A communication device operation and maintenance management method and system

Abstract

The application provides a kind of communication equipment operation and maintenance management method and system, belong to intelligent operation and maintenance technical field, its method includes operation and maintenance management center receives operation and maintenance task, and the operation and maintenance task is modeled and analyzed, generates task execution atlas, obtains the resource demand attribute of operation and maintenance task in combination with the resource state information acquired in real time, dynamically decides the target execution position of operation and maintenance task;Operation and maintenance task is unloaded to corresponding target execution position, after execution is completed, draw key results, classify key results, and upload classification results asynchronously;According to the asynchronous upload result, generate operation and maintenance insight report and strategy optimization suggestion, close loop feedback and model optimization are carried out to operation and maintenance insight report and strategy optimization suggestion.The power resource cross-regional cross-level intelligent scheduling is realized, the task demand and resource state are dynamically perceived, the power mismatch problem is solved, the resource utilization is improved, the 5G / 6G edge application low latency requirement is met, and the power network operation and maintenance new paradigm is constructed.

Description

Technical Field

[0001] This invention relates to the field of intelligent operation and maintenance technology, and in particular to a method and system for operation and maintenance management of communication equipment. Background Technology

[0002] The national push for on-demand computing resources, similar to water and electricity, is a key technology for achieving intelligent scheduling and efficient collaboration of computing resources across regions and levels (edge-cloud), and is essential for building the operation and maintenance support system for computing networks. 5G / 6G networks have spurred the development of massive low-latency, high-bandwidth applications such as autonomous driving and VR / AR, which rely on edge computing. Traditional centralized cloud operations and maintenance (O&M) cannot meet these requirements, urgently necessitating a new O&M paradigm that can decentralize O&M capabilities to the edge and achieve global collaboration. Traditional O&M, based on static rules for scheduling, fails to perceive the actual needs of tasks and the real-time status of resources, leading to computing power mismatch, rigid scheduling, and low resource utilization.

[0003] Therefore, the present invention provides a method and system for operation and maintenance management of communication equipment. Summary of the Invention

[0004] This invention provides a communication equipment operation and maintenance management method and system. The operation and maintenance management center generates a task execution map through modeling and analysis. Based on the resource requirement attributes of the task execution map and real-time acquired resource status information, it intelligently schedules operation and maintenance tasks to the target execution location. After execution, key results are categorized and uploaded asynchronously, generating operation and maintenance insight reports and strategy optimization suggestions. Finally, model optimization is achieved through closed-loop feedback. This enables intelligent scheduling of computing resources across regions and levels, solves the problem of computing power mismatch by dynamically sensing task requirements and resource status, improves resource utilization, meets the low latency requirements of 5G / 6G edge applications, and constructs a new paradigm for computing power network operation and maintenance.

[0005] This invention provides a method for operation and maintenance management of communication equipment, comprising: Step 1: The operation and maintenance management center receives the operation and maintenance tasks, models and analyzes the operation and maintenance tasks, and generates a task execution graph; Step 2: Obtain the resource requirement attributes of the operation and maintenance task according to the task execution map. Based on the resource requirement attributes and the real-time resource status information, dynamically determine the target execution location of the operation and maintenance task through an intelligent scheduling algorithm. Step 3: Unload the operation and maintenance task to the corresponding target execution location. After execution, obtain the key results, classify the key results, and upload the classification results asynchronously. Step 4: Generate an operation and maintenance insight report and strategy optimization suggestions based on the asynchronous upload results, and perform closed-loop feedback and model optimization on the operation and maintenance insight report and strategy optimization suggestions.

[0006] This invention provides a communication equipment operation and maintenance management method, in which an operation and maintenance management center receives operation and maintenance tasks, performs modeling and analysis on the operation and maintenance tasks, and generates a task execution map, including: The operations and maintenance management center receives operations and maintenance tasks, parses the high-level business intent of the operations and maintenance tasks, and identifies the core objectives and task constraints of the operations and maintenance tasks. The high-level business intent is decomposed into a set of atomic task sequences that can be directly executed by the system, and the logical dependencies between the atomic task sequences are defined. Based on the core objectives and task constraints, determine the resource requirement attributes corresponding to each atomic task sequence, and form a resource requirement model for each atomic task sequence based on the resource requirement attributes. Based on the atomic task sequence, the corresponding resource requirement model, and logical dependencies, a task execution graph is constructed.

