Flight operation data quality evaluation analysis method and system

Abstract

The invention provides a flight operation data quality evaluation analysis method and system, and belongs to the technical field of civil aviation management and data quality evaluation. The method comprises the steps that original flight operation data are sent to data middleware, and the data middleware receives the original flight operation data based on a data pipeline and stores the original flight operation data in an annular data queue; when the annular data queue is full, obtaining all flight operation data stored in the annular data queue, inputting the flight operation data into a pre-trained flight agent model, and outputting a first flight simulation scheduling result by the flight agent model; when the similarity between the first flight simulation scheduling result and the current actual flight scheduling result meets a preset condition, it is judged that the original flight operation data passes data quality evaluation; otherwise, checking the original flight operation data. The pre-trained flight agent model is an action path prediction model based on a Boolean network model. According to the invention, the acquisition and warehousing quality level of flight operation data can be ensured.

Description

TECHNICAL FIELD

[0001] The present application belongs to the technical field of civil aviation management and data quality evaluation, and particularly relates to a flight operation data quality evaluation and analysis method and system, a computer readable storage medium for implementing the method, a computer program product, and an electronic device. BACKGROUND

[0002] In the civil aviation management system, flight plans and flight dynamic data are collectively referred to as flight operation data; flight operation data includes flight plans and flight dynamics, as well as the current position and state of the flight in the air traffic primary and secondary radar track data.

[0003] In the prior art, there are methods for using big data technology to analyze historical flight operation conditions to find rules and develop flight schedule adjustment strategies to improve the accuracy of schedule adjustment, such as the airport group flight schedule optimization method based on deep reinforcement learning proposed in Chinese authorized patent CN118798840B, which uses DQN to optimize the flight schedule and uses a deep neural network to replace the traditional reinforcement learning table, combining the decision-making ability of reinforcement learning and the data analysis ability of deep learning to make the task of processing large-scale data space more efficient; the flight delay recovery method based on a multi-agent system proposed in Chinese authorized patent CN113947296B uses the communication negotiation mechanism between agents to simulate the communication and collaboration process between various functional departments to complete the integrated solution of the flight delay recovery problem, solving the technical problems in the communication and negotiation process when the airline actually performs flight delay recovery.

[0004] Whether it is to use historical data for prediction or to perform simulation based on real-time data, the quality of the data itself must be guaranteed to a certain extent, such as the accuracy of data collection, the representativeness of the data itself, and the completeness of the data, which must meet certain standards. For flight operation data, quality evaluation is even more important, and only data that has passed quality evaluation is suitable for being collected into the database for subsequent inspection or continued use for updating and training of the flight agent model.

[0005] Therefore, there is a need for an improved technical solution that can ensure the quality level of flight operation data collection and database storage. SUMMARY

[0006] To solve the above technical problems, the present application proposes a flight operation data quality evaluation and analysis method and system, a computer readable storage medium for implementing the method, a computer program product, and an electronic device.

[0007] In the first aspect of the present application, a flight operation data quality evaluation and analysis method is proposed, which is implemented based on a data middleware;

[0008] The method comprises the following steps:

[0009] The original flight operation data is sent to the data middleware, and the data middleware receives the original flight operation data based on a data pipeline and stores it to a ring data queue;

[0010] When the ring data queue is full, all flight operation data stored in the ring data queue is obtained and input to a pre-trained flight intelligent agent model, and the flight intelligent agent model outputs a first flight simulation scheduling result;

[0011] When the similarity of the first flight simulation scheduling result and the current actual flight scheduling result meets a preset condition, it is determined that the original flight operation data passes the data quality evaluation.

[0012] When the similarity of the first flight simulation scheduling result and the current actual flight scheduling result does not meet the preset condition, the original flight operation data is checked.

[0013] The data middleware receives the original flight operation data based on a data pipeline, and also receives the current actual flight scheduling result corresponding to the original flight operation data and stores it to the ring data queue.

