Method and device for repairing wireless network based on multiple agents, and electronic equipment

Abstract

The invention discloses a method and a device for repairing a wireless network based on multiple agents, and electronic equipment. The method comprises the following steps: receiving target data; a corresponding first intelligent agent is matched from the multiple intelligent agents according to the data type of target data, fault analysis is conducted on the target data through the first intelligent agent, an analysis result is obtained, and the analysis result is at least used for indicating the reason for abnormity of the target data; the analysis result serves as input, a wireless network repairing scheme is generated through a second intelligent agent, and the wireless network repairing scheme is used for indicating a disposal scheme for repairing the wireless network. According to the method and the device, the technical problem that the urgent demand of a modern network on high availability and low operation and maintenance cost is difficult to meet due to the fact that network repair depending on manual intervention is lagged in response, inaccurate in positioning, high in cost and incomplete in coverage is solved.

Description

Technical Field

[0001] This application relates to the field of mobile network communication, and more specifically, to a method, apparatus, and electronic device for repairing a wireless network based on multiple agents. Background Technology

[0002] As wireless networks grow increasingly large and complex in topology, manual troubleshooting and repair methods face severe challenges such as delayed response, inaccurate location, high costs, and incomplete coverage, making it difficult to meet the urgent needs of modern networks for high availability and low maintenance costs. Traditional manual wireless network maintenance suffers from the following main problems:

[0003] 1. Current wireless networks are large in scale and complex in architecture. They are designed for co-construction and sharing by 4G, 5G and operators. They rely on manual fault repair, which takes a long time. There are many things to be investigated across operators and wireless / transmission domains, making it difficult to quickly and accurately locate the root cause of the fault.

[0004] 2. Manual monitoring requires a large number of highly skilled engineers to work in shifts. Engineers' time is occupied by a large number of inefficient and repetitive troubleshooting tasks, making it impossible for them to focus on higher-value network optimization work.

[0005] 3. Manual troubleshooting is usually a "reactive response," making it difficult to conduct continuous and comprehensive health checks on the network. Some potential problems may go undetected before they develop into serious malfunctions.

[0006] There is currently no effective solution to the above problems. Summary of the Invention

[0007] This application provides a method, apparatus, and electronic device for repairing wireless networks based on multi-agent systems, which at least solves the technical problems of delayed response, inaccurate positioning, high cost, and incomplete coverage in network repair that rely on manual intervention, making it difficult to meet the urgent needs of modern networks for high availability and low maintenance costs.

[0008] According to one aspect of the embodiments of this application, a method for repairing a wireless network based on multiple agents is provided, comprising: receiving target data; matching a first agent from multiple agents according to the data type of the target data, and performing fault analysis on the target data through the first agent to obtain analysis results, wherein the analysis results are at least used to indicate the cause of the abnormality in the target data; and generating a wireless network repair scheme through a second agent using the analysis results as input, wherein the wireless network repair scheme is used to indicate a handling scheme for repairing the wireless network.

[0009] Optionally, a corresponding agent is matched from multiple agents based on the data type of the target data, and the target data is analyzed by the corresponding agent to obtain the analysis result. This includes: when the data type of the target data is alarm data, the alarm analysis agent is identified as the first agent; the alarm analysis agent performs alarm analysis on the alarm data to obtain a first analysis result, which includes alarm root cause location information and suggested repair solutions; historical alarm information and historical fault information of the network element to which the alarm data belongs are obtained, and a second analysis result is determined based on the historical alarm information and historical fault information, wherein the second analysis result is used to quantify the health of the network element to which the alarm data belongs; and the set of the first analysis result and the second analysis result is determined as the analysis result.

[0010] Optionally, a corresponding agent is matched from multiple agents based on the data type of the target data, and the target data is analyzed by the corresponding agent to obtain the analysis result. This includes: if the data type of the target data is performance index data, the performance analysis agent is identified as the first agent; the performance analysis agent performs cause analysis on the abnormal performance index data in the performance index data to obtain a third analysis result, which at least indicates the cause of the abnormal performance index data in the target cell, wherein the target cell is the cell with network fault to which the abnormal performance index data belongs; the performance analysis agent predicts the performance index data to obtain a prediction result, and the fault repair probability of the target cell is determined based on the prediction result; the third analysis result and the fault repair probability of the target cell are determined as the analysis result.

[0011] Optionally, using the analysis results as input, a wireless network repair plan is generated by a second intelligent agent, including: determining the neighboring cells and uplink bearer devices of the target cell from the cross-domain network topology map based on the analysis results, wherein the target cell is the cell with network faults indicated in the analysis results; obtaining the status information corresponding to the neighboring cells and uplink bearer devices respectively; and determining the wireless network repair plan based on the analysis results and the status information corresponding to the neighboring cells and uplink bearer devices respectively, wherein the wireless network repair plan is divided into two categories: remote repair plan or local solution. The local solution includes a repair procedure list for guiding maintenance personnel to perform on-site repairs, and the remote repair plan includes a repair script for performing network repairs through a third intelligent agent.

[0012] Optionally, based on the analysis results and the status information corresponding to neighboring cells and uplink bearer equipment, a wireless network repair plan is determined, including: determining the analysis results and the status information corresponding to neighboring cells and uplink bearer equipment as target conditions; matching the solutions corresponding to the target conditions in the wireless network fault handling knowledge base through the search enhancement generation search engine of the second intelligent agent, and determining the wireless network repair plan based on the solutions. The wireless network fault handling knowledge base is a database consisting of historical fault cases corresponding to the wireless network, network equipment operation manuals, network maintenance procedures, fault handling manuals, and network parameter configuration guidelines.

[0013] Optionally, the method further includes: receiving a remote repair plan through a third intelligent agent; performing wireless network repair by having the third intelligent agent send the command of the repair script in the remote repair plan to the network management of the operation and maintenance center; receiving the execution result corresponding to the remote repair plan from the network management of the operation and maintenance center through the third intelligent agent, and sending the execution result to the second intelligent agent, wherein the execution result is at least used to guide the second intelligent agent to perform incremental updates.