[0007] This invention provides a communication equipment operation and maintenance management method, which obtains the resource requirement attributes of the operation and maintenance tasks based on the task execution map, and dynamically determines the target execution location of the operation and maintenance tasks based on the resource requirement attributes and real-time acquired resource status information, using an intelligent scheduling algorithm. The method includes: The task execution graph is analyzed to extract the sub-requirements of each atomic task sequence. The sub-requirements are aggregated and calculated in combination with logical dependencies to obtain the resource requirement attributes of the operation and maintenance task. The resource requirement attributes are compared with the resource status information obtained in real time from the terminal side, edge side and cloud side to generate a set of candidate execution locations; Using the latency sensitivity and task priority in the resource demand attributes as the core optimization objectives and other resource demands as demand constraints, the resource matching matrix is ​​evaluated using an intelligent scheduling algorithm to solve for the Pareto optimal solution set. The resource matching matrix is ​​derived by using the candidate execution positions in the candidate execution position set as rows and the quantified resource requirement attributes and the real-time acquired resource status information as columns. From the Pareto optimal solution set, a final target execution location is selected according to a preset scheduling strategy.

[0008] This invention provides a communication equipment operation and maintenance management method, which selects a final target execution location from the Pareto optimal solution set according to a preset scheduling strategy, including: Based on the metadata of the operation and maintenance task, identify the tenant identity to which the operation and maintenance task belongs, and query the customized scheduling policy bound to the corresponding tenant identity. Using the customized scheduling strategy as the evaluation criterion, the priority of each candidate execution position in the Pareto optimal solution set is ranked. Select the candidate execution position that has the highest priority and use it as the target execution position.

[0009] This invention provides a communication equipment operation and maintenance management method, which offloads operation and maintenance tasks to corresponding target execution locations, obtains key results after execution, classifies the key results, and asynchronously uploads the classification results, including: Based on the scheduling decision, the operation and maintenance management center generates task execution instructions and unloads the operation and maintenance tasks and atomic task sequences to the corresponding target execution locations via the control channel; After receiving instructions, each target execution location loads the operation and maintenance task context and performs collaborative calculations based on the logical dependencies in the task execution graph to obtain key results. The key results obtained at each target execution location are divided into two categories: real-time control data and analysis and optimization data. The encapsulated real-time control data is synchronously fed back to the operation and maintenance management center through the control channel, while the encapsulated analysis and optimization data is uploaded to the cloud center asynchronously through the data channel as a structured dataset.

[0010] This invention provides a communication equipment operation and maintenance management method, which generates an operation and maintenance insight report and strategy optimization suggestions based on asynchronous upload results, and performs closed-loop feedback and model optimization on the operation and maintenance insight report and strategy optimization suggestions, including: Draw a global data view based on the asynchronous upload result set; Based on the predefined operation and maintenance meta-model, the entities and their corresponding relationships in the global data view are instantiated to construct and dynamically update the global operation and maintenance knowledge graph. Calculate the correlation strength between different data dimensions based on the global operation and maintenance knowledge graph, and identify potential causal relationships; When an abnormality in service quality is detected, starting from the abnormal node, the root cause analysis algorithm is executed in reverse on the operation and maintenance knowledge graph based on the potential causal relationship to calculate the probability that each node in the global operation and maintenance knowledge graph is the root cause and to determine the possible fault propagation path. Based on the possible fault propagation paths, the abnormal locations are located, and the trends of the abnormal locations are predicted. The results of the abnormal locations and the trend predictions are combined to form a complete insight. Integrate all complete insights to generate an operations and maintenance insight report; Based on the operation and maintenance insight report, strategy optimization suggestions are generated, and the operation and maintenance insight report and strategy optimization suggestions are sent to the operation and maintenance management center.

