Personal portrayal-based controller matching method and device

Abstract

The embodiment of the invention discloses a controller matching method and device based on figure portraits. The method comprises the following steps: constructing an index system of an air traffic controller figure portrait; designing a mapping rule based on non-numeric indexes in the index system; inputting controller information, and performing preprocessing, numerical type conversion and representation vector construction on the controller information to obtain a controller representation vector; performing feature extraction by using a pre-constructed auto-encoder model to obtain a low-dimensional dense vector; and based on the low-dimensional dense feature vector, setting an optimization target according to a matching demand, and generating a controller matching scheme by using a greedy algorithm. According to the method, the problems that the number and types of indexes of an existing controller portrait index system are not rich enough, and accurate portraying cannot be achieved can be solved, and the obtained controller matching scheme is more objective.

Description

Technical Field

[0001] The present invention relates to the technical field of air traffic control, and particularly relates to a controller matching method and device based on a personal portrait. Background Art

[0002] With the rapid development of the national economy, the civil aviation transportation volume has increased sharply, resulting in increasingly busy air traffic, thus increasing the potential safety hazards and the occurrence of unsafe events in air traffic control. How to effectively reduce these unsafe events has become an urgent problem to be solved. Air traffic control work requires close cooperation between various seats to command aircraft. Improving the complementary and balanced capabilities of the seat duty personnel matching can enhance their insight and handling ability of unsafe events, thereby effectively improving the air traffic control service level and reducing the risk of unsafe events.

[0003] However, the current controller matching scheme is mainly formulated based on experience, with a large degree of subjectivity and insufficient consideration of the scientificity and rationality of personnel matching. Specifically, the existing controller matching has the following obvious drawbacks:

[0004] (1) In terms of indicators: The quantity and type of indicators are not rich enough to accurately portrait.

[0005] (2) In terms of research methods: There are more qualitative studies and fewer quantitative studies in the existing technology, and fewer objective conclusions can be drawn. Summary of the Invention

[0006] Aiming at the defects of the existing technology mentioned in the background art, the purpose of the embodiments of the present invention is to provide a controller matching method and device based on a personal portrait.

[0007] To achieve the above purpose, in the first aspect, the embodiments of the present invention provide a controller matching method based on a personal portrait, including:

[0008] Constructing an index system for the personal portrait of air traffic controllers; the index system is used to objectively evaluate controllers from multiple perspectives; the index system includes non-numerical indicators;

[0009] Designing a mapping rule based on the non-numerical indicators;

[0010] Inputting controller information, preprocessing, numerically converting and constructing a representation vector for the controller information to obtain a controller representation vector; the numerical conversion is implemented based on the mapping rule;

[0011] Using a pre-constructed autoencoder model to extract abstract features from the controller representation vector and performing compression processing on the abstract features to obtain a low-dimensional dense vector;

[0012] Based on the low-dimensional dense feature vectors, an optimization objective is set according to the matching requirements, and a controller matching scheme is generated using a greedy algorithm.

[0013] As a specific implementation manner of the present application, the index system for constructing the portrait of an air traffic controller is specifically as follows:

[0014] Based on business knowledge and expert experience, an index system for constructing the portrait of an air traffic controller is constructed;

[0015] Among them, the index system includes first-level indexes, second-level indexes, and third-level indexes.

[0016] As a specific implementation manner of the present application, obtaining the controller representation vector is specifically as follows:

[0017] Enter the controller information according to the index system of the portrait of the air traffic controller;

[0018] Perform missing value processing and outlier processing on the controller information to obtain the data to be processed; among them, the missing value processing uses mean filling, and the outlier processing uses boundary value replacement;

[0019] Based on the mapping rule, convert the non-numeric index values in the data to be processed into numeric scores;

[0020] Use the maximum-minimum normalization method to standardize the numeric scores;

[0021] For each controller, construct a controller representation vector based on the standardized data.

[0022] As a specific implementation manner of the present application, the autoencoder model includes an encoder and a decoder both using two-layer fully connected networks; obtaining the low-dimensional dense vector is specifically as follows:

[0023] Use the encoder to perform feature extraction on the input controller representation vector and output the low-dimensional dense vector.

[0024] As a specific implementation manner of the present application, generating a controller matching scheme is specifically as follows:

[0025] Based on the low-dimensional dense feature vectors, calculate the cosine similarity between each controller;

[0026] According to the matching requirements, set an optimization objective and use a greedy algorithm to generate a controller matching scheme.