[0014] Optionally, the method further includes: receiving user natural language information uploaded in response to user input instructions through a fourth intelligent agent; performing intent recognition on the user natural language information through the fourth intelligent agent to obtain user intent, wherein the user intent is at least used to indicate the problem to be solved by the user; obtaining data corresponding to the user intent through the fourth intelligent agent, and calling the intelligent agent corresponding to the data corresponding to the user intent to process the data corresponding to the user intent, obtaining feedback results corresponding to the user intent, and pushing the feedback results to the user terminal.

[0015] According to another aspect of the embodiments of this application, a repair device for a wireless network based on multiple agents is also provided, comprising: a receiving module for receiving target data; an analysis module for matching a corresponding first agent from multiple agents according to the data type of the target data, and performing fault analysis on the target data through the first agent to obtain analysis results, wherein the analysis results are at least used to indicate the cause of the abnormality in the target data; and a generation module for generating a wireless network repair scheme through a second agent using the analysis results as input, wherein the wireless network repair scheme is used to indicate a handling scheme for repairing the wireless network.

[0016] According to another aspect of the embodiments of this application, a non-volatile storage medium is also provided, wherein a program is stored in the non-volatile storage medium, and the program controls the device where the non-volatile storage medium is located to execute the above-mentioned multi-agent wireless network repair method when it runs.

[0017] According to another aspect of the embodiments of this application, an electronic device is also provided, including: a memory and a processor, the processor being configured to run a program stored in the memory, wherein the program executes the above-described multi-agent-based wireless network repair method during runtime.

[0018] According to another aspect of the embodiments of this application, a computer program product is also provided, including computer instructions that, when executed by a processor, implement the above-described multi-agent-based wireless network repair method.

[0019] In this embodiment, the following steps are taken: receiving target data; matching the corresponding first agent from multiple agents based on the data type of the target data; and performing fault analysis on the target data through the first agent to obtain analysis results, wherein the analysis results are at least used to indicate the cause of the target data anomaly; using the analysis results as input, generating a wireless network repair scheme through a second agent, wherein the wireless network repair scheme is used to indicate the method of handling the repair of the wireless network. By performing fault analysis on the target data through the first agent to obtain analysis results, and then using the analysis results as input, generating a wireless network repair scheme through the second agent, the goal of accurately determining the wireless network repair strategy without relying on manual intervention is achieved. This solves the technical problems of delayed response, inaccurate positioning, high cost, and incomplete coverage in network repair that rely on manual intervention, making it difficult to meet the urgent needs of modern networks for high availability and low maintenance costs. Attached Figure Description

[0020] The accompanying drawings, which are included to provide a further understanding of this application and form part of this application, illustrate exemplary embodiments and are used to explain this application, but do not constitute an undue limitation of this application. In the drawings:

[0021] Figure 1 This is a hardware structure block diagram of a computer terminal for implementing a multi-agent-based wireless network repair method according to an embodiment of this application;

[0022] Figure 2 This is a flowchart of a repair method for a multi-agent wireless network provided according to an embodiment of this application;

[0023] Figure 3 This is a schematic diagram of the architecture of a multi-agent wireless network repair system provided in an embodiment of this application;

[0024] Figure 4 This is an interactive schematic diagram of a multi-agent wireless network repair system provided according to an embodiment of this application;

[0025] Figure 5This is a schematic diagram of a repair device for a multi-agent wireless network provided in an embodiment of this application. Detailed Implementation

[0026] To enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present application, and not all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative effort should fall within the scope of protection of the present application.

[0027] The information collected in this application embodiment is information and data authorized by the user or fully authorized by all parties. The collection, storage, use, processing, transmission, provision, disclosure and application of the relevant data all comply with the relevant laws, regulations and standards of the relevant regions, and necessary confidentiality measures have been taken. It does not violate public order and good morals, and provides corresponding operation entry points for users to choose to authorize or reject the automated decision results. If the user chooses to reject, the process will proceed to the expert decision-making process.

[0028] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this application described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0029] To better understand the embodiments of this application, the technical terms involved in the embodiments of this application are explained below:

[0030] Multi-Agent System (MAS): A multi-agent system is a computer system composed of multiple agents that can perform tasks independently while also communicating and cooperating with each other to complete complex tasks or solve problems.

[0031] Retrieval-Augmented Generation (RAG) is a model that combines retrieval and generation techniques. It retrieves relevant information from a large document collection before generating output, then uses this information to enrich the generated content, improving the quality and relevance of the generated text.

[0032] Long Short-Term Memory (LSTM) networks are a special type of recurrent neural network (RNN) particularly well-suited for processing and predicting important events with very long intervals and delays in time series. LSTMs effectively overcome the vanishing gradient problem in RNNs by introducing a gating mechanism, enabling the network to remember information for longer periods.

[0033] Operations and Maintenance Center (OMC): The OMC is a critical management system primarily used to monitor, manage, and maintain the operational status of network equipment. It provides a range of tools and functions to help network operators perform tasks such as network monitoring, fault diagnosis, performance analysis, and configuration management, ensuring network stability and quality of service.

[0034] Key Performance Indicators (KPIs): KPIs are metrics used to measure the performance of an organization, process, or individual, particularly when quantifying objectives. In network operations, KPIs can include network latency, packet loss rate, connection success rate, throughput, etc., used to assess network health and service quality.

[0035] In related technologies, fault diagnosis and repair modes relying on manual intervention face severe challenges such as delayed response, inaccurate location, high cost, and incomplete coverage, making it difficult to meet the urgent needs of modern networks for high availability and low maintenance costs. Therefore, there is a technical problem with network repair relying on manual intervention, characterized by delayed response, inaccurate location, high cost, and incomplete coverage, failing to meet the urgent needs of modern networks for high availability and low maintenance costs. To address this problem, this application provides relevant solutions in its embodiments, which are detailed below.

[0036] According to an embodiment of this application, an embodiment of a repair method for a wireless network based on multiple agents is provided. It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions. Furthermore, although a logical order is shown in the flowchart, in some cases, the steps shown or described may be executed in a different order than that shown here.