[0011] This invention provides a communication equipment operation and maintenance management method, which generates strategy optimization suggestions based on the operation and maintenance insight report, and performs closed-loop feedback and model optimization on the operation and maintenance insight report and strategy optimization suggestions, including: Based on the diagnostic conclusions and prediction results of the aforementioned operation and maintenance insight report, executable strategy code is automatically generated according to a predefined strategy logic template, serving as a strategy optimization suggestion. While generating executable strategy code, a metadata link is created and stored between the executable strategy code and the corresponding operation and maintenance insight report to form a decision knowledge graph. Based on the target execution location, the strategy optimization suggestions and executable strategy code are layered to form different hierarchical strategy packages; The hierarchical strategy package is sent to the corresponding target execution location via a secure channel; The effectiveness of the proposed strategy optimization suggestions is tracked, and closed-loop feedback and model optimization are performed in conjunction with the decision knowledge graph.

[0012] This invention provides a communication equipment operation and maintenance management system, comprising: Map generation module: The operation and maintenance management center receives operation and maintenance tasks, performs modeling and analysis on the operation and maintenance tasks, and generates a task execution map; Location determination module: Based on the task execution map, obtain the resource requirement attributes of the operation and maintenance task, and based on the resource requirement attributes and the real-time resource status information, dynamically determine the target execution location of the operation and maintenance task through an intelligent scheduling algorithm; Asynchronous upload module: Unloads the operation and maintenance task to the corresponding target execution location, and after execution, obtains key results, classifies the key results, and uploads the classification results asynchronously; Distribution module: Generates operation and maintenance insight reports and strategy optimization suggestions based on asynchronous upload results, and performs closed-loop feedback and model optimization on the operation and maintenance insight reports and strategy optimization suggestions.

[0013] Compared with existing technologies, the beneficial effects of this application are as follows: The operation and maintenance management center generates a task execution map through modeling and analysis. Based on the resource demand attributes of the task execution map and the real-time acquired resource status information, it intelligently schedules operation and maintenance tasks to the target execution location. After execution, key results are classified and uploaded asynchronously, thereby generating operation and maintenance insight reports and strategy optimization suggestions. Finally, model optimization is achieved through closed-loop feedback. Relying on edge computing technology, operation and maintenance capabilities are moved from the traditional centralized cloud to the edge side, enabling low-latency sensitive operation and maintenance tasks to be scheduled to the nearest edge node for execution without relying on remote cloud computing power transmission and processing. This fully leverages the core technological advantages of edge computing, such as close-range deployment, low transmission latency, and localized processing, to achieve intelligent scheduling of computing resources across regions and levels. By dynamically sensing task requirements and resource status, it solves the problem of computing power mismatch, improves resource utilization, meets the low-latency requirements of 5G / 6G edge applications, and constructs a new paradigm for computing network operation and maintenance.

[0014] The features and advantages of the present invention will be set forth in the following description, and will be apparent in part from the description, or may be learned by practicing the invention. The objects and other advantages of the invention may be realized and obtained by means of the structures particularly pointed out in the written description and the accompanying drawings.

[0015] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description

[0016] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings: Figure 1 This is a flowchart illustrating a communication equipment operation and maintenance management method provided in an embodiment of the present invention; Figure 2 This is a schematic diagram of the structure of a communication equipment operation and maintenance management system provided in an embodiment of the present invention. Detailed Implementation

[0017] The preferred embodiments of the present invention will be described below with reference to the accompanying drawings. It should be understood that the preferred embodiments described herein are for illustration and explanation only and are not intended to limit the present invention. Example 1:

