Multi-level dynamic ability assessment method and system for operation and maintenance personnel

Abstract

The invention belongs to the technical field of management, and relates to a multi-level dynamic ability assessment method and system for operation and maintenance personnel, and the method comprises the steps: firstly collecting, storing and processing multi-dimensional data; the method comprises the following steps: establishing a graph neural network, performing data processing and layering, constructing a multi-level dynamic evaluation model based on the graph neural network, training the multi-level dynamic evaluation model, and performing dynamic evaluation on operation and maintenance personnel according to the trained multi-level dynamic evaluation model; finally, according to operation modes of different parking lots, weights of all indexes are set, comprehensive evaluation is carried out in combination with evaluation results of the multi-level dynamic evaluation model, the system is combined with multi-dimensional data, multi-level capability evaluation is provided through comprehensive analysis of the multi-level dynamic evaluation model, the comprehensive capability and work performance of operation and maintenance personnel are comprehensively reflected, and the operation and maintenance efficiency is improved. Moreover, differential evaluation can be carried out according to the operation requirements of different parking lots, and the evaluation result is more accurate.

Description

Technical Field

[0001] The present invention belongs to the field of management technology and relates to a multi-level dynamic capability evaluation method and system for operation and maintenance personnel. Background Art

[0002] The ability assessment of operation and maintenance personnel is of great significance and role in intelligent operation and maintenance management. By accurately analyzing the performance of operation and maintenance personnel in multiple dimensions such as repair processing time, equipment failure rate and service user evaluation, it can help managers fully understand the actual work ability of operation and maintenance personnel, which not only helps to improve the efficiency of operation and maintenance management, but also provides managers with scientific decision-making support, thereby optimizing resource allocation and improving service quality and customer satisfaction.

[0003] Traditional evaluation methods for operation and maintenance personnel mainly rely on static performance appraisals, which mainly include annual or quarterly appraisals and single-dimensional appraisals. Annual or quarterly appraisals are usually based on a fixed annual or quarterly appraisal cycle to evaluate the overall performance of the operation and maintenance personnel during this cycle. Managers will regularly summarize the work of the operation and maintenance personnel and score them according to preset standards; single-dimensional appraisal content mainly focuses on single dimensions such as repair processing time, equipment failure rate and service user evaluation. Managers usually measure the performance of operation and maintenance personnel through these single indicators, such as the average time for statistically processing faults, the monthly equipment failure rate and user feedback scores, which cannot reflect the dynamic performance of the operation and maintenance personnel in real time.

[0004] Some of these methods usually rely on simple data analysis and statistical models, such as frequency statistics, cluster analysis and preset thresholds. Frequency statistics uses simple frequency statistics methods to analyze the number of faults handled by operation and maintenance personnel within a certain period of time and the repair response speed. For example, it counts how many faults each operation and maintenance personnel handles in a month, the average time for each fault handling, etc. Cluster analysis uses cluster analysis technology to divide operation and maintenance personnel into different performance levels according to their performance data. By analyzing these data, managers classify operation and maintenance personnel into different levels such as excellent performance, average performance and need for improvement, so as to carry out more targeted management. Preset thresholds use preset thresholds to evaluate the performance of operation and maintenance personnel. For example, it is stipulated that the time for handling faults cannot exceed a certain value, or the monthly equipment failure rate cannot exceed a certain percentage. Exceeding these thresholds is considered to be non-compliant. Most of these methods are used to identify operation and maintenance performance within a specific time period. They have obvious limitations in multi-dimensional data integration and dynamic evaluation. It is difficult to comprehensively analyze multi-dimensional data such as repair processing time, equipment failure rate, and service user evaluation, which leads to one-sided evaluation results.