[0037] The methods and embodiments provided in this application can be executed on a computer terminal or similar computing device. Figure 1 A hardware block diagram of a computer terminal for implementing a repair method for a multi-agent wireless network is shown. Figure 1 As shown, the computer terminal 10 may include one or more processors 102 (shown as 102a, 102b, ..., 102n in the figure) 102 (processor 102 may include, but is not limited to, a microprocessor MCU or a programmable logic device FPGA, etc.), a memory 104 for storing data, and a transmission device 106 for communication functions. In addition, it may also include: a display, an input / output interface (I / O interface), a universal serial bus (USB) port (which may be included as one of the ports of a BUS bus), a network interface, a power supply, and / or a camera. Those skilled in the art will understand that... Figure 1 The structure shown is for illustrative purposes only and does not limit the structure of the aforementioned electronic device. For example, computer terminal 10 may also include... Figure 1 The more or fewer components shown, or having the same Figure 1 The different configurations shown.

[0038] It should be noted that the aforementioned one or more processors 102 and / or other data processing circuits are generally referred to herein as "data processing circuits". These data processing circuits may be embodied, in whole or in part, in software, hardware, firmware, or any other combination thereof. Furthermore, the data processing circuits may be a single, independent processing module, or may be integrated, in whole or in part, into any other element within the computer terminal 10. As involved in the embodiments of this application, the data processing circuits serve as processor control (e.g., selection of a variable resistor termination path connected to an interface).

[0039] The memory 104 can be used to store software programs and modules of application software, such as the program instructions / data storage device corresponding to the multi-agent-based wireless network repair method in this embodiment of the application. The processor 102 executes various functional applications and data processing by running the software programs and modules stored in the memory 104, thereby realizing the above-mentioned multi-agent-based wireless network repair method. The memory 104 may include high-speed random access memory, and may also include non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some instances, the memory 104 may further include memory remotely located relative to the processor 102, and these remote memories can be connected to the computer terminal 10 via a network. Examples of the above-mentioned networks include, but are not limited to, the Internet, corporate intranets, local area networks, mobile communication networks, and combinations thereof.

[0040] The transmission device 106 is used to receive or send data via a network. Specific examples of the network described above may include a wireless network provided by the communication provider of the computer terminal 10. In one example, the transmission device 106 includes a Network Interface Controller (NIC), which can connect to other network devices via a base station to communicate with the Internet. In another example, the transmission device 106 may be a Radio Frequency (RF) module, used for wireless communication with the Internet.

[0041] The display can be, for example, a touchscreen liquid crystal display (LCD) that allows the user to interact with the user interface of the computer terminal 10.

[0042] Under the aforementioned operating environment, embodiments of this application provide a method for repairing a wireless network based on multiple agents, such as... Figure 2 The diagram shown is a flowchart of a multi-agent-based wireless network repair method according to an embodiment of this application, including:

[0043] Step S202: Receive target data.

[0044] In some embodiments of this application, the multi-agent wireless network repair method of steps S202-S206 can be executed through a multi-agent wireless network repair system. The multi-agent architecture of this system is divided into two layers: a perception layer and a decision layer. The perception layer includes an orchestration and classification agent (i.e., the fourth agent), an alarm analysis agent, a performance analysis agent, and a repair execution agent (the third agent). Deployed on the end network element device or OMC, the orchestration and classification agent (the fourth agent) serves as the starting point for data reception, used to receive external information and also to achieve real-time monitoring of the device. Therefore, in step S202, the orchestration and classification agent (i.e., the fourth agent) receives target data. The target data is a type of data source in the method of this application embodiment. The target data is structured data, specifically divided into two categories: alarm data and performance indicator data. The alarm data is alarm data automatically generated by the network element device or OMC, and the performance indicator data is KPI data automatically generated by the network element device or OMC.

[0045] Step S204: Match the first agent from multiple agents according to the data type of the target data, and perform fault analysis on the target data through the first agent to obtain the analysis results.

[0046] In the technical solution provided in step S204, the first analysis result is at least used to indicate the cause of the anomaly in the target data. There are multiple ways to match the corresponding intelligent agent according to the data type of the target data, and to perform fault analysis on the target data through the corresponding intelligent agent to obtain the analysis result. For example: when the data type of the target data is alarm data, the alarm analysis intelligent agent is determined as the first intelligent agent; the alarm analysis intelligent agent performs alarm analysis on the alarm data to obtain the first analysis result, which includes alarm root cause location information and suggested repair solutions; historical alarm information and historical fault information of the network element to which the alarm data belongs are obtained, and a second analysis result is determined based on the historical alarm information and historical fault information, wherein the second analysis result is used to quantify the health of the network element to which the alarm data belongs; the set of the first analysis result and the second analysis result is determined as the analysis result. When the target data is performance index data, the performance analysis agent is designated as the first agent. The performance analysis agent performs causal analysis on the abnormal performance index data, obtaining a third analysis result. This third analysis result at least indicates the cause of the abnormal performance index data in the target cell, where the target cell is the cell with a network fault to which the abnormal performance index data belongs. The performance analysis agent predicts the performance index data, obtaining a prediction result, and determines the fault repair probability of the target cell based on the prediction result. The third analysis result and the fault repair probability of the target cell are then used as the analysis result. The execution process of step S204 is explained in detail below.

[0047] After receiving the target data, the system first matches the corresponding first agent (alarm analysis agent or performance analysis agent) in the perception layer based on the data type of the target data: if the data type of the target data is alarm data, the alarm analysis agent is determined as the first agent; if the data type of the target data is performance indicator data, the performance analysis agent is determined as the first agent.

[0048] When the target data is alarm data, the orchestration and classification agent (i.e., the fourth agent) sends the alarm data to the alarm analysis agent. When the alarm analysis agent performs alarm analysis, since the fields of the alarm data are fixed, it first extracts the alarm information, extracting key information to determine the alarm device, network standard, alarm type, alarm device model, and operating environment. For example, the following is an example of alarm data: Vendor Name: Vendor 1, Network Type: 4G, Data Collection Time: Time A, Synchronization Status: 0 (0 indicates data is not synchronized), Alarm Title: Device Input Power Disconnected, Alarm Status: Status 1. The following alarm information can be extracted: At time A, the 4G network device manufactured by Vendor 1 reported an alarm titled "Input Power Disconnected". The next step is to perform alarm analysis on the alarm information. This analysis can be performed using the alarm analysis algorithm built into the alarm analysis agent: The alarm type is identified based on the alarm title field. Then, historical alarm cases corresponding to the alarm type are retrieved from the alarm database. Using these historical cases as a reference, and based on the alarm device model and operating environment indicated in the alarm information, root cause analysis is performed using a large model (e.g., domain LLM) + RAG knowledge base approach. This analysis determines whether the alarm is caused by a fault, software problem, or external factors. Pre-defined repair strategies (i.e., suggested repair solutions, which may include one or more pre-defined repair strategies from the pre-defined repair strategy library) are then selected based on the scenario and cause of the alarm. The first analysis result is obtained (alarm root cause location information (the root cause of the alarm refers to the fundamental reason for the alarm) and suggested repair solutions).