[0018] This invention provides a method for operation and maintenance management of communication equipment, such as... Figure 1 As shown, it includes: Step 1: The operation and maintenance management center receives the operation and maintenance tasks, models and analyzes the operation and maintenance tasks, and generates a task execution graph; Step 2: Obtain the resource requirement attributes of the operation and maintenance task according to the task execution map. Based on the resource requirement attributes and the real-time resource status information, dynamically determine the target execution location of the operation and maintenance task through an intelligent scheduling algorithm. Step 3: Unload the operation and maintenance task to the corresponding target execution location. After execution, obtain the key results, classify the key results, and upload the classification results asynchronously. Step 4: Generate an operation and maintenance insight report and strategy optimization suggestions based on the asynchronous upload results, and perform closed-loop feedback and model optimization on the operation and maintenance insight report and strategy optimization suggestions.

[0019] In this embodiment, the operations and maintenance management center is the entity that receives operations and maintenance tasks and performs modeling and analysis on them; it is the initiation and control center of the entire process. For example, a cloud-based software platform that manages a nationwide base station network.

[0020] In this embodiment, the operation and maintenance task is a specific work request received by the operation and maintenance management center, and it is the object of modeling and analysis. For example, improving the recognition accuracy of the video surveillance system in XX Industrial Park at night.

[0021] In this embodiment, the resource requirement attributes describe the specific quantitative requirements of various resources needed to execute a sequence of atomic tasks. For example, for the atomic task sequence compressing and transmitting data packets, its resource requirement attributes include: high computational complexity requiring one CPU core, and low latency sensitivity accepting a processing time of 100ms.

[0022] In this embodiment, the task execution graph is the final output, a machine-readable global execution blueprint that integrates atomic task sequences, logical dependencies, and resource requirement models. For example, a visual flowchart where nodes represent atomic task sequences with resource requirement models and arrows represent logical dependencies clearly describes the complete execution path of the entire complex operations and maintenance task.

[0023] In this embodiment, resource status information is real-time data on available resources collected from edge devices, edge nodes, and cloud data centers. This information includes, but is not limited to, dynamic metrics such as real-time CPU utilization, available memory capacity, GPU load status, current network bandwidth, and storage I / O performance of each computing node. For example, the number of currently available GPU computing units on edge nodes, the remaining virtual machine quota in the cloud data center, and the remaining power of edge devices are all important bases for resource matching.

[0024] In this embodiment, the intelligent scheduling algorithm is an advanced algorithm based on multi-objective optimization theory, which comprehensively weighs the core optimization objectives while satisfying various constraints. Such algorithms can simultaneously consider multiple conflicting optimization objectives, such as genetic algorithms and particle swarm optimization, and find the optimal matching solution between resource demand and resource supply through intelligent search in the solution space.

[0025] In this embodiment, the target execution location is a specific computing location ultimately selected from the Pareto optimal solution set based on a preset scheduling strategy. This location could be an edge computing node, a cloud virtual machine, or an edge device; it is the optimal task execution location determined after comprehensive evaluation from multiple dimensions. For example, a resource-rich and low-cost regional edge node might be selected as the final task execution location based on a cost-optimal strategy.

[0026] In this embodiment, the resource requirements of atomic task sequences are extracted by parsing the task execution graph, and a set of candidate execution locations is generated by combining the real-time resource status. With latency sensitivity and task priority as the core objectives, an intelligent scheduling algorithm is used to solve for the Pareto optimal solution set, and finally the target execution location is selected according to a preset strategy.

[0027] In this embodiment, the key results are the important output data generated after each target execution location completes its task calculations, reflecting the core state and main achievements of the task execution. These results include basic information such as the success or failure status of the task, execution time, and resource consumption, as well as task-specific business outputs, such as fault codes detected by the equipment and target lists identified by video analysis, which are data with decision-making value.

[0028] In this embodiment, after the operation and maintenance management center unloads the operation and maintenance tasks to the target execution location, each target execution location collaboratively calculates key results based on the task execution map and classifies them into real-time control data and analysis and optimization data, which are synchronously fed back to the management center through the control channel and asynchronously uploaded to the cloud center through the data channel.