[0005] There are also some methods that try to combine machine learning technology, such as regression analysis and classification models. The regression analysis model predicts the future performance of operation and maintenance personnel based on historical performance data. Through linear regression or multivariate regression models, the relationship between repair processing time, equipment failure rate and user evaluation is analyzed, and future performance is predicted; classification models, such as decision trees, random forests and support vector machines, classify and evaluate operation and maintenance personnel based on multi-dimensional historical data. Through these models, operation and maintenance personnel can be divided into different categories such as efficient, general and inefficient, helping managers to carry out targeted management. However, these methods have limited capabilities in processing complex spatiotemporal data and high-dimensional data. These models often cannot capture the complex dynamic relationship of operation and maintenance personnel's behavior, resulting in the lack of accuracy and depth in the evaluation results.

[0006] In summary, existing methods are usually unable to effectively handle the dynamic interactions of operation and maintenance personnel at different times and places, and it is difficult to reveal the true work performance and behavior patterns of operation and maintenance personnel in complex environments. Evaluation methods that lack spatiotemporal dynamics cannot provide a deep understanding of the behavior of operation and maintenance personnel, limiting the comprehensiveness and accuracy of the evaluation. Therefore, a new multi-level dynamic capability evaluation method and system for operation and maintenance personnel is urgently needed. Summary of the invention

[0007] In view of the shortcomings of the existing methods, the present invention proposes a multi-level dynamic capability evaluation method and system for operation and maintenance personnel, which are used to deeply analyze and evaluate the achievement of various task indicators of operation and maintenance personnel.

[0008] To achieve the above object, the present invention adopts the following technical solution: In a first aspect, the present invention provides a multi-level dynamic capability assessment method for operation and maintenance personnel, which specifically comprises the following steps: S1. Data collection: Collect multi-dimensional data such as repair processing time, equipment failure rate, and service user evaluation, and use secure communication protocols (such as HTTPS) to transmit the collected data in real time to the central database for storage and processing; S2. Data processing and stratification: The collected data is cleaned, outlier detected and standardized in turn, and each indicator is stratified according to the preset standards; S3. Multi-level dynamic evaluation: Build a multi-level dynamic evaluation model based on graph neural networks (GCNs), train the multi-level dynamic evaluation model, and dynamically evaluate the operation and maintenance personnel based on the trained multi-level dynamic evaluation model; S4. Differentiated weight allocation and evaluation: according to the operation modes of different parking lots, the weights of various indicators are set, and a comprehensive evaluation is conducted in combination with the evaluation results of step S3.

[0009] As a further technical solution of the present invention, the repair processing time described in step S1 is automatically obtained through the API interface of the integrated repair system; the equipment failure rate is obtained through sensors and IoT devices, and the service user evaluation is obtained through a user feedback system or application.

[0010] As a further technical solution of the present invention, the specific process of step S2 is: S21, clean the data collected in step S1 to remove noise data and outliers. Data cleaning includes missing value processing. The mean filling method is used for missing value processing. ,in is the missing value neighboring data, is the number of adjacent data of the missing value; S22. Use the Isolation Forest algorithm to detect outliers. The Isolation Forest algorithm builds a tree by randomly selecting a subset of the data and randomly selecting features, and detects outliers by path length. The anomaly score , in, It’s data The average path length in the isolation tree, is the adjustment factor; S23, perform data standardization processing to convert the data into a standard normal distribution with a mean of 0 and a standard deviation of 1; S24, perform interval stratification, divide each indicator into different levels according to historical data and industry standards, and use K-means clustering algorithm to determine the boundaries of each interval. ,in is a cluster set, For the Cluster The center of mass.