[0027] In a second aspect, the embodiments of the present application further provide a controller matching device based on a portrait, including:

[0028] An index system construction unit for constructing an index system for the portrait of an air traffic controller; the index system is used to objectively evaluate controllers from multiple perspectives; the index system includes non-numerical indexes;

[0029] A rule design unit for designing a mapping rule based on the non-numerical indexes;

[0030] A characterization vector extraction unit for inputting controller information, preprocessing the controller information, performing numerical conversion and constructing a characterization vector to obtain a controller characterization vector; the numerical conversion is implemented based on the mapping rule;

[0031] A low-dimensional dense vector extraction unit for extracting abstract features from the controller characterization vector by using a pre-constructed autoencoder model and performing compression processing on the abstract features to obtain a low-dimensional dense vector;

[0032] A matching scheme generation unit for setting an optimization target according to the matching requirements based on the low-dimensional dense feature vector and using a greedy algorithm to generate a controller matching scheme.

[0033] In a third aspect, an embodiment of the present invention further provides another controller matching device based on a portrait, including a processor, an input device, an output device and a memory, the processor, the input device, the output device and the memory are interconnected, wherein the memory is used to store a computer program, the computer program includes program instructions, and the processor is configured to call the program instructions to execute the method in the first aspect above.

[0034] Implementing the embodiments of the present invention has the following advantages:

[0035] 1. The present invention constructs an evaluation index system for the portrait of an air traffic controller, which can solve the problem that the existing controller portrait index system has insufficient quantity and types of indexes and cannot accurately portrait. Specifically, the evaluation index system can accurately portrait controllers from six aspects: basic information, health status, service awareness, safety record, business ability, and experience.

[0036] 2. The present invention generates a controller characterization vector based on the constructed air traffic controller portrait index system, extracts features and reduces the dimension of the characterization vector through an autoencoder model to generate a low-dimensional dense characterization of the controller portrait information, then calculates the similarity between controllers based on the low-dimensional dense characterization, sets an optimization target according to the matching requirements, and uses a greedy algorithm to scientifically solve the controller matching scheme; the obtained controller matching scheme is more objective. Description of the Drawings

[0037] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the following will briefly introduce the accompanying drawings required for the description of the specific embodiments or the prior art.

[0038] Figure 1 is the flowchart of the controller matching method based on the portrait of the person provided by the embodiment of the present invention;

[0039] Figure 2a and Figure 2b is the schematic diagram of the air traffic controller portrait index system;

[0040] Figure 3 is the schematic diagram of the structure of the autoencoder model;

[0041] Figure 4 is the schematic diagram of the visualization of the controller's low-dimensional dense vector;

[0042] Figure 5 is the schematic diagram of the controller matching scheme

[0043] Figure 6 is the structure diagram of the controller matching device based on the portrait of the person provided by the embodiment of the present invention;

[0044] Figure 7 is Figure 6 another structure diagram of the shown device. Specific embodiments

[0045] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0046] It should be understood that when used in this specification and the appended claims, the terms "comprises" and "comprising" indicate the presence of the described features, wholes, steps, operations, elements, and / or components, but do not preclude the presence or addition of one or more other features, wholes, steps, operations, elements, components, and / or their combinations.

[0047] Please refer to Figure 1 , which is the controller matching method based on the portrait of the person provided by the embodiment of the present invention, including:

[0048] S1. Construct an air traffic controller portrait index system to objectively evaluate the controller from multiple dimensions.

[0049] In specific implementation, based on business knowledge and expert experience, this embodiment constructs a three-level portrait index system for air traffic controllers. As Figure 2a and Figure 2b shown, this portrait index system specifically includes six first-level indicators, namely basic information, health status, service awareness, safety record, professional ability, and experience, twenty-nine second-level indicators, and sixty-three third-level indicators.