[0014] The original flight operation data includes the destination of the current flight, the route of the current flight, the full load rate of the current flight, the unit passenger kilometer income of the current flight, the unit seat kilometer cost, all arrival flight data of the current airport, all departure flight data, etc., covering flight plans and flight dynamics, as well as the current position and state of the flight in the air traffic control primary and secondary radar track data.

[0015] The pre-trained flight intelligent agent model is an action path prediction model based on a Boolean network model, which generates an optimal flight scheduling sequence based on the original flight operation data.

[0016] The Boolean network model is constructed in the following manner:

[0017] Collect original flight operation data and define the logical variables of each flight at the current airport;

[0018] Use Boolean algebra to represent the logical variables of all flights and their relationships, and establish a flight operation Boolean network model.

[0019] The flight intelligent agent model outputs a first flight simulation scheduling result, which specifically includes:

[0020] All flight operation data stored in the ring data queue is input to a pre-trained action path prediction model based on a Boolean network model;

[0021] solving the optimal control sequence of the Boolean network model based on at least one optimal solution solving algorithm, obtaining an optimal action path, the optimal action path representing an optimal action order of all flights of the current airport under the current original flight operation data.

[0022] In the second aspect of the present application, in order to implement the method of the first aspect, a flight operation data quality evaluation analysis system is further provided, and the system comprises:

[0023] Data middleware: receiving original flight operation data and current actual flight scheduling results corresponding to the original flight operation data based on a data pipeline and storing them to a ring data queue;

[0024] Flight intelligent agent model: when the ring data queue is full, obtaining all flight operation data from the ring data queue and outputting a first flight simulation scheduling result;

[0025] Quality evaluation unit: when the similarity between the first flight simulation scheduling result and the current actual flight scheduling result meets a preset condition, determining that the original flight operation data passes the data quality evaluation; otherwise, checking the original flight operation data.

[0026] The flight intelligent agent model is a pre-trained action path prediction model based on a Boolean network model, and the action path prediction model generates an optimal flight scheduling order based on the original flight operation data.

[0027] In the third aspect of the present application, a computer readable storage medium is further provided, which is used for storing computer instructions, and when the computer instructions are executed on an electronic device, the electronic device is caused to execute all or part of the steps of the aforementioned flight operation data quality evaluation analysis method.

[0028] In the fourth aspect of the present application, a computer device is further provided, which comprises a processor and a memory, the memory is used for storing instructions, and the processor is used for calling the instructions in the memory, so that the computer device executes the aforementioned flight operation data quality evaluation analysis method.

[0029] In the fifth aspect of the present application, a computer program product is further provided, which comprises a computer program, and when the computer program is executed, all or part of the steps of the aforementioned flight operation data quality evaluation analysis method are implemented.

[0030] The application sends original flight operation data to data middleware, the data middleware receives the original flight operation data based on a data pipeline and stores to a ring data queue; when the ring data queue is full, all flight operation data stored in the ring data queue is obtained and input to a pre-trained flight intelligent agent model, the flight intelligent agent model outputs a first flight simulation scheduling result; when the similarity of the first flight simulation scheduling result and the current actual flight scheduling result meets a preset condition, it is determined that the original flight operation data passes the data quality evaluation; otherwise, the original flight operation data is checked. The pre-trained flight intelligent agent model is an action path prediction model based on a Boolean network model. The application can ensure the collection and storage quality level of flight operation data.

[0031] Further advantages of the application will be further illustrated in detail in the specific embodiment part combined with the drawings of the specification. BRIEF DESCRIPTION OF DRAWINGS

[0032] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the drawings needed to be used in the embodiments will be briefly introduced below. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative effort on the basis of these drawings.