[0011] As a further technical solution of the present invention, the specific process of step S3 is: S31. Construct a graph model that reflects the relationship between operation and maintenance personnel, parking lots, and equipment, and map the interactive relationship between operation and maintenance personnel at different times and locations into a graph structure. Specifically, define the operation and maintenance personnel, parking lots, and equipment as nodes in the graph, and define the relationship between nodes (such as the equipment and parking lots that the operation and maintenance personnel are responsible for) as edges. Construct an adjacency matrix A to represent the relationship between nodes in the graph, and construct a feature matrix X to represent the feature vector of each node. The feature vector of the node includes historical performance data, equipment status data, and vehicle factory operation data. S32. Use graph neural networks (GCNs) to process the constructed graph model and design a graph convolution layer to capture high-order dependencies and dynamic interactions between nodes to obtain a multi-level dynamic evaluation model. The graph convolution layer propagates information through the product of the adjacency matrix and the feature matrix: ,in, , is the adjacency matrix after adding self-loops, is the identity matrix, yes The degree matrix of It is The feature matrix of the layer, It is The weight matrix of the layer, is the activation function ReLU; high-order features are extracted by stacking three layers of graph convolutional layers. The output of each layer is used as the input of the next layer, and the final output is the embedded representation of each node: ; S33. Based on historical data, the cross entropy loss function and the Adam gradient descent algorithm are used to perform model training and parameter optimization on the multi-level dynamic evaluation model to obtain a trained multi-level dynamic evaluation model; S34. Input real-time data, perform dynamic evaluation through the trained multi-level dynamic evaluation model, and generate the ability score and performance evaluation of the operation and maintenance personnel. That is, input the feature data collected in real time into the trained multi-level dynamic evaluation model to generate the embedded representation of the node.

[0012] As a further technical solution of the present invention, the specific process of step S4 is: S41. According to the operational needs and management priorities of different parking lots, the weights of various indicators are determined through the Analytic Hierarchy Process (AHP). Specifically, the relative importance of each indicator is calculated based on expert scores or historical data, and a judgment matrix B is constructed, where Indicator Relative to the indicator Importance: , Then calculate the maximum eigenvalue of the judgment matrix and its corresponding eigenvector , and normalize the feature vector Get the weight of each indicator : , ; S42: Perform a comprehensive capability assessment based on the output of the multi-level dynamic assessment model and the weights set in S41, and calculate the comprehensive score of each operation and maintenance personnel: ,in, Indicates The comprehensive score of each operation and maintenance personnel, Indicates The operation and maintenance personnel are The scores on the indicators, Indicates The weight of each indicator; S43. Comprehensive score of each operation and maintenance personnel And the individual scores of each indicator Output: Generate a detailed evaluation report based on the evaluation results. The evaluation report includes the comprehensive score, scores of each indicator, and weight distribution details for managers to review and analyze.

[0013] In a second aspect, the present invention provides a multi-level dynamic capability evaluation system for operation and maintenance personnel, comprising: The data collection module is used to collect multi-dimensional data such as repair processing time, equipment failure rate, and service user evaluation, and use a secure communication protocol (such as HTTPS) to transmit the collected data in real time to the central database for storage and processing; The data processing and stratification module is used to perform data cleaning, outlier detection and data standardization on the collected data in turn, and to stratify the indicators into intervals according to the preset standards; The multi-level dynamic evaluation module is used to build a multi-level dynamic evaluation model, train the multi-level dynamic evaluation model, and dynamically evaluate the operation and maintenance personnel based on the trained multi-level dynamic evaluation model; The differentiated weight allocation and evaluation module is used to set the weights of various indicators according to the operating modes of different parking lots, and conduct a comprehensive evaluation based on the evaluation results of the multi-level dynamic evaluation module.

[0014] Compared with the prior art, the present invention has the following advantages: 1. Strong real-time performance. Through the data collection module, various operation and maintenance data can be obtained in real time and accurately, and dynamically updated and evaluated, so that managers can promptly understand the latest performance of operation and maintenance personnel and quickly adjust management strategies.

[0015] 2. Comprehensive evaluation: The system combines multi-dimensional data (such as repair processing time, equipment failure rate and service user evaluation) and conducts comprehensive analysis through a multi-level dynamic evaluation model to provide multi-level capability evaluation, which fully reflects the comprehensive capabilities and work performance of operation and maintenance personnel.

[0016] 3. High-precision outlier detection: Use machine learning algorithms (such as the isolation forest algorithm) to detect outliers in data, improve data quality and evaluation accuracy, and be able to more accurately identify and process abnormal data to ensure data reliability.