[0050] Among them, the first-level indicator of basic information includes nine second-level indicators, namely name, age, political status, education level, English level, rank, qualification, solo flight status, and solo flight years;

[0051] The first-level indicator of health status includes four second-level indicators, namely fatigue level, physical health, mental health, and workload;

[0052] The first-level indicator of service awareness includes four second-level indicators, namely active announcements in the past 30 days, corrections to flight crews in the past 30 days, voluntary reports in the past 30 days, and safety suggestions in the past 30 days;

[0053] The first-level indicator of safety record includes five second-level indicators, namely abnormal events, non-consequential violations, abnormal behaviors, bad records, and consecutive safe operation duration;

[0054] The first-level indicator of professional ability includes three second-level indicators, namely control standardization, control efficiency, and control stability;

[0055] The first-level indicator of experience includes four second-level indicators, namely training performance, work experience, awards, and special skills;

[0056] The second-level indicator of rank includes three third-level indicators, namely job rank, professional title, and position;

[0057] The second-level indicator of qualification includes four third-level indicators, namely license qualification, instructor qualification, inspector qualification, and special skills;

[0058] The second-level indicator of workload includes four third-level indicators, namely cumulative working hours in the past 30 days, consecutive working hours, average call load in the past day, and proportion of high call load hours in the past day;

[0059] The second-level indicator of abnormal events includes three third-level indicators, namely the number of abnormal events in the past 7 days, the number of abnormal events in the past 30 days, and the number of abnormal events in the past 90 days;

[0060] The second-level indicator of non-consequential violations includes three third-level indicators, namely the number of non-consequential violations in the past 7 days, the number of non-consequential violations in the past 30 days, and the number of non-consequential violations in the past 90 days;

[0061] The secondary indicator of abnormal behavior includes nine tertiary indicators, namely the number of sleeping-on-duty times in the past 1 day, the number of leaving-post times in the past 1 day, the number of other abnormal behavior times in the past 1 day, the number of sleeping-on-duty times in the past 7 days, the number of leaving-post times in the past 7 days, the number of other abnormal behavior times in the past 7 days, the number of sleeping-on-duty times in the past 30 days, the number of leaving-post times in the past 30 days, and the number of other abnormal behavior times in the past 30 days;

[0062] The secondary indicator of bad records includes three tertiary indicators, namely the number of bad record times in the past 7 days, the number of bad record times in the past 30 days, and the number of bad record times in the past 90 days;

[0063] The secondary indicator of training performance includes twenty tertiary indicators, namely the cumulative number of training times for pre-job training, the cumulative training duration for pre-job training, the one-time passing rate of the pre-job training assessment, the excellent rate of the pre-job training assessment, the cumulative number of training times for on-the-job training, the cumulative training duration for on-the-job training, the one-time passing rate of the on-the-job training assessment, the excellent rate of the on-the-job training assessment, the cumulative number of training times for safety education training, the cumulative training duration for safety education training, the one-time passing rate of the safety education training assessment, the excellent rate of the safety education training assessment, the cumulative number of training times for retraining, the cumulative training duration for retraining, the one-time passing rate of the retraining assessment, the excellent rate of the retraining assessment, the cumulative number of training times for business training, the cumulative training duration for business training, the one-time passing rate of the business training assessment, and the excellent rate of the business training assessment;

[0064] The secondary indicator of work experience includes three tertiary indicators, namely the cumulative duty duration, the duration of communication with the airspace, and the proportion of English communication;

[0065] The secondary indicator of award situation includes six tertiary indicators, namely department-level awards, sub-bureau-level awards, regional-level awards, air traffic control bureau-level / municipal-level awards, Civil Aviation Administration of China-level / provincial-level / association-level awards, and national-level awards;

[0066] The secondary indicator of special work includes five tertiary indicators, namely department-level special work, sub-bureau-level special work, regional-level special work, air traffic control bureau-level special work, and Civil Aviation Administration of China-level special work.

[0067] S2. According to the non-numerical indicators in the air traffic controller persona index system, design the corresponding mapping rules.

[0068] S3. Input the controller information, and construct its vectorized representation after corresponding data processing.

[0069] In specific implementation, step S3 specifically includes:

[0070] S31. According to the tertiary air traffic controller persona index system constructed in S1, input the controller information.

[0071] S32. Process the missing values and outliers. The missing values are filled with the mean value, and the outliers are replaced with the boundary values.

[0072] S33. Use the mapping rules in S2 to convert non - numerical index values into numerical scores.

[0073] S34. Standardize the data using the min - max normalization method. The specific process is as follows:

[0074]

[0075] where \(x\) is the input value of any index, \(x\) min is the minimum value of this index, \(x\) max is the maximum value of this index, is the output value of this index after min - max normalization.