[0033] Figure 1 is the main flow chart of the flight operation data quality evaluation and analysis method of an embodiment of the present application;

[0034] Figure 2 is Figure 1 is the part flow chart of the method data collection, storage and processing;

[0035] Figure 3 is the flow chart of constructing the flight intelligent agent model based on the Boolean network model;

[0036] Figure 4 is the functional unit combination diagram of the flight operation data quality evaluation and analysis system for implementing the flight operation data quality evaluation and analysis method of the present application. DETAILED DESCRIPTION

[0037] In the specific embodiments of the present application, if the embodiments of the related technical solutions involve user-related data, when the embodiments of the present application are applied to specific products or technologies, the user's permission or consent is required, and the collection, use and processing of related data need to comply with relevant laws, regulations and standards of relevant countries and regions.

[0038] Referring to Figure 1 , Figure 1is the main flow chart of a flight operation data quality evaluation analysis method of an embodiment of the present application.

[0039] Figure 1 The method is shown to include three main flow blocks. For convenience of description, it is numbered as steps S100-S300 and described as follows (the accompanying drawings are not shown) Figure 1 The step numbers are omitted in the following description:

[0040] S100: send the original flight operation data to the data middleware, the data middleware receives the original flight operation data based on the data pipeline and stores it to the ring data queue;

[0041] S200: when the ring data queue is full, obtain all the flight operation data stored in the ring data queue, input it to the pre-trained flight intelligent agent model, and the flight intelligent agent model outputs the first flight simulation scheduling result;

[0042] S300: when the similarity of the first flight simulation scheduling result and the current actual flight scheduling result meets the preset condition, it is determined that the original flight operation data passes the data quality evaluation and can be stored as selected flight training data or record data.

[0043] Next, on the basis of Figure 1 further combined with Figure 2 and Figure 3 , the specific implementation of each step is described in detail.

[0044] Firstly, the method is realized based on the data middleware. The data middleware can be in the form of a transmission storage queue, a storage stack, etc. containing a data pipeline.

[0045] By adopting the form of data middleware, the production end and the consumption end of the data can be decoupled, and the production end is only responsible for publishing (producing, sending) data, which is received and stored by the data middleware. The consumption end obtains data from the data middleware based on a certain mechanism for consumption (such as subsequent processing, data evaluation, etc.). This mode of “data publishing end-data middleware-data consumption end” has been widely used in the field.

[0046] However, in the prior art, the "publish-subscribe" mode is mostly based on complete "decoupling". That is, the publishing end is "only responsible" for publishing (producing, sending) data storage to a certain queue, and the consuming end is "only responsible" for consuming data, and there is no data communication (communication) process between the publishing end and the consuming end. The consuming end pre-registers the message data type it is interested in, and then scans the storage queue at a fixed frequency to see if the message data type it is interested in has arrived. Under this mechanism, the storage queue usually needs a large space to ensure that the message data of the publishing end is stored enough for the consuming end to get it without being overwritten by new data before the old data is consumed.

[0047] To this end, in order to ensure the decoupling of data production and consumption, thereby expanding the applicability of the method, while avoiding the problem of needing a large storage queue space, in the embodiment, step S100 sends the original flight operation data to the data middleware, and the data middleware receives the original flight operation data based on the data pipeline and stores it to a ring-shaped data queue. Figure 1

[0048] Preferably, the data middleware can be a RocketMQ-based cloud storage data queue; preferably, the cloud storage data queue is a ring-shaped data queue with a preset storage space size.

[0049] The reason for using a RocketMQ-based cloud storage data queue is that RocketMQ is a low-latency, high-concurrency, high-availability, and high-reliability distributed message. Due to its simple architecture, rich business functions, and strong scalability, it is widely used by many enterprise developers and cloud vendors, especially in specific flight operation data processing scenarios, which need to meet the standards of low latency, high concurrency, high availability, and high reliability, and can be expanded.

[0050] The reason for using a ring-shaped data queue instead of a linked queue is that the linked queue is a first-in, first-out queue, which may cause old data to be overwritten by new data before it is consumed if the storage space is limited. If the overwrite is prohibited, it may cause data overflow. The use of a ring-shaped queue does not have the problem of data overflow.