[0017] 4. Capturing dynamic behavior patterns: Graph neural networks can capture the dynamic interactions of operation and maintenance personnel at different times and places, revealing their behavior patterns and trends. This ability to capture dynamic behavior patterns makes the evaluation results more accurate and in-depth, reflecting the true performance of operation and maintenance personnel in complex environments.

[0018] 5. Differentiated weight allocation, using the analytic hierarchy process (AHP) to set the weights of various indicators, can be differentiated according to the operational needs of different parking lots, ensuring the accuracy and pertinence of the evaluation results. This can provide more personalized evaluation results for different operating models and management needs. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 A flow chart of a multi-level dynamic capability assessment method for operation and maintenance personnel provided by the present invention.

[0020] Figure 2 This is a structural block diagram of a multi-level dynamic capability evaluation system for operation and maintenance personnel provided by the present invention. DETAILED DESCRIPTION

[0021] It should be noted that the following detailed descriptions are exemplary and are intended to provide further explanation of the present invention. Unless otherwise specified, all technical and scientific terms used herein have the same meanings as those commonly understood by those skilled in the art to which the present invention belongs.

[0022] It should be noted that the terms used herein are only for describing specific embodiments, and are not intended to limit exemplary embodiments according to the present invention. As used herein, unless the context clearly indicates otherwise, the singular form is also intended to include the plural form. In addition, it should be understood that the terms "include" and "have" and any of their variations are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device comprising a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0023] In the absence of conflict, the embodiments of the present invention and the features of the embodiments may be combined with each other.

[0024] Embodiment 1: like Figure 1 As shown, this embodiment provides a multi-level dynamic capability assessment method for operation and maintenance personnel, which specifically includes the following steps: S1. Data collection: Collect multi-dimensional data on repair processing time, equipment failure rate, and service user evaluation, and use secure communication protocols (such as HTTPS) to transmit the collected data in real time to the central database for storage and processing. The repair processing time is automatically obtained through the API interface of the integrated repair system; the equipment failure rate is obtained through sensors and IoT devices, and the service user evaluation is obtained through the user feedback system or application.

[0025] S2. Data processing and stratification: The collected data is cleaned, outlier detected and standardized in turn, and each indicator is stratified according to the preset standards; the specific process is as follows: S21, clean the data collected in step S1 to remove noise data and outliers. Data cleaning includes missing value processing. The missing value processing adopts the mean filling method. ,in is the missing value neighboring data, is the number of adjacent data of the missing value; S22. Use the Isolation Forest algorithm to detect outliers. The Isolation Forest algorithm builds a tree by randomly selecting a subset of the data and randomly selecting features, and detects outliers by path length. The anomaly score , in, It’s data The average path length in the isolation tree, is the adjustment factor; S23, perform data standardization processing to convert the data into a standard normal distribution with a mean of 0 and a standard deviation of 1; S24, perform interval stratification, divide each indicator into different levels according to historical data and industry standards, and use K-means clustering algorithm to determine the boundaries of each interval. ,in is a cluster set, For the Cluster The center of mass.

[0026] S3. Multi-level dynamic evaluation: A multi-level dynamic evaluation model is constructed based on graph neural networks (GCNs), and the multi-level dynamic evaluation model is trained. The operation and maintenance personnel are dynamically evaluated according to the trained multi-level dynamic evaluation model. The specific process is as follows: S31. Construct a graph model that reflects the relationship between operation and maintenance personnel, parking lots, and equipment, and map the interactive relationship between operation and maintenance personnel at different times and locations into a graph structure. Specifically, define the operation and maintenance personnel, parking lots, and equipment as nodes in the graph, and define the relationship between nodes (such as the equipment and parking lots that the operation and maintenance personnel are responsible for) as edges. Construct an adjacency matrix A to represent the relationship between nodes in the graph, and construct a feature matrix X to represent the feature vector of each node. The feature vector of the node includes historical performance data, equipment status data, and vehicle factory operation data. S32. Use graph neural networks (GCNs) to process the constructed graph model and design a graph convolution layer to capture high-order dependencies and dynamic interactions between nodes to obtain a multi-level dynamic evaluation model. The graph convolution layer propagates information through the product of the adjacency matrix and the feature matrix: ,in, , is the adjacency matrix after adding self-loops, is the identity matrix, yes The degree matrix of It is The feature matrix of the layer, It is The weight matrix of the layer, is the activation function ReLU; high-order features are extracted by stacking three layers of graph convolutional layers. The output of each layer is used as the input of the next layer, and the final output is the embedded representation of each node: ; S33. Based on historical data, the cross entropy loss function and the Adam gradient descent algorithm are used to perform model training and parameter optimization on the multi-level dynamic evaluation model to obtain a trained multi-level dynamic evaluation model; S34. Input real-time data, perform dynamic evaluation through the trained multi-level dynamic evaluation model, and generate the ability score and performance evaluation of the operation and maintenance personnel. That is, input the feature data collected in real time into the trained multi-level dynamic evaluation model to generate the embedded representation of the node.