[0049] Furthermore, this alarm analysis agent is equipped with a network element health algorithm. Based on historical alarms and faults, it assesses the health of network elements and uses this assessment to determine the likelihood of future faults. It retrieves historical alarm and fault information for the network element to which the alarm data belongs from the OMC center. From this information, it extracts feature vectors influencing network element health (features include alarm frequency (e.g., the total number of alarms in the past 30 days), fault interval, repair time (average repair time for each fault in the past year), device age, and software version change records). A pre-trained random forest model based on the random forest algorithm is used. Taking the feature vectors influencing network element health as input, each tree in the random forest model predicts the input feature vectors. The final prediction is determined by a majority vote of all tree predictions; the health status level with the highest frequency is selected as the final prediction. The output network element health status is then categorized into low, medium, and high levels based on the probability of the output and the corresponding thresholds for low, medium, and high levels. Finally, the set of the first and second analysis results is determined as the analysis result.

[0050] When the target data is performance index data, the performance analysis agent is designated as the first agent, and the orchestration and classification agent (i.e., the fourth agent) sends the performance index data to the performance analysis agent. The performance index data specifically includes KPI data (e.g., load (including traffic, number of connected users, Physical Resource Block (PRB) utilization, etc.), user uplink and downlink experience rates, cell load indicators, cell coverage indicators, interference indicators, etc.). The performance analysis agent is mainly used to analyze the KPI performance of the device, primarily performing preliminary analysis and root cause localization of abnormal KPI fluctuations in a single cell (the target cell, which is the cell with network faults to which the abnormal performance index data belongs). By associating multiple KPI relationships within the cell, preliminary causes are analyzed. Correlation analysis is performed on various data points in the performance index data to determine the reasons for the abnormal performance index data in the target cell. A multi-dimensional feature space is constructed, including user uplink and downlink experience rates, cell load indicators, cell coverage indicators, interference indicators, etc. For each KPI, the performance analysis agent extracts its time-series characteristics, statistics (such as mean and variance), and correlations with other KPIs. Multivariate statistical analysis methods or machine learning algorithms, such as multivariate parametric linear regression prediction models, are used to construct the relationships between KPIs. Model training is based on historical KPI datasets, with the goal of identifying patterns of mutual influence between KPIs.

[0051] The performance analysis agent also incorporates a lightweight LSTM prediction model to predict performance metrics (such as load, including traffic, number of connected users, PRB utilization, etc.) and obtain prediction results. These prediction results indicate the predicted performance metrics of the target cell over a predetermined time period (e.g., the next 72 hours). The predicted data includes at least the trend of load and other metrics during this period. The next step is to determine the fault recovery probability of the target cell based on the prediction results, judging whether the cell will automatically recover and whether there will be a sudden drop in performance. Based on the predicted performance metrics, combined with network configuration parameters, weather conditions, historical fault cases, and repair records, the performance analysis agent uses a pre-trained fault recovery probability model to assess the probability of the target cell automatically recovering from a fault within a certain timeframe. For example, suppose the performance analysis agent detects that a cell has recently experienced high load. Using the LSTM prediction model, it predicts that the load will remain high for the next 24 hours, but traffic will show a slight downward trend over the next 72 hours. The fault recovery probability model assessment concludes that the cell's load problem can be alleviated through load balancing adjustments. Finally, the third analysis result and the fault repair probability of the target cell were determined as the analysis results.

[0052] In step S204, different intelligent agents perform preliminary analysis on different types of target data to obtain analysis results. These preliminary analysis results significantly improve network availability and operation and maintenance efficiency, while reducing labor costs and resource waste. The next step is to pass the results to the second intelligent agent (decision intelligent agent) to generate the final wireless network repair solution based on the preliminary analysis results. This multi-level intelligent agent approach effectively solves the key technical problems of delayed response, inaccurate positioning, high cost, and incomplete coverage in network repair that rely on manual intervention.

[0053] Step S206: Using the analysis results as input, a wireless network repair plan is generated by the second intelligent agent, wherein the wireless network repair plan is used to indicate the handling plan for repairing the wireless network.

[0054] In the technical solution provided in step S206, there are multiple ways to generate a wireless network repair scheme through the second intelligent agent using the analysis results as input. For example: based on the analysis results, determine the neighboring cells and uplink bearer devices of the target cell from the cross-domain network topology map, where the target cell is the cell with network faults indicated in the analysis results (the cell with network faults indicated in the analysis results (e.g., the cell with network faults to which the abnormal performance index data belongs)); obtain the status information corresponding to the neighboring cells and uplink bearer devices respectively; based on the analysis results and the status information corresponding to the neighboring cells and uplink bearer devices respectively, determine the wireless network repair scheme, where the wireless network repair scheme is divided into two categories: remote repair scheme or near-end solution. The near-end solution includes a repair handling list for guiding maintenance personnel to perform on-site repairs, and the remote repair scheme includes a repair script for performing network repairs through a third intelligent agent.

[0055] Based on the analysis results and the status information of neighboring cells and uplink bearer devices, the above steps determine the implementation methods of the wireless network repair scheme in various ways. For example, the analysis results and the status information of neighboring cells and uplink bearer devices are determined as target conditions; the search engine generated by the second intelligent agent matches the solution corresponding to the target conditions in the wireless network fault handling knowledge base, and the wireless network repair scheme is determined based on the solution. The wireless network fault handling knowledge base is a database consisting of historical fault cases of wireless networks, network equipment operation manuals, network maintenance procedures, fault handling manuals, and network parameter configuration guidelines.