[0051] As for the data overwrite problem that may occur when the storage space is limited, further refer to step S200: when the ring-shaped data queue is full, all flight operation data stored in the ring-shaped data queue is obtained and input into a pre-trained flight intelligent agent model, and the flight intelligent agent model outputs a first flight simulation scheduling result.

[0052] In this step, the method improves the traditional publish-subscribe communication-free message mechanism to a "queue full to remind" mechanism. ​

[0053] At this time, the pre-trained flight agent model can be regarded as a "consumer"; when the ring-shaped data queue is full, a "full" signal is generated, reminding the pre-trained flight agent model to perform a "consumption" operation; at this time, the "ring-shaped data queue stored all flight operation data" can be actively pushed to the "consumer"; or the "consumer" pulls data from the ring-shaped data queue when receiving the "full" reminder signal.

[0054] Preferably, a one-way data transmission channel is used between the "consumer" and the "ring-shaped data queue".

[0055] Through the joint improvement of the above aspects, the storage cost of data (private cloud storage space is usually charged according to the size of the storage space) can be controlled, and the timeliness and effectiveness of data can be avoided (avoiding data being overwritten before being obtained).

[0056] After step S200, the ring-shaped data queue is emptied to perform the next data collection (publishing)-consumption phase.

[0057] Referring to Figure 2 , Figure 2 is Figure 1 The method data collection, storage and processing part flow diagram.

[0058] Specifically, the data middleware receives the original flight operation data based on the data pipeline, and also receives the current actual flight scheduling result corresponding to the original flight operation data and stores it to the ring-shaped data queue.

[0059] The original flight operation data includes: the destination of the current flight, the flight path of the current flight, the full load rate of the current flight, the unit passenger kilometer income of the current flight, the unit seat kilometer cost, all arrival flight data of the current airport, all departure flight data, etc. It includes flight planning and flight dynamics, as well as the current position and state of the flight in the primary and secondary radar track data.

[0060] As an example, when the method of the application is applied to a certain airport, for a certain target flight (airline, airport), a plurality of flight agent models capable of optimizing scheduling need to be pre-trained, such as flight time optimization model, flight sorting model, flight delay recovery model, etc. These models can be used to assist the airport management and control party to specify the flight scheduling scheme.

[0061] Specific to a certain flight, at the current time, the airport management has derived the scheduling scheme (scheduling order) that needs to be executed and sent to the on-duty aircraft crew of the flight, and the on-duty aircraft crew needs to strictly execute the received scheduling scheme (scheduling order) that needs to be executed, that is, the current actual flight scheduling result.

[0062] In another aspect, in synchronization with this, the step S200 can also be performed, based on all the flight operation data stored in the ring data queue, input into a pre-trained flight intelligent agent model, and the flight intelligent agent model outputs a first flight simulation scheduling result. This mechanism is similar to digital twin driven intelligent decision support.

[0063] Both practice and theory prove that only when the data quality meets certain evaluation standards, the prediction model trained based on the data, the prediction output or simulation output based on the data will basically meet the actual situation. On the contrary, if the data quality is low, the trained model (intelligent agent) or the model (intelligent agent) output result will be greatly different from the actual situation, and the batch of data itself may have certain problems, which should be re-verified before being stored.

[0064] Based on this, the method performs step S300: when the similarity between the first flight simulation scheduling result and the current actual flight scheduling result meets the preset condition, it is determined that the original flight operation data passes the data quality evaluation; when the similarity between the first flight simulation scheduling result and the current actual flight scheduling result does not meet the preset condition, the original flight operation data is checked.

[0065] The data quality itself affects the model training and the model output result, but whether the model itself matches the actual scene is also a problem that needs to be concerned.

[0066] Based on this, as a further improvement, in the technical solution of the present application, the pre-trained flight intelligent agent model is an action path prediction model based on a Boolean network model, and the action path prediction model generates an optimal flight scheduling order based on the original flight operation data.

