data civil airport boarding application dynamic planning allocation system, characterized in that: The system implements the method for dynamically planning and allocating civil airport boarding applications of multi-service data as claimed in any one of claims 1 to 9, and the system comprises: A data collection and cleaning processing module (1) is used for data collection and preliminary cleaning of the data; Feature selection and feature extraction module (2), used to select the most influential subset of historical data for the prediction target, and extract new feature representations that better reflect the internal laws by transforming or mapping the historical data; The data modeling and model training module (3) is responsible for training the prediction model based on the extracted feature values. It uses a deep learning algorithm to build a prediction model based on the flight boarding application business scenario. It makes plans for future flight boarding applications through analysis and training of historical data. The decision support or prediction analysis module (4) applies the trained prediction model to the actual production process. By comprehensively analyzing flight history records, passenger flow changes, and weather conditions, it predicts delays or congestion in advance and generates the optimal scheduling plan based on this.

Images