[0029] In this embodiment, the operations and maintenance insight report is a structured diagnostic document based on in-depth analysis. This report systematically presents the root causes of problems, the scope of impact, development trends, and handling recommendations, including visual charts and data support, providing comprehensive reference for operations and maintenance decisions. Examples include analysis reports containing elements such as fault propagation path diagrams and performance trend prediction curves.

[0030] In this embodiment, the strategy optimization suggestions are targeted improvement plans generated based on the operation and maintenance insight report. These suggestions are specific and feasible, involving multiple aspects such as parameter adjustment, rule optimization, and resource configuration. For example, it is suggested to adjust the maximum number of database connection pool connections from 100 to 150 to cope with the current business growth pressure and provide clear guidance for system optimization.

[0031] In this embodiment, a full-domain data view and operation and maintenance knowledge graph are constructed based on asynchronous upload results. Potential causal relationships are identified through correlation analysis. Root cause location and trend prediction are performed when business anomalies occur. Finally, operation and maintenance insight reports and strategy optimization suggestions are generated and sent to the management center, realizing the transformation from massive operation and maintenance data to intelligent decision-making.

[0032] The working principle and beneficial effects of the above technical solution are as follows: The operation and maintenance management center generates a task execution map through modeling and analysis. Based on the resource demand attributes of the task execution map and the real-time acquired resource status information, it intelligently schedules operation and maintenance tasks to the target execution location. After execution, key results are classified and uploaded asynchronously, generating operation and maintenance insight reports and strategy optimization suggestions. Finally, model optimization is achieved through closed-loop feedback. This enables intelligent scheduling of computing resources across regions and levels, solves the problem of computing power mismatch by dynamically sensing task requirements and resource status, improves resource utilization, meets the low latency requirements of 5G / 6G edge applications, and constructs a new paradigm for computing power network operation and maintenance. Example 2:

[0033] This invention provides a method for operation and maintenance management of communication equipment. The operation and maintenance management center receives operation and maintenance tasks, performs modeling and analysis on the tasks, and generates a task execution graph, including: The operations and maintenance management center receives operations and maintenance tasks, parses the high-level business intent of the operations and maintenance tasks, and identifies the core objectives and task constraints of the operations and maintenance tasks. The high-level business intent is decomposed into a set of atomic task sequences that can be directly executed by the system, and the logical dependencies between the atomic task sequences are defined. Based on the core objectives and task constraints, determine the resource requirement attributes corresponding to each atomic task sequence, and form a resource requirement model for each atomic task sequence based on the resource requirement attributes. Based on the atomic task sequence, the corresponding resource requirement model, and logical dependencies, a task execution graph is constructed.

[0034] In this embodiment, the high-level business intent is an abstract description of the operational task objectives, focusing on business value rather than specific technical implementation. For example, ensuring the real-time communication reliability of autonomous vehicles on highways.

[0035] In this embodiment, the core objective and task constraints are specific, measurable optimization directions and mandatory limitations derived from the high-level business intent. The core objective can be quantified as communication reliability with a packet loss rate of less than 0.001%. The constraint can be defined as the task completion cost not exceeding Y yuan.

[0036] In this embodiment, an atomic task sequence is a minimal set of independently executable tasks obtained by decomposing high-level business intentions. For example, to ensure communication reliability, it can be decomposed into a series of atomic task sequences such as monitoring base station signal strength, switching to a backup communication link, and compressing and transmitting data packets.

[0037] In this embodiment, logical dependencies define the execution order or triggering conditions of each task in the atomic task sequence. For example, the atomic task sequence can only switch to the backup communication link after the atomic task sequence detects an interruption in the main link signal.

[0038] In this embodiment, the inputs to the resource requirement model include technical characteristics parsed from the atomic task sequence, core objectives and task constraints derived from high-level business intent, and actual resource consumption data recorded during historical task execution. These inputs collectively form the foundation for model construction. The output of the resource requirement model is a structured, quantitatively described task resource profile, specifically represented by a set of standardized resource requirement attribute key-value pairs, including, but not limited to: the required type and quantity of computing units, memory capacity, storage performance indicators, network bandwidth requirements, latency upper limit threshold, and task priority values. This output provides a directly comparable decision-making basis for intelligent scheduling algorithms.