[0067] Boolean networks are systems of binary-valued variables interacting through directed connections, which are useful for modeling complex systems with threshold behavior and multiple feedbacks. Boolean network model was first used to study the dynamic behavior of gene regulatory networks. It is assumed that the state of a gene can be divided into two kinds: "0" and "1"; wherein 0 represents that the gene is not expressed; 1 represents that the gene is expressed. Therefore, each node in the Boolean network can represent the state of a gene. The interaction rule between nodes is represented by a Boolean function. The infinite range of optimal control problem refers to selecting an optimal control input strategy to make the network travel along each feasible path from the initial state, and minimize the expected total control cost when the target state is met.

[0068] It should be pointed out that the action state mechanism of the Boolean network model belongs to the prior art in mathematics itself, and the present application does not expand on this in detail, and those skilled in the art can refer to the following related literature:

[0069] "Implicit Boolean Network for Planning Actions", G. A. Oparin, et al; MIPRO 2022.

[0070] The unique improvement of the present application is that it first combines the action optimal path state mechanism of the Boolean network, applies it to the flight data scheduling scene, and gives a specific combination application means. In particular, based on the above mechanism of the Boolean network model, combined with the characteristics of airport flight scheduling, the present application first applies the Boolean network model to the construction and training solution process of the intelligent model of the flight model.

[0071] Specifically, the state of a flight can also be expressed as two states: "0" and "1"; wherein 0 represents that the flight is not expressed (for example: not entered into the scheduling sequence); 1 represents that the flight is expressed (for example, entered into the scheduling sequence). Each flight is a node in the Boolean network, and the connection relationship between nodes is represented by a Boolean function.

[0072] Specifically, the interaction rule is represented by a Boolean function. The Boolean network can be represented as B=(G,F); G=(V,E) is a directed graph;

[0073] Wherein,

[0074] V={x1,x2,……,xn} represents the set of nodes in the network; E represents the association relationship between nodes; each node xi represents the Boolean state value of node i (flight i);

[0075] F = {f1, f2, …, fn} is a set of Boolean functions representing the rules of regulating interaction between all interrelated flights; each Boolean function f i realizes a mapping of a state, determining the state of node i (flight i) at the next time;

[0076] If xi(t) represents the Boolean state value of node i (flight i) at time t, then

[0077] xi(t+1) = fi(xi 1(t), xi 2(t), …xin(t)) represents the Boolean state value of all flights at time t, thereby representing the logical variables of all flights and their relationships to build a flight operation Boolean network model, and converting it into an algebraic equation set.

[0078] By solving the algebraic equation set, the optimal path action state of each node can be obtained, that is, the current best state (whether to enter the scheduling state) and order (in which position to schedule) of each flight. That is, the pre-trained flight intelligent agent model is an action path prediction model based on a Boolean network model, and the action path prediction model generates an optimal flight scheduling order based on the original flight operation data.

[0079] Specifically, the Boolean network model is constructed in the following manner:

[0080] Collecting original flight operation data, defining the logical variables of each flight at the current airport;

[0081] Using Boolean algebra to represent the logical variables of all flights and their relationships, establishing a flight operation Boolean network model.

[0082] Specifically, referring to Figure 3 , Figure 3 is a flowchart of constructing a flight intelligent agent model based on a Boolean network model, specifically including:

[0083] Collecting original flight operation data;

[0084] Defining the logical variables of each flight at the current airport, for example, the state of a flight can also be expressed as two states: "0" and "1"; wherein 0 represents that the flight is not expressed (for example: not entering the scheduling sequence); 1 represents that the flight is expressed (for example: can enter the scheduling sequence). Each flight is regarded as a node in the Boolean network, and the connection relationship between nodes is represented by a Boolean function, and the connection order (precedence) of nodes represents the scheduling order (precedence) of flights.