[0056] When the wireless network repair scheme is a remote repair scheme, the remote repair scheme is received by a third intelligent agent; the wireless network is repaired by sending the commands of the repair script in the remote repair scheme to the network management of the operation and maintenance center through the third intelligent agent; the execution result corresponding to the remote repair scheme is received by the network management of the operation and maintenance center through the third intelligent agent, and the execution result is sent to the second intelligent agent, wherein the execution result is at least used to guide the second intelligent agent to perform incremental updates.

[0057] In some embodiments of this application, the second intelligent agent is located in the decision layer of the multi-agent-based wireless network repair system. The decision layer mainly includes an intelligent decision agent (the second intelligent agent), deployed on a central cloud server. Using the analysis results as input (the target data can simultaneously include alarm data and performance indicator data; in this case, the input is a set of analysis results corresponding to both types of data. If the target data is either alarm data or performance indicator data, then the corresponding analysis result is used as input), the second intelligent agent determines the target cell with anomalies based on the analysis results, and identifies the neighboring cells and uplink bearer devices (A / B devices) of the target cell from the cross-domain network topology map. Device A refers to the first-level aggregation device in the bearer network, directly connected to the wireless base station, responsible for aggregating and forwarding the base station's data to higher-level network devices or the core network. Device A acts as a bridge between the base station and the bearer network in the communication network. Device B, as the second-level aggregation point of the bearer network, further aggregates and transmits the data to the core network. Device B is responsible for deeper data forwarding and processing in the bearer network, ensuring efficient data transmission. The target cell is the cell indicated in the analysis results as having network faults (i.e., the cell that generates alarm data or has abnormal performance indicator data). The next step is to send information to the perception layer to obtain the status information of neighboring cells and uplink bearer devices (obtaining real-time KPIs, alarm status, network configuration parameters, etc., to determine if there are any abnormalities in neighboring cells or A / B devices). The cross-domain network topology map is a complete cross-domain wireless-bearer network topology view maintained by the intelligent decision-making agent.

[0058] When determining a wireless network repair plan, the analysis results, along with the status information of neighboring cells and uplink bearer devices, are defined as target conditions (the analysis results and status information of neighboring cells and uplink bearer devices in the target conditions are structured information). For example, key information such as fault type, fault location, and impact range indicated in the analysis results are encoded into structured tuples or dictionaries for easy retrieval. Information such as the current KPI status, configuration details, alarm records, and correlation strength between neighboring cells and A / B devices and the target cell is converted into a retrieval-enhanced generative search engine (RAG search engine) readable format. The target conditions are a multidimensional dataset integrating fault conditions, network topology information, and device status, designed to provide comprehensive search criteria for the RAG search engine.

[0059] The wireless network fault handling knowledge base is a database consisting of historical fault cases, network equipment operation manuals, network maintenance procedures, fault handling manuals, and network parameter configuration guidelines for wireless networks. It requires pre-built indexes, including but not limited to multi-dimensional indexes based on fault type, equipment model, network standard, and repair strategy, to ensure the RAG engine can quickly locate relevant knowledge entries. A second-agent-based enhanced search engine matches solutions corresponding to target conditions within the wireless network fault handling knowledge base. Based on the target conditions, a search query is constructed, which includes, but is not limited to, the following elements: fault description, including fault type, specific manifestations, and impact scope; status information of neighboring cells and uplink bearer equipment, including KPI fluctuations, configuration details, and hardware status; and other related information from the fault analysis results, such as possible fault causes, fault frequency, and fault urgency.

[0060] The search query is input into the Retrieval Enhancement Generation (RAG) search engine, which uses a content-based retrieval mechanism to match solutions corresponding to the target conditions in the wireless network fault handling knowledge base (by calculating the similarity or correlation between the query and entries in the knowledge base to retrieve relevant solutions (using cosine similarity) as the solutions corresponding to the target conditions). The solution matching the target conditions in the wireless network fault handling knowledge base is the one with the highest cosine similarity. This solution is then used as a candidate repair strategy. Real-time network status data (device status, network configuration, resource availability, etc.) of the current target cell is obtained, and the candidate repair strategies are adjusted based on this data to obtain a wireless network repair solution. There are two types of wireless network repair solutions: remote repair solutions and local solutions. Local solutions include a repair and handling checklist to guide maintenance personnel in on-site repairs. Local solutions require maintenance personnel to go to the site for repairs and will include detailed repair and handling suggestions. The remote repair solution includes repair scripts for network repair via a third-party intelligent agent. The remote handling solution refers to a solution that eliminates faults by adjusting parameters in the background without going to the site. This type of solution will generate corresponding repair scripts and connect to the OMC southbound interface to realize automatic script execution.

[0061] For remote repair solutions, the third-party intelligent agent (repair execution agent) receives the remote repair plan and sends the commands of the repair script in the remote repair plan to the network management system of the operation and maintenance center for wireless network repair. The method of sending the commands of the repair script in the remote repair plan to the network management system of the operation and maintenance center involves the repair execution agent sending commands through the multi-connection control server (MCP server, or MCP for short) or API capabilities via the southbound interface of the OMC to repair the fault. After receiving the command, the OMC forwards it to the corresponding network device through the southbound interface. The device executes the command to perform repair operations such as parameter adjustment and configuration change, and after the repair operation is completed, it feeds back the execution result of the remote repair plan to the repair execution agent.

[0062] The third intelligent agent receives the execution results of the remote repair plan from the network management system of the operation and maintenance center, extracts key information, and sends the execution results to the second intelligent agent according to the second intelligent agent's requirements. This execution result is used at least for incremental updates by the second intelligent agent: the intelligent decision-making agent combines all data (analysis results, execution results, wireless network repair plans, etc.) to make a comprehensive judgment, which includes confirming the existence of alarm data, performance data, analysis results, execution results, and the wireless network repair plan (including suggestions on the execution order and resource allocation of the handling plan). Finally, the judgment result is pushed to the corresponding cloud-based scheduling system (or work order system). After receiving the relevant information, the scheduling system assigns the work order to the appropriate personnel. After completing the processing, the personnel provide feedback on the accuracy of the intelligent decision-making agent's handling plan. This feedback continuously optimizes the intelligent decision-making agent and guides its incremental updates.