[0027] S4. Differentiated weight allocation and evaluation: according to the operation mode of different parking lots, the weights of various indicators are set, and combined with the evaluation results of step S3, a comprehensive evaluation is carried out. The specific process is as follows: S41. According to the operational needs and management priorities of different parking lots, the weights of various indicators are determined through the Analytic Hierarchy Process (AHP). Specifically, the relative importance of each indicator is calculated based on expert scores or historical data, and a judgment matrix B is constructed, where Indicators Relative to the indicator Importance: , Then calculate the maximum eigenvalue of the judgment matrix and its corresponding eigenvector , and normalize the feature vector Get the weight of each indicator : , ; S42: Perform a comprehensive capability assessment based on the output of the multi-level dynamic assessment model and the weights set in S41, and calculate the comprehensive score of each operation and maintenance personnel: ,in, Indicates The comprehensive score of each operation and maintenance personnel, Indicates The operation and maintenance personnel are The scores on the indicators, Indicates The weight of each indicator; S43. Comprehensive score of each operation and maintenance personnel And the individual scores of each indicator Output: Generate a detailed evaluation report based on the evaluation results. The evaluation report includes the comprehensive score, scores of each indicator, and weight distribution details for managers to review and analyze.

[0028] This embodiment constructs a graph model that reflects the relationship between operation and maintenance personnel, parking lots, and equipment, and uses graph neural networks to accurately capture the dynamic interactions of operation and maintenance personnel at specific times and places, thereby revealing the behavioral patterns and trends of the operation and maintenance personnel group. Managers can understand the work performance of operation and maintenance personnel in real time and dynamically, adjust management strategies in a timely manner, and improve the overall operation and maintenance management efficiency. Accurate capability assessment helps to identify efficient and excellent operation and maintenance personnel, and conduct targeted training and incentives. At the same time, it can identify problematic operation and maintenance personnel and weak links in their work, and make timely improvements, thereby improving service quality and customer satisfaction. In the context of intelligent parking lot management, this method not only has significant commercial value, but also provides strong technical support for the development of intelligent transportation and smart cities.

[0029] Embodiment 2: like Figure 2 As shown, this embodiment provides a multi-level dynamic capability evaluation system for operation and maintenance personnel, including: The data collection module is used to collect multi-dimensional data such as repair processing time, equipment failure rate, and service user evaluation, and use a secure communication protocol (such as HTTPS) to transmit the collected data in real time to the central database for storage and processing; The data processing and stratification module is used to perform data cleaning, outlier detection and data standardization on the collected data in turn, and to stratify the indicators into intervals according to the preset standards; The multi-level dynamic evaluation module is used to build a multi-level dynamic evaluation model, train the multi-level dynamic evaluation model, and dynamically evaluate the operation and maintenance personnel based on the trained multi-level dynamic evaluation model; The differentiated weight allocation and evaluation module is used to set the weights of various indicators according to the operating modes of different parking lots, and conduct a comprehensive evaluation based on the evaluation results of the multi-level dynamic evaluation module.