[0039] The working principle and beneficial effects of the above technical solution are as follows: The operation and maintenance management center analyzes the high-level business intent of operation and maintenance tasks, identifies core objectives and constraints, decomposes them into atomic task sequences with logical dependencies, and establishes a resource requirement model for each atomic task sequence. Finally, it constructs a machine-understandable task execution graph, realizing the standardization and intelligent decomposition of operation and maintenance tasks. This lays a solid foundation for the subsequent accurate scheduling and efficient collaboration of resources, and greatly improves operation and maintenance efficiency. Example 3:

[0040] This invention provides a communication equipment operation and maintenance management method, which obtains the resource requirement attributes of the operation and maintenance tasks based on the task execution map, and dynamically determines the target execution location of the operation and maintenance tasks based on the resource requirement attributes and real-time acquired resource status information, using an intelligent scheduling algorithm. The method includes: The task execution graph is analyzed to extract the sub-requirements of each atomic task sequence. The sub-requirements are aggregated and calculated in combination with logical dependencies to obtain the resource requirement attributes of the operation and maintenance task. The resource requirement attributes are compared with the resource status information obtained in real time from the terminal side, edge side and cloud side to generate a set of candidate execution locations; Using the latency sensitivity and task priority in the resource demand attributes as the core optimization objectives and other resource demands as demand constraints, the resource matching matrix is ​​evaluated using an intelligent scheduling algorithm to solve for the Pareto optimal solution set. The resource matching matrix is ​​derived by using the candidate execution positions in the candidate execution position set as rows and the quantified resource requirement attributes and the real-time acquired resource status information as columns. From the Pareto optimal solution set, a final target execution location is selected according to a preset scheduling strategy.

[0041] In this embodiment, sub-requirements refer to the specific resource requirements extracted from the atomic task sequences of the task execution graph. Each atomic task sequence requires specific computing, storage, network, and other resources during execution, and these detailed resource requirements constitute sub-requirements. For example, in a video intelligent analysis task, the video decoding atomic task sequence requires GPU decoding resources, the feature extraction atomic task sequence requires AI inference computing power, and the result reporting atomic task sequence requires network bandwidth. These different resource requirements are the sub-requirements of each atomic task sequence.

[0042] In this embodiment, aggregate computation refers to the computational process of systematically integrating the sub-requirements of each atomic task sequence based on the logical dependencies between atomic task sequences. For atomic task sequences executed in parallel, the peak value of each resource requirement is taken as the overall requirement; for atomic task sequences executed sequentially, the resource requirements of each stage are accumulated according to the execution order; for atomic task sequences executed conditionally, the expected resource requirements are calculated based on a probability model. For example, in data processing tasks, the storage requirement of multiple parallel data preprocessing tasks should be the maximum value, while the computation time of multiple sequentially executed processing stages needs to be accumulated.

[0043] In this embodiment, the candidate execution location set is a set of computing locations with basic execution capabilities selected by initially matching the resource requirement attributes of the operation and maintenance task with real-time resource status information. This set includes all computing nodes on the edge, cloud, and device sides that meet the minimum resource requirements. For example, for an AI inference task that requires 8GB of memory, all edge servers and cloud servers with currently available memory greater than 8GB will be included in the candidate execution location set.

[0044] In this embodiment, the core optimization objective is to prioritize and optimize key performance indicators during resource scheduling. In this scheme, latency sensitivity and task priority are identified as the core optimization objectives. Latency sensitivity requires tasks to be completed within a specific timeframe, such as requiring autonomous driving perception tasks to be completed within 100 milliseconds; task priority reflects the importance of the service, such as prioritizing emergency fault handling over routine data backup tasks.