[0076] S35. Based on the standardized data, construct a vectorized representation of the portrait information for each controller The feature dimension is 80. Where \(X\) i represents the feature vector of the \(i\) - th controller, represents the value of the first index of the \(i\) - th controller after standardization.

[0077] S4. Construct an auto - encoder model to extract the abstract features in the controller representation vector and compress it into a low - dimensional dense vector containing rich semantic information.

[0078] When specifically implemented, step S4 specifically includes:

[0079] S41. Construct an auto - encoder model. As Figure 3 shown, the model is divided into two parts: an encoder and a decoder. Both the encoder and the decoder use two - layer fully - connected networks; among them, the activation function uses Relu, the loss function uses MSE, the optimization algorithm uses Adam, the number of iterations is 100, and the learning rate is 0.0005;

[0080] S42. The encoder performs data encoding on the input data to achieve feature extraction; the decoder performs data decoding on the output of the encoder to achieve data reconstruction. The processes of encoding and decoding are as follows:

[0081] \(H\) i \(=\sigma\) f (W1X i +b1)

[0082] \(Y\) i \(=\sigma\) g (W2H i +b2)

[0083] where \(H\) i is the output of the encoder, \(\sigma\) fis the activation function of the encoder, W1 is the weight matrix of the encoder, and b1 is the bias vector of the encoder. Y i is the output of the decoder, s g is the activation function of the decoder, W2 is the weight matrix of the encoder, and b2 is the bias vector of the encoder.

[0084] S43. Using the output H of the encoder i as the low-dimensional dense representation of the controller portrait information, as Figure 4 shown, which is the visualization result of the low-dimensional dense representation.

[0085] S5. Based on the extracted low-dimensional dense feature vectors, set the optimization goal according to the matching requirements, and use the greedy algorithm to generate the controller matching scheme.

[0086] Specifically, in implementation, step S5 specifically includes:

[0087] S51. Based on the low-dimensional dense representation of the controller portrait information generated in S4, calculate the cosine similarity between each controller. The specific process is as follows:

[0088]

[0089] Among them, S(H i , H j ) represents the cosine similarity between the two vectors H i and H j , n is the dimension of the vector H i , represents the k-th dimensional feature of the vector H i .

[0090] S52. Set the optimization goal according to the matching requirements. In this embodiment, the goal is to minimize the cosine similarity of the controller matching combination and maximize the degree of complementary capabilities, generate the most complementary matching combination, and use the greedy algorithm to solve the controller matching scheme, as Figure 5 shown.

[0091] From the above description, it can be known that implementing the embodiments of the present invention has the following advantages:

[0092] 1. The present invention constructs an evaluation index system for the portrait of air traffic controllers, which can solve the problem that the existing controller portrait index system has insufficient quantity and types of indicators and cannot accurately portrait. Specifically, the evaluation index system can accurately portrait controllers from six aspects: basic information, health status, service awareness, safety record, business ability, and experience.

[0093] 2. Based on the constructed index system of the air traffic controller's portrait, the present invention generates a controller representation vector, extracts features and reduces the dimension of the representation vector through an autoencoder model to generate a low-dimensional dense representation of the controller portrait information, then calculates the similarity between controllers based on the low-dimensional dense representation, sets an optimization objective according to the matching requirements, and uses a greedy algorithm to scientifically solve the controller matching scheme; the obtained controller matching scheme is more objective.

[0094] Based on the same inventive concept, as Figure 6 shown, an embodiment of the present invention further provides a controller matching device based on a portrait, including:

[0095] An index system construction unit, configured to construct an index system of the air traffic controller's portrait; the index system is used to objectively evaluate the controller from multiple perspectives; the index system includes non-numerical indexes;

[0096] A rule design unit, configured to design a mapping rule based on the non-numerical index;

[0097] A representation vector extraction unit, configured to input controller information, perform preprocessing, numerical conversion and representation vector construction on the controller information to obtain a controller representation vector; the numerical conversion is implemented based on the mapping rule;

[0098] A low-dimensional dense vector extraction unit, configured to extract abstract features in the controller representation vector by using a pre-constructed autoencoder model, and perform compression processing on the abstract features to obtain a low-dimensional dense vector;

[0099] A matching scheme generation unit, configured to set an optimization objective according to the matching requirements based on the low-dimensional dense feature vector, and use a greedy algorithm to generate a controller matching scheme.