[0030] In further embodiments, there is also provided: An electronic device includes a memory and a processor, and is equipped with computer instructions running on the processor. When these instructions are executed by the processor, the method steps described in the embodiment can be accurately implemented. For the sake of simplicity, the detailed method content will not be repeated here.

[0031] It should be clear that the processor in this embodiment is not limited to the central processing unit CPU, it can also be a general-purpose processor core, a digital signal processor DSP, an application-specific integrated circuit ASIC, a field programmable gate array FPGA or other programmable logic device PLD, a discrete gate circuit or transistor logic component, a discrete hardware module, etc. In addition, the general-purpose processor can be a microprocessor MPU, or any conventional computing core.

[0032] The memory may include a read-only memory and a random access memory, and provide instructions and data to the processor. A portion of the memory may also include a non-volatile random access memory. For example, the memory may also store information about the device type.

[0033] A computer-readable storage medium is used to store computer instructions. When the computer instructions are executed by a processor, the method described in the embodiment is completed.

[0034] The method in the embodiment can be directly embodied as being executed by a hardware processor, or being executed by a combination of hardware and software modules in the processor. The software module can be located in a mature storage medium in the art such as a random access memory, a flash memory, a read-only memory, a programmable read-only memory, or an electrically erasable programmable memory, a register, etc. The storage medium is located in a memory, and the processor reads the information in the memory and completes the steps of the above method in combination with its hardware. To avoid repetition, it will not be described in detail here.

[0035] The algorithms, network structures, integrated repair reporting systems, sensors, Internet of Things devices, user feedback systems, and applications not described in detail in this embodiment are all existing technologies or general technologies in the field.

[0036] Those skilled in the art will appreciate that the units, i.e., algorithm steps, of the various examples described in conjunction with this embodiment can be implemented in electronic hardware or in a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0037] The above are only preferred embodiments of the present invention and are not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

[0038] Although the above describes the specific implementation mode of the present invention in conjunction with the accompanying drawings, it is not intended to limit the scope of protection of the present invention. Those skilled in the art should understand that various modifications or variations that can be made by those skilled in the art on the basis of the technical solution of the present invention without creative work are still within the scope of protection of the present invention.

Claims

1. A multi-level dynamic capability assessment method for operation and maintenance personnel, characterized in that: The specific steps include: S1. Data collection: Collect multi-dimensional data such as repair processing time, equipment failure rate and service user evaluation, and use secure communication protocols to transmit the collected data to the central database in real time for storage and processing; S2. Data processing and stratification: The collected data is cleaned, outlier detected and standardized in turn, and each indicator is stratified according to the preset standards; S3. Multi-level dynamic evaluation: Build a multi-level dynamic evaluation model based on graph neural network, train the multi-level dynamic evaluation model, and dynamically evaluate the operation and maintenance personnel according to the trained multi-level dynamic evaluation model; S4. Differentiated weight allocation and evaluation: according to the operation modes of different parking lots, the weights of various indicators are set, and a comprehensive evaluation is conducted in combination with the evaluation results of step S3.

2. The multi-level dynamic capability evaluation method for operation and maintenance personnel according to claim 1 is characterized in that: The repair processing time described in step S1 is automatically obtained through the API interface of the integrated repair system; the equipment failure rate is obtained through sensors and IoT devices, and the service user evaluation is obtained through the user feedback system or application.

3. The multi-level dynamic capability evaluation method for operation and maintenance personnel according to claim 2 is characterized in that: The specific process of step S2 is: S21, clean the data collected in step S1 to remove noise data and outliers. Data cleaning includes missing value processing. The missing value processing adopts the mean filling method. ,in is the missing value neighboring data, is the number of adjacent data of the missing value; S22. Use the isolation forest algorithm to detect outliers. The isolation forest constructs a tree by randomly selecting a subset of the data and randomly selecting features, and detects outliers by path length, anomaly score , in, It’s data The average path length in the isolation tree, is the adjustment factor; S23, perform data standardization processing to convert the data into a standard normal distribution with a mean of 0 and a standard deviation of 1; S24, perform interval stratification, divide each indicator into different levels according to historical data and industry standards, and use K-means clustering algorithm to determine the boundaries of each interval. ,in is a cluster set, For the Cluster The center of mass.