[0063] In some embodiments of this application, in addition to receiving target data, it also supports customized and personalized user-generated questions, receiving natural language input in response to user input instructions, and answering user needs. For example, in addition to target data, it can also receive user natural language information uploaded in response to user input instructions through a fourth intelligent agent; perform intent recognition on the user natural language information through the fourth intelligent agent to obtain the user intent, wherein the user intent is at least used to indicate the problem to be solved by the user; obtain the data corresponding to the user intent through the fourth intelligent agent, and call the intelligent agent corresponding to the data corresponding to the user intent to process the data corresponding to the user intent, obtain the feedback result corresponding to the user intent, and push the feedback result to the user terminal. For example, the fourth intelligent agent (the orchestration and classification agent in the perception layer) performs intent recognition on the user's natural language information to obtain the user's intent. If the user's intent is to query alarms generated near a certain device and provide processing opinions, the corresponding data is alarm data. The fourth intelligent agent then obtains the data corresponding to the user's intent and sends the corresponding alarm data to the corresponding intelligent agents (alarm analysis agent and decision agent). The corresponding data is processed to obtain the feedback result corresponding to the user's intent, and the feedback result is pushed to the user terminal. The specific data processing methods of the alarm analysis agent and decision agent are described in steps S204-S206, and will not be repeated here.

[0064] Figure 3This is a schematic diagram of the architecture of a multi-agent wireless network repair system according to an embodiment of this application. It illustrates the architecture of the multi-agent wireless network repair system used to execute the multi-agent wireless network repair method in this embodiment. The system consists of two layers: a perception layer and a decision layer. The perception layer includes an orchestration and classification agent (i.e., the fourth agent mentioned above), an alarm analysis agent, a performance analysis agent, and a repair execution agent. It is deployed on the end network element device or OMC. The perception layer includes four types of agents, capable of semantic recognition, optimization problem investigation, maintenance problem investigation, and repair command execution. The orchestration and classification agent supports semantic recognition, fault analysis, performance data analysis, and MCP server (the main function of the MCP server in the orchestration and classification agent is to act as a communication bridge, responsible for interacting with the northbound interface of the OMC to obtain necessary network status, device information, alarm data, and performance indicator data). The Alarm Analysis Agent is used for root cause localization and determining network element health (i.e., obtaining historical alarm and fault information of the network element to which the alarm data belongs, and determining the second analysis result based on the historical alarm and fault information, whereby the second analysis result is used to quantify the health of the network element to which the alarm data belongs). The Performance Analysis Agent is used for root cause localization and LSTM prediction (i.e., predicting performance index data through the performance analysis agent, obtaining prediction results, and determining the fault repair probability of the target cell based on the prediction results). The Repair Execution Agent has southbound interface capabilities (i.e., the ability to execute repair commands), connects to the repair plan generated by the Intelligent Decision Agent, interacts directly with the OMC's southbound interface through the MCP server, and issues specific repair instructions to the faulty device to achieve remote repair.

[0065] The decision-making layer acts as the brain, analyzing neighboring cell conditions through the overall network topology, interacting with the real-time perception layer to obtain information about current or neighboring network elements, and obtaining root cause localization and remediation solutions through multi-dimensional comprehensive analysis. The decision-making layer includes an intelligent decision-making agent (i.e., the aforementioned second intelligent agent). The intelligent decision-making agent is responsible for maintaining the cross-domain network topology map and the wireless fault handling knowledge base. The intelligent decision-making agent integrates the analysis results from the perception layer agents with the status information of neighboring cells and uplink bearer devices. Through deep learning and retrieval-enhanced generation (RAG) technology, it customizes and distributes policies (i.e., the analysis results and the status information of neighboring cells and uplink bearer devices are determined as target conditions; the retrieval-enhanced generation search engine of the second intelligent agent matches solutions corresponding to the target conditions in the wireless network fault handling knowledge base, and determines the wireless network repair plan based on the solutions). The intelligent decision-making agent considers global impact and resource allocation to maximize repair efficiency and network stability, coordinating multiple agents, and distributing the generated policies to the perception layer for execution through the MCPSERVER (same as the MCP server).

[0066] Each agent in the perception layer specializes in a specific domain, with different agents equipped with different small analytical models, enabling them to achieve professional-level analysis in their respective vertical domains. The decision-making agent in the decision layer aggregates all information and makes the final decision. Agents at each layer have clearly defined responsibilities and interact and collaborate on tasks through predefined communication protocols. This architecture design implements a decision-making mechanism of "local self-intelligence, regional coordination, and global optimization." The perception layer can respond to local faults within minutes, significantly improving real-time processing. The decision layer aggregates local information to solve complex problems across device nodes while ensuring global policy consistency. This design effectively overcomes the single point of failure and response latency issues of traditional centralized systems, greatly improving the system's scalability, response speed, and overall robustness. Lightweight machine learning models are embedded in the alarm analysis agent and performance analysis agent in the perception layer. Device data is used at edge device nodes to train the models in each agent, improving the accuracy of small models in vertical domains and enabling collaboration between large and small models. The advantage of this design is that the intelligent agent can learn from historical fault handling and continuously optimize its self-healing strategy, which solves the problems of poor adaptability and lagging updates of traditional static rules or centralized AI models, and improves the intelligence level and long-term effectiveness of the self-healing strategy.