[0100] Specifically, the representation vector extraction unit is specifically configured to:

[0101] Enter controller information according to the index system of the air traffic controller's portrait;

[0102] Perform missing value processing and outlier processing on the controller information to obtain data to be processed; wherein, the missing value processing uses mean filling, and the outlier processing uses boundary value replacement;

[0103] Based on the mapping rule, convert the non-numerical index values in the data to be processed into numerical scores;

[0104] Use the maximum-minimum normalization method to standardize the numerical scores;

[0105] For each controller, a controller representation vector is constructed based on the standardized data.

[0106] Specifically, the autoencoder model includes an encoder and a decoder, both of which adopt two-layer fully connected networks; the low-dimensional dense vector is specifically used for:

[0107] The encoder is used to extract features from the input controller representation vector and output the low-dimensional dense vector.

[0108] Specifically, the matching scheme generation unit is specifically used for:

[0109] Based on the low-dimensional dense feature vector, calculate the cosine similarity between each pair of controllers;

[0110] Set an optimization goal according to the matching requirements and use the greedy algorithm to generate a controller matching scheme.

[0111] Specifically, the matching scheme generation unit is specifically used for:

[0112] Based on the low-dimensional dense feature vector, calculate the cosine similarity between each pair of controllers;

[0113] Set an optimization goal according to the matching requirements and use the greedy algorithm to generate a controller matching scheme.

[0114] It should be noted that for the specific working process of this embodiment, please refer to the method embodiment part above and will not be elaborated here.

[0115] Furthermore, another embodiment of the present invention also provides a controller matching device based on a person portrait. As Figure 7 shown, the device may include: one or more processors 101, one or more input devices 102, one or more output devices 103, and a memory 104. The above-mentioned processors 101, input devices 102, output devices 103, and memory 104 are interconnected through a bus 105. The memory 104 is used to store a computer program, and the computer program includes program instructions. The processor 101 is configured to call the program instructions to execute the method in the method embodiment part above.

[0116] It should be understood that in the embodiments of the present invention, the so-called processor 101 may be a central processing unit (CPU), and the processor may also be other general-purpose processors, digital signal processors (DSPs), application specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or the processor may also be any conventional processor, etc.

[0117] The input device 102 may include a keyboard, etc., and the output device 103 may include a display (such as an LCD), a speaker, etc.

[0118] The memory 104 may include a read-only memory and a random access memory, and provide instructions and data to the processor 101. A part of the memory 104 may also include a non-volatile random access memory. For example, the memory 104 may also store information about the device type.

[0119] In specific implementation, the processor 101, input device 102, and output device 103 described in the embodiments of the present invention may execute the implementation manners described in the embodiments of the controller matching method based on a person portrait provided by the embodiments of the present invention, which will not be elaborated here.

[0120] Those of ordinary skill in the art can realize that the units and algorithm steps of each example described in combination with the embodiments disclosed in this article can be implemented by electronic hardware, computer software, or a combination of the two. To clearly illustrate the interchangeability of hardware and software, the components and steps of each example have been generally described according to functions in the above description. Whether these functions are executed in a hardware or software manner depends on the specific application and design constraints of the technical solution. Professional technicians can use different methods to implement the described functions for each specific application, but such implementation should not be considered to exceed the scope of the present invention.

[0121] In several embodiments provided in this application, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are merely illustrative. For example, the division of the units is only a logical function division. In actual implementation, there may be other division methods. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the displayed or discussed couplings or direct couplings or communication connections to each other can be indirect couplings or communication connections through some interfaces, devices or units, or can also be electrical, mechanical or other forms of connection.

[0122] The units described as separate components may or may not be physically separated. The components displayed as units may or may not be physical units, that is, they can be located in one place, or can also be distributed to multiple network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of the embodiments of the present invention.

[0123] In addition, each functional unit in various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically alone, or two or more units can be integrated into one unit. The above-mentioned integrated units can be implemented in the form of hardware or in the form of software functional units.

[0124] If the above-mentioned integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of this technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to enable a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in various embodiments of the present invention. The foregoing storage medium includes: various media such as USB flash drives, mobile hard disks, read-only memories (ROMs), random access memories (RAMs), magnetic disks, or optical discs that can store program codes.

[0125] The above is only the specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present invention can easily think of various equivalent modifications or substitutions, and these modifications or substitutions should all be covered within the protection scope of the present invention. Therefore, the protection scope of the present invention should be subject to the protection scope of the claims.