[0085] All flights are represented by logical variables and their relationships using Boolean algebra, for example, V = {x1, x2, …, xn} represents a set of nodes in the network; E represents the association between nodes; each node xi represents the Boolean state value of node i (flight i); F = {f1, f2, …, fn} is a set of Boolean functions representing the rules for regulating the interaction between all related flights; each Boolean function fi realizes a state mapping, which determines the state of node i (flight i) at the next time;

[0086] A flight operation Boolean network model is constructed;

[0087] If xi(t) represents the Boolean state value of node i (flight i) at time t, then

[0088] xi(t+1) = fi(xi 1(t), xi 2(t), …, xin(t)) represents the Boolean state value of all flights at time t, thereby representing the logical variables and their relationships of all flights;

[0089] A flight operation Boolean network model is further constructed and converted into an algebraic equation system;

[0090] The vector Xt = {x1(t), x2(t), …, xn(t)} is used to represent the Boolean state value of all nodes (flights) at time t, and the state m of the network at time t is defined, so that the state X(t+1) = F(x1(t), x2(t), …, xn(t)) of the network at the next time is determined.

[0091] At this time, X(t+1) = Lu(t)x(t), and Lu(t) is the state transition matrix of the Boolean network, which can be determined by the initial state and the target state.

[0092] Preferably, the dynamic expression of the flight operation Boolean network model can be converted into an algebraic equation system based on the semi-tensor product technique of matrices.

[0093] After the flight operation Boolean network model is constructed and the dynamic expression of the flight operation Boolean network model is converted into an algebraic equation system, the optimal control series can be obtained by solving the algebraic equation system, and the optimal flight scheduling sequence can be generated.

[0094] Therefore, in the step S200, the flight intelligent agent model outputs a first flight simulation scheduling result, which specifically includes:

[0095] All flight operation data stored in the ring data queue are input into a pre-trained action path prediction model based on a Boolean network model;

[0096] Solve an optimal control sequence of the Boolean network model based on at least one optimal solution solving algorithm, and obtain an optimal action path, which represents an optimal action order of all flights of the current airport under constraints of current original flight operation data.

[0097] In Figures 1-3 Based on the detailed introduction of the method embodiment, further refer to Figure 4 , Figure 4 is a functional unit combination schematic diagram of a flight operation data quality evaluation and analysis system for implementing the flight operation data quality evaluation and analysis method.

[0098] Figure 4 Three functional modules are simply shown: data middleware, flight intelligent agent model and quality evaluation unit.

[0099] It can be understood that the related principles and steps of the system and device embodiments correspond to the method embodiment, so they do not need to be repeated and expanded, and they can be mutually referenced and referred to.

[0100] Specifically, the specific functions of the three functional modules are:

[0101] Data middleware: based on the data pipeline, receive original flight operation data and current actual flight scheduling results corresponding to the original flight operation data and store them to a ring-shaped data queue;

[0102] Flight intelligent agent model: when the ring-shaped data queue is full, obtain all flight operation data from the ring-shaped data queue, and output a first flight simulation scheduling result;

[0103] Quality evaluation unit: when the similarity of the first flight simulation scheduling result and the current actual flight scheduling result meets a preset condition, it is determined that the original flight operation data passes the data quality evaluation and can be stored as selected flight training data or record data; otherwise, the original flight operation data is checked.

[0104] The flight intelligent agent model is a pre-trained action path prediction model based on a Boolean network model, and the action path prediction model generates an optimal flight scheduling order based on the original flight operation data.

[0105] Other technologies, principles, algorithms or models not expanded in detail in the present application can refer to prior art.

[0106] The application sends original flight operation data to data middleware, the data middleware receives the original flight operation data based on a data pipeline and stores to a ring data queue; when the ring data queue is full, all flight operation data stored in the ring data queue is obtained and input to a pre-trained flight intelligent agent model, the flight intelligent agent model outputs a first flight simulation scheduling result; when the similarity of the first flight simulation scheduling result and the current actual flight scheduling result meets a preset condition, it is determined that the original flight operation data passes the data quality evaluation; otherwise, the original flight operation data is checked. The pre-trained flight intelligent agent model is an action path prediction model based on a Boolean network model. The application can ensure the collection and storage quality level of flight operation data.