[0067] Figure 4 This is an interactive schematic diagram of a multi-agent wireless network-based repair system according to an embodiment of this application, illustrating... Figure 3The system described illustrates an interactive process: ① The orchestration and classification agent at the perception layer receives data (which can be natural language input from the user or structured data corresponding to base station fault events (i.e., the aforementioned target data)). ② The orchestration and classification agent performs semantic analysis on the user's natural language, identifies the user's intent, and retrieves the corresponding alarm data or raw KPI data from the network management OMC. ③ Based on the structured data input in ① and the structured data obtained from the OMC in ②, the orchestration and classification agent, if it is alarm data, calls the alarm analysis agent; if it is performance indicator data, it calls the performance analysis agent, performing single-node, single-domain root cause analysis respectively. ④ The alarm analysis agent and the performance analysis agent inform the intelligent decision-making agent at the decision layer of the analysis results for each single node. This includes, for example, root cause localization of single-node alarms, network element health analysis, root cause localization of performance indicator anomalies, and prediction of future performance indicator trends. The results are then transmitted to the intelligent decision-making agent via JSON. ⑤. When the intelligent decision-making agent in the decision-making layer analyzes neighboring node failures, KPI performance, or needs to perform data prediction, it will notify each agent in the perception layer to execute corresponding actions. When the intelligent decision-making agent generates a background repair plan or needs to schedule the OMC system to query more base station parameter configurations, it will generate relevant command scripts. These scripts will then be scheduled by the perception layer repair execution agent through process ⑥ to execute the script plan and wait for the results. ⑦. After receiving the execution script from the intelligent decision-making agent, the repair execution agent will send the command to the OMC network management system. After executing the command, the OMC network management system will provide feedback on the execution results. The repair execution agent will parse the execution results from the OMC, extract keywords, and provide feedback on the results according to the requirements of the intelligent decision-making agent in process ⑥. Process 8: After the intelligent decision-making agent combines all the data to make a comprehensive judgment, it will push the judgment result and other information to the corresponding cloud dispatch system (or work order system). After receiving the relevant information, the dispatch system will assign the task to the corresponding personnel. After completing the processing, the personnel will provide feedback on the accuracy of the intelligent decision-making agent's processing solution. This feedback will continuously optimize the intelligent decision-making agent. Process 9: If the user asks a question through dialogue, the analysis results of the faulty cell will be fed back to the user in natural language.

[0068] Figure 5 This is a schematic diagram of a repair device for a multi-agent wireless network according to an embodiment of this application, comprising:

[0069] The receiving module 502 is used to receive target data.

[0070] Analysis module 504 is used to match the corresponding first agent from multiple agents according to the data type of the target data, and to perform fault analysis on the target data through the first agent to obtain analysis results. The first analysis results are used to indicate at least the cause of the abnormality in the target data.

[0071] The analysis module 504 is further configured to: When the target data is alarm data, identify the alarm analysis agent as the first agent; perform alarm analysis on the alarm data using the alarm analysis agent to obtain a first analysis result, which includes alarm root cause location information and suggested repair solutions; acquire historical alarm information and historical fault information of the network element to which the alarm data belongs, and determine a second analysis result based on the historical alarm information and historical fault information, wherein the second analysis result is used to quantify the health of the network element to which the alarm data belongs; and determine the set of the first and second analysis results as the analysis result. When the target data is performance indicator data, identify the performance analysis agent as the first agent; perform cause analysis on abnormal performance indicator data in the performance indicator data using the performance analysis agent to obtain a third analysis result, which at least indicates the cause of abnormal performance indicator data in the target cell, wherein the target cell is the cell with network faults to which the abnormal performance indicator data belongs; predict the performance indicator data using the performance analysis agent to obtain a prediction result, and determine the fault repair probability of the target cell based on the prediction result; and determine the third analysis result and the fault repair probability of the target cell as the analysis result.

[0072] The generation module 506 is used to generate a wireless network repair plan through a second intelligent agent, taking the analysis results as input. The wireless network repair plan is used to indicate the handling method for repairing the wireless network.

[0073] The generation module 506 is also used to determine the neighboring cells and uplink bearer devices of the target cell from the cross-domain network topology map based on the analysis results. The target cell is the cell with network faults indicated in the analysis results. The module also obtains the status information corresponding to the neighboring cells and uplink bearer devices. Based on the analysis results and the status information corresponding to the neighboring cells and uplink bearer devices, the module determines a wireless network repair plan. The wireless network repair plan is divided into two categories: a remote repair plan or a local solution. The local solution includes a repair and handling list to guide maintenance personnel to perform on-site repairs. The remote repair plan includes a repair script for network repair through a third-party intelligent agent.

[0074] The generation module 506 is also used to determine the analysis results and the status information corresponding to the neighboring cells and the uplink bearer equipment as target conditions; through the retrieval enhancement of the second intelligent agent, the generation search engine matches the solution corresponding to the target conditions in the wireless network fault handling knowledge base, and determines the wireless network repair plan based on the solution. The wireless network fault handling knowledge base is a database consisting of historical fault cases corresponding to the wireless network, network equipment operation manuals, network maintenance procedures, fault handling manuals, and network parameter configuration guidelines.

[0075] The generation module 506 is also used to receive a remote repair plan through a third intelligent agent; to perform wireless network repair by sending the command of the repair script in the remote repair plan to the network management of the operation and maintenance center through the third intelligent agent; to receive the execution result corresponding to the remote repair plan from the network management of the operation and maintenance center through the third intelligent agent, and to send the execution result to the second intelligent agent, wherein the execution result is at least used to guide the second intelligent agent to perform incremental updates.

[0076] The generation module 506 is also used to receive user natural language information uploaded in response to user input instructions through a fourth intelligent agent; to perform intent recognition on the user natural language information through the fourth intelligent agent to obtain the user intent, wherein the user intent is at least used to indicate the problem to be solved by the user; to obtain the data corresponding to the user intent through the fourth intelligent agent, and to call the intelligent agent corresponding to the data corresponding to the user intent to process the data corresponding to the user intent, to obtain the feedback result corresponding to the user intent, and to push the feedback result to the user terminal.

[0077] It should be noted that, Figure 5 The repair device for the multi-agent wireless network shown is used to perform... Figure 2 The repair method for multi-agent-based wireless networks shown herein, therefore Figure 2 The relevant explanations in the multi-agent-based wireless network repair method also apply to the multi-agent-based wireless network repair device, and will not be repeated here.

[0078] It should be noted that each module in the above-mentioned multi-agent wireless network repair device can be a program module (e.g., a set of program instructions to implement a certain function) or a hardware module. For the latter, it can be manifested in the following forms, but is not limited to them: each of the above modules is manifested as a processor, or the functions of each of the above modules are implemented by a processor.

[0079] This application also provides a non-volatile storage medium, which includes a stored program, wherein, when the program is running, it controls the device where the non-volatile storage medium is located to execute the multi-agent-based wireless network repair method of any of the above embodiments.

[0080] This application also provides an electronic device, which includes a processor for running a program, wherein the program executes the multi-agent-based wireless network repair method of any of the above embodiments.

[0081] According to another aspect of the embodiments of this application, a computer program product is also provided, including a computer program that, when executed by a processor, implements the multi-agent-based wireless network repair method of any of the above embodiments.