4. The multi-level dynamic capability evaluation method for operation and maintenance personnel according to claim 3 is characterized in that: The specific process of step S3 is: S31. Construct a graph model that reflects the relationship between operation and maintenance personnel, depots, and equipment, and map the interactive relationship between operation and maintenance personnel at different times and locations into a graph structure. Specifically, define the operation and maintenance personnel, depots, and equipment as nodes in the graph, define the relationship between nodes as edges, construct an adjacency matrix A to represent the relationship between nodes in the graph, and construct a feature matrix X to represent the feature vector of each node. The feature vector of the node includes historical performance data, equipment status data, and depot operation data. S32. Use graph neural network to process the constructed graph model, design graph convolution layer, capture high-order dependencies and dynamic interactions between nodes, and obtain a multi-level dynamic evaluation model. The graph convolution layer propagates information through the product of adjacency matrix and feature matrix: ,in, , is the adjacency matrix after adding self-loops, is the identity matrix, yes The degree matrix of It is The feature matrix of the layer, It is The weight matrix of the layer, is the activation function ReLU; high-order features are extracted by stacking three layers of graph convolutional layers. The output of each layer is used as the input of the next layer, and the final output is the embedded representation of each node: ; S33. Based on historical data, the cross entropy loss function and the Adam gradient descent algorithm are used to perform model training and parameter optimization on the multi-level dynamic evaluation model to obtain a trained multi-level dynamic evaluation model; S34. Input real-time data, perform dynamic evaluation through the trained multi-level dynamic evaluation model, and generate the ability score and performance evaluation of the operation and maintenance personnel. That is, input the feature data collected in real time into the trained multi-level dynamic evaluation model to generate the embedded representation of the node.

5. The multi-level dynamic capability evaluation method for operation and maintenance personnel according to claim 4 is characterized in that: The specific process of step S4 is: S41. According to the operational needs and management priorities of different parking lots, the weights of various indicators are determined by the hierarchical analysis method. Specifically, the relative importance of each indicator is calculated based on expert scores or historical data, and a judgment matrix B is constructed, where Indicators Relative to the indicator Importance: ， Then calculate the maximum eigenvalue of the judgment matrix and its corresponding eigenvector , and normalize the feature vector Get the weight of each indicator : , ; S42: Perform a comprehensive capability assessment based on the output of the multi-level dynamic assessment model and the weights set in S41, and calculate the comprehensive score of each operation and maintenance personnel: ,in, Indicates The comprehensive score of each operation and maintenance personnel, Indicates The operation and maintenance personnel are The scores on the indicators, Indicates The weight of each indicator; S43. Comprehensive score of each operation and maintenance personnel And the individual scores of each indicator Output: Generate a detailed evaluation report based on the evaluation results. The evaluation report includes the comprehensive score, scores of each indicator, and weight distribution details for managers to review and analyze.

6. A multi-level dynamic capability evaluation system for operation and maintenance personnel, characterized in that: The method according to any one of claims 1 to 5 is performed, comprising: The data collection module is used to collect multi-dimensional data such as repair processing time, equipment failure rate and service user evaluation, and use a secure communication protocol to transmit the collected data in real time to the central database for storage and processing; The data processing and stratification module is used to perform data cleaning, outlier detection and data standardization on the collected data in turn, and to stratify the indicators into intervals according to the preset standards; The multi-level dynamic evaluation module is used to build a multi-level dynamic evaluation model, train the multi-level dynamic evaluation model, and dynamically evaluate the operation and maintenance personnel based on the trained multi-level dynamic evaluation model; The differentiated weight allocation and evaluation module is used to set the weights of various indicators according to the operating modes of different parking lots, and conduct a comprehensive evaluation based on the evaluation results of the multi-level dynamic evaluation module.

Images