[0045] In this embodiment, the demand constraints are the hard resource requirements that must be met in addition to the core optimization objective. These conditions constitute the basic threshold for task execution, including computing resource constraints, storage capacity constraints, and network bandwidth constraints. For example, the task must run on a node equipped with specific AI acceleration hardware, or the node must have a network transmission capacity of no less than 10Gbps; these are all demand constraints.

[0046] In this embodiment, the Pareto optimal solution set is a collection of optimal solutions obtained by the intelligent scheduling algorithm. In this solution set, any improvement in the performance of one objective of a solution will inevitably lead to a decrease in the performance of other objectives, and all solutions are in an optimal balance state. For example, in the trade-off between latency and cost, the solution set includes multiple non-dominated solutions such as the shortest latency but higher cost and the lowest cost but slightly longer latency.

[0047] In this embodiment, the preset scheduling strategy is a final decision rule pre-set according to the characteristics of different business scenarios and operation and maintenance management requirements. These strategies are used to select the most suitable solution for the current business needs from the Pareto optimal solution set, such as always prioritizing latency requirements, selecting the lowest cost solution while meeting basic performance requirements, or prioritizing the use of green energy nodes, etc.

[0048] The working principle and beneficial effects of the above technical solution are as follows: Resource requirements of atomic task sequences are extracted by analyzing the task execution graph, and a set of candidate execution locations is generated by combining real-time resource status. With latency sensitivity and task priority as the core objectives, an intelligent scheduling algorithm is used to solve for the Pareto optimal solution set. Finally, the target execution location is selected according to a preset strategy, achieving precise matching and multi-objective optimized scheduling of edge-cloud resources, effectively improving resource utilization, ensuring the service quality of low-latency, high-priority tasks, and providing optimal execution location decisions for complex operation and maintenance scenarios. Example 4:

[0049] This invention provides a communication equipment operation and maintenance management method, which selects a final target execution location from the Pareto optimal solution set according to a preset scheduling strategy, including: Based on the metadata of the operation and maintenance task, identify the tenant identity to which the operation and maintenance task belongs, and query the customized scheduling policy bound to the corresponding tenant identity. Using the customized scheduling strategy as the evaluation criterion, the priority of each candidate execution position in the Pareto optimal solution set is ranked. Select the candidate execution position that has the highest priority and use it as the target execution position.

[0050] In this embodiment, metadata is structured data describing the basic attributes of operation and maintenance tasks, including fundamental information such as task origin, creation time, and task type. During tenant identification, metadata carries key identifiers that distinguish different user groups; for example, identity tokens in task request headers, business department codes in task descriptions, or specific access keys used when creating tasks. This data provides the system with the basic basis for identifying task ownership.

[0051] In this embodiment, tenant identity refers to a unique identifier for a user or user group that is isolated from each other within the shared infrastructure. Each tenant represents an independent business entity or department, such as an autonomous driving business unit, a video surveillance business team, or an external enterprise customer. The system achieves logical isolation and differentiated management of resources by identifying tenant identities, ensuring that the tasks of different tenants are executed according to their respective quality of service requirements.

[0052] In this embodiment, the customized scheduling strategy is a set of resource scheduling rules pre-configured for a specific tenant, reflecting the tenant's unique service quality requirements and optimization goals. For example, an ultra-low latency strategy is configured for autonomous driving services, requiring tasks to be scheduled to the edge node closest to the data source; a cost-optimized strategy is configured for data analysis services, prioritizing cloud resources with lower pricing.

[0053] In this embodiment, the evaluation criteria are a concrete quantitative manifestation of the customized scheduling strategy, transforming the strategy into a computable evaluation standard. For example, in a low-latency-first strategy, the evaluation criteria might include a specific configuration of a latency weight of 0.8 and a cost weight of 0.2; in a high-reliability strategy, the evaluation criteria might require candidate nodes to have redundancy backup capabilities. These criteria provide clear metrics for ranking.

[0054] In this embodiment, priority ranking is a process of comprehensively scoring candidate execution positions in the Pareto optimal solution set based on evaluation criteria. The system calculates a comprehensive score based on the performance of each candidate position in various dimensions of the evaluation criteria. For example, it considers multiple factors such as latency performance, resource cost, and node reliability, and sorts the candidate positions from high to low according to the total score to form a priority selection sequence.