[0082] In the above embodiments of this application, the descriptions of each embodiment have different focuses. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0083] In the several embodiments provided in this application, it should be understood that the disclosed technical content can be implemented in other ways. The device embodiments described above are merely illustrative; for example, the division of units can be a logical functional division, and in actual implementation, there may be other division methods. For instance, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the displayed or discussed mutual coupling, direct coupling, or communication connection may be through some interfaces; the indirect coupling or communication connection between units or modules may be electrical or other forms.

[0084] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0085] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0086] If the integrated unit is implemented as 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 this application, in essence, or the part that contributes to related technologies, or all or part of the 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 cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, read-only memory (ROM), random access memory (RAM), portable hard drives, magnetic disks, or optical disks.

[0087] The above description is only a preferred embodiment of this application. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principle of this application, and these improvements and modifications should also be considered within the scope of protection of this application.

Claims

1. A method for repairing a wireless network based on multi-agent technology, characterized in that, include: Receive target data; The first agent is matched from multiple agents according to the data type of the target data, and the first agent performs fault analysis on the target data to obtain analysis results, wherein the analysis results are at least used to indicate the cause of the abnormality in the target data; Using the analysis results as input, a wireless network repair plan is generated by a second intelligent agent, wherein the wireless network repair plan is used to indicate the handling method for repairing the wireless network.

2. The method according to claim 1, characterized in that, The process of matching the target data with a corresponding agent from multiple agents based on the data type, and then performing fault analysis on the target data using the corresponding agent to obtain analysis results includes: If the data type of the target data is alarm data, the alarm analysis agent is determined to be the first agent; The alarm analysis agent performs alarm analysis on the alarm data to obtain a first analysis result, which includes alarm root cause location information and suggested repair solutions. The system obtains historical alarm information and historical fault information of the network element to which the alarm data belongs, and determines a second analysis result based on the historical alarm information and historical fault information. The second analysis result is used to quantify the health of the network element to which the alarm data belongs. The set of the first analysis result and the second analysis result is defined as the analysis result.

3. The method according to claim 1, characterized in that, The process of matching the target data with a corresponding agent from multiple agents based on the data type, and then performing fault analysis on the target data using the corresponding agent to obtain analysis results includes: When the data type of the target data is performance index data, the performance analysis agent is determined to be the first agent; The performance analysis agent performs cause analysis on the abnormal performance index data in the performance index data to obtain a third analysis result. The third analysis result is used to at least indicate the cause of the abnormal performance index data in the target cell, wherein the target cell is the cell with network failure to which the abnormal performance index data belongs. The performance analysis agent predicts the performance index data to obtain prediction results, and determines the fault repair probability of the target cell based on the prediction results. The third analysis result and the fault repair probability of the target cell are determined as the analysis result.

4. The method according to claim 1, characterized in that, The step of generating a wireless network repair scheme through a second intelligent agent, using the analysis results as input, includes: Based on the analysis results, the neighboring cells and uplink bearer devices of the target cell are determined from the cross-domain network topology map, wherein the target cell is the cell with network faults indicated in the analysis results; Obtain the status information corresponding to the neighboring cells and the uplink bearer devices respectively; Based on the analysis results and the status information corresponding to the neighboring cells and the uplink bearer devices, the wireless network repair scheme is determined. The wireless network repair scheme is divided into two categories: remote repair scheme or local solution. The local solution includes a repair and handling list to guide maintenance personnel to perform on-site repairs, and the remote repair scheme includes a repair script for network repair through a third-party intelligent agent.

5. The method according to claim 4, characterized in that, Based on the analysis results and the status information corresponding to the neighboring cells and the uplink bearer devices, the wireless network repair scheme is determined, including: The analysis results, along with the status information corresponding to the neighboring cells and the uplink bearer equipment, are determined as the target conditions. The second intelligent agent enhances the search engine to match the solution corresponding to the target condition in the wireless network fault handling knowledge base, and determines the wireless network repair solution based on the solution. The wireless network fault handling knowledge base is a database consisting of historical fault cases, network equipment operation manuals, network maintenance procedures, fault handling manuals, and network parameter configuration guidelines corresponding to the wireless network.

6. The method according to claim 4, characterized in that, The method further includes: The remote repair plan is received through the third intelligent agent; Wireless network repair is performed by sending commands for the repair script in the remote repair scheme to the network management system of the operation and maintenance center through the third intelligent agent. The third agent receives the execution result corresponding to the remote repair scheme from the network management system of the operation and maintenance center, and sends the execution result to the second agent. The execution result is used to guide the second agent to perform incremental updates.

7. The method according to claim 1, characterized in that, The method further includes: The fourth intelligent agent receives user natural language information uploaded in response to user input commands; The fourth intelligent agent performs intent recognition on the user's natural language information to obtain the user's intent, wherein the user intent is at least used to indicate the problem to be solved by the user; The fourth agent obtains the data corresponding to the user's intent, calls the agent corresponding to the data corresponding to the user's intent to process the data corresponding to the user's intent, obtains the feedback result corresponding to the user's intent, and pushes the feedback result to the user's end.

8. A repair device based on a multi-agent wireless network, characterized in that, include: The receiving module is used to receive target data; An analysis module is used to match a first agent from multiple agents based on the data type of the target data, and to perform fault analysis on the target data through the first agent to obtain analysis results, wherein the analysis results are at least used to indicate the cause of the anomaly in the target data; A generation module is used to generate a wireless network repair plan through a second intelligent agent, using the analysis results as input. The wireless network repair plan is used to indicate a handling scheme for repairing the wireless network.

9. A non-volatile storage medium, characterized in that, The non-volatile storage medium stores a program, wherein when the program is executed, it controls the device containing the non-volatile storage medium to perform the repair method for a multi-agent wireless network as described in any one of claims 1 to 7.

10. An electronic device, characterized in that, include: A memory and a processor, the processor being configured to run a program stored in the memory, wherein the program, when running, executes the repair method for a multi-agent wireless network as described in any one of claims 1 to 7.

11. A computer program product comprising computer instructions, characterized in that, When the computer instructions are executed by the processor, they implement the repair method for a multi-agent wireless network as described in any one of claims 1 to 7.

Images