Claims

1. A controller matching method based on a portrait of a person, characterized in that Including: Construct an index system for building the portrait of an air traffic controller; the index system is used to objectively evaluate the controller from multiple perspectives; the index system includes non-numerical indexes; Design a mapping rule based on the non-numerical indexes; Input the controller information, preprocess, numerically transform, and construct a representation vector for the controller information to obtain a controller representation vector; the numerical transformation is implemented based on the mapping rule; Use a pre-constructed autoencoder model to extract the abstract features in the controller representation vector and perform compression processing on the abstract features to obtain a low-dimensional dense vector; Based on the low-dimensional dense feature vector, set an optimization objective according to the matching requirements and use a greedy algorithm to generate a controller matching scheme.

2. The controller allocation method according to claim 1, wherein Specifically, constructing the index system for building the portrait of an air traffic controller is as follows: Based on business knowledge and expert experience, construct an index system for building the portrait of an air traffic controller; Among them, the index system includes first-level indexes, second-level indexes, and third-level indexes.

3. The controller matching method according to claim 1, characterized in that, Specifically, obtaining the controller representation vector is as follows: Input the controller information according to the index system for building the portrait of an air traffic controller; Perform missing value processing and outlier processing on the controller information to obtain data to be processed; among them, the missing value processing uses mean filling, and the outlier processing uses boundary value replacement; Based on the mapping rule, convert the non-numerical index values in the data to be processed into numerical scores; Use the maximum-minimum normalization method to standardize the numerical scores; For each controller, construct a controller representation vector based on the standardized data.

4. The controller allocation method according to claim 1, wherein The autoencoder model includes an encoder and a decoder both using two-layer fully connected networks; specifically, obtaining the low-dimensional dense vector is as follows: Use the encoder to extract features from the input controller representation vector and output the low-dimensional dense vector.

5. The controller collocation method according to claim 1, characterized in that Specifically, generating the controller matching scheme is as follows: Based on the low-dimensional dense feature vector, calculate the cosine similarity between each controller; Set an optimization objective according to the matching requirements and use a greedy algorithm to generate a controller matching scheme.

6. A controller collocation device based on a portrait of a person, characterized in that, Including: An index system construction unit for constructing an index system for building the portrait of an air traffic controller; the index system is used to objectively evaluate the controller from multiple perspectives; the index system includes non-numerical indexes; A rule design unit for designing a mapping rule based on the non-numerical indexes; A representation vector extraction unit for inputting the controller information, preprocessing, numerically transforming, and constructing a representation vector for the controller information to obtain a controller representation vector; the numerical transformation is implemented based on the mapping rule; A low-dimensional dense vector extraction unit for using a pre-constructed autoencoder model to extract the abstract features in the controller representation vector and performing compression processing on the abstract features to obtain a low-dimensional dense vector; A matching scheme generation unit for setting an optimization objective according to the matching requirements based on the low-dimensional dense feature vector and using a greedy algorithm to generate a controller matching scheme.

7. The controller matching device according to claim 6, wherein Specifically, the representation vector extraction unit is used for: Input the controller information according to the index system for building the portrait of an air traffic controller; Perform missing value processing and outlier processing on the controller information to obtain the data to be processed; wherein, the missing value processing uses mean filling, and the outlier processing uses boundary value replacement; Based on the mapping rule, convert the non-numerical index values in the data to be processed into numerical scores; Use the maximum-minimum normalization method to standardize the numerical scores; For each controller, construct a controller representation vector based on the standardized data.

8. The controller matching device according to claim 6, wherein, The autoencoder model includes an encoder and a decoder both using two-layer fully connected networks; the low-dimensional dense vector is specifically used for: Use the encoder to extract features from the input controller representation vector and output the low-dimensional dense vector.

9. The controller matching device according to claim 6, characterized in that, The matching scheme generation unit is specifically used for: Based on the low-dimensional dense feature vector, calculate the cosine similarity between each pair of controllers; Set the optimization objective according to the matching requirements and use the greedy algorithm to generate the controller matching scheme.

10. A controller matching device based on a portrait of a person, characterized in that, It includes a processor, an input device, an output device and a memory. The processor, input device, output device and memory are interconnected. Among them, the memory is used to store computer programs. The computer programs include program instructions. The processor is configured to call the program instructions to execute the method according to any one of claims 1-5.

Images