[0107] In the foregoing embodiment section, the application gives a plurality of embodiments, each of which can constitute an independent technical solution and can contribute to the prior art and solve the corresponding technical problems. However, it should be pointed out that different embodiments can be combined with each other without violating the logic; at the same time, each embodiment can solve at least one technical problem, but it does not require each individual embodiment to solve multiple or all technical problems.

[0108] The method embodiments and system of the application have been shown and described in the foregoing, but it can be understood by those skilled in the art that various changes, modifications, replacements and variations can be made to these embodiments without departing from the principles and spirits of the application, and the scope of the application is defined by the appended claims and their equivalents.

Claims

1. A flight operation data quality assessment and analysis method, which is implemented based on data middleware and is characterized by: The method comprises the following steps: Sending the original flight operation data to the data middleware, the data middleware receives the original flight operation data based on a data pipeline and stores the data in a circular data queue; When the circular data queue is full, all flight operation data stored in the circular data queue is obtained and input into a pre-trained flight agent model, and the flight agent model outputs a first flight simulation scheduling result; When the similarity between the first flight simulation scheduling result and the current actual flight scheduling result meets a preset condition, it is determined that the original flight operation data passes the data quality assessment.

2. The flight operation data quality assessment and analysis method according to claim 1, characterized in that: The data middleware receives the original flight operation data based on the data pipeline, and also receives the current actual flight scheduling result corresponding to the original flight operation data and stores it in the circular data queue.

3. The flight operation data quality assessment and analysis method according to claim 1, characterized in that: The original flight operation data includes: the destination of the current flight, the route of the current flight, the load factor of the current flight, the revenue per passenger kilometer of the current flight, the cost per seat kilometer, all arrival flight data of the current airport, and all departure flight data.

4. The flight operation data quality assessment and analysis method according to claim 1, characterized in that: When the similarity between the first flight simulation scheduling result and the current actual flight scheduling result does not meet a preset condition, the original flight operation data is checked.

5. The flight operation data quality assessment and analysis method according to claim 1, characterized in that: The pre-trained flight agent model is a motion path prediction model based on a Boolean network model, and the motion path prediction model generates an optimal flight scheduling sequence based on the original flight operation data.

6. The flight operation data quality assessment and analysis method according to claim 5, characterized in that: The Boolean network model is constructed as follows: Collect raw flight operation data and define logical variables for each flight at the current airport; Boolean algebra is used to represent the logical variables and their relationships of all flights, and a Boolean network model of flight operations is established.

7. The flight operation data quality assessment and analysis method according to claim 5, characterized in that: The flight agent model outputs a first flight simulation scheduling result, specifically including: Inputting all flight operation data stored in the circular data queue into a pre-trained motion path prediction model based on a Boolean network model; Based on at least one optimal solution algorithm, the optimal control sequence of the Boolean network model is solved to obtain an optimal action path, where the optimal action path represents the optimal action sequence of all flights at the current airport based on the current original flight operation data.

8. A flight operation data quality assessment and analysis system, characterized by: The system comprises: Data middleware: receives original flight operation data and current actual flight scheduling results corresponding to the original flight operation data based on the data pipeline and stores them in a circular data queue; Flight agent model: when the circular data queue is full, it obtains all flight operation data from the circular data queue and outputs the first flight simulation scheduling result; Quality assessment unit: When the similarity between the first flight simulation scheduling result and the current actual flight scheduling result meets the preset conditions, it is determined that the original flight operation data passes the data quality assessment; otherwise, the original flight operation data is verified.

9. The flight operation data quality assessment and analysis system according to claim 8, characterized in that: The flight agent model is a pre-trained action path prediction model based on a Boolean network model, and the action path prediction model generates an optimal flight scheduling sequence based on the original flight operation data.

10. A computer-readable storage medium having a computer program stored thereon, wherein: When the computer program is executed by a processor, the steps of the flight operation data quality assessment and analysis method according to any one of claims 1 to 7 are implemented.

Images