[0055] The working principle and beneficial effects of the above technical solution are as follows: Tenant identities are identified based on operation and maintenance task metadata, and their customized scheduling strategies are obtained. These customized scheduling strategies are then used as evaluation criteria to prioritize candidate execution positions in the Pareto optimal solution set, ultimately selecting the highest-ranked position as the target execution position. This achieves differentiated service quality assurance in multi-tenant scenarios, ensuring that the personalized scheduling needs of different tenants are met, and improving service satisfaction and resource utilization efficiency in a resource-sharing environment. Example 5:

[0056] This invention provides a method for operation and maintenance management of communication equipment, which involves offloading operation and maintenance tasks to corresponding target execution locations, obtaining key results after execution, classifying the key results, and asynchronously uploading the classification results, including: Based on the scheduling decision, the operation and maintenance management center generates task execution instructions and unloads the operation and maintenance tasks and atomic task sequences to the corresponding target execution locations via the control channel; After receiving instructions, each target execution location loads the operation and maintenance task context and performs collaborative calculations based on the logical dependencies in the task execution graph to obtain key results. The key results obtained at each target execution location are divided into two categories: real-time control data and analysis and optimization data. The encapsulated real-time control data is synchronously fed back to the operation and maintenance management center through the control channel, while the encapsulated analysis and optimization data is uploaded to the cloud center asynchronously through the data channel as a structured dataset.

[0057] In this embodiment, the scheduling decision result is the final output of the intelligent scheduling algorithm after multi-objective optimization calculation, which clearly specifies the specific location information where the operation and maintenance task and its atomic task sequence should be executed. This result is the optimal solution formed after comprehensively considering multiple factors such as task resource requirements, real-time resource status, and tenant policies. For example, the algorithm calculates and decides to offload the three atomic tasks of the video analysis task to edge node A, edge node B, and cloud center C for collaborative execution.

[0058] In this embodiment, the task execution command is an operable control command generated by the operation and maintenance management center based on the scheduling decision results, containing all the necessary information required for task execution. These commands are sent through the control channel, carrying not only target location information but also key configuration data such as task execution map, resource allocation parameters, and timeout settings, ensuring that the receiver can execute the task accurately.

[0059] In this embodiment, the operation and maintenance task context is the complete state information and environmental data that needs to be carried during task execution, providing all the background knowledge required for task execution at each execution location. This includes the task's initial input parameters, user identity credentials, service access permissions, historical execution status, etc., ensuring that each atomic task in distributed execution can maintain a consistent execution environment and state awareness.

[0060] In this embodiment, the key results are the important output data generated after each target execution location completes its task calculations, reflecting the core state and main achievements of the task execution. These results include basic information such as the success or failure status of the task, execution time, and resource consumption, as well as task-specific business outputs, such as fault codes detected by the equipment and target lists identified by video analysis, which are data with decision-making value.

[0061] In this embodiment, real-time control data is critical information that needs to be reported immediately for real-time system monitoring and rapid response, and it has the characteristics of high timeliness and high priority. This type of data typically includes task execution status, key performance indicators, and emergency alarm information, such as equipment temperature over-limit alarms and network connection interruption notifications, which are information that the operation and maintenance management center needs to know and handle immediately.

[0062] In this embodiment, the analysis and optimization data is process data used for long-term storage and in-depth analysis, characterized by large capacity and low timeliness requirements. This type of data includes detailed operation logs, complete performance indicator curves, resource consumption details, etc., such as historical CPU utilization curves and detailed network traffic statistics. This data will be uploaded to the cloud center for trend analysis and model optimization.

[0063] In this embodiment, the control channel is a dedicated communication channel for transmitting high-priority control information and real-time data, featuring low latency and high reliability. This channel is specifically used to transmit data with extremely high timeliness requirements, such as task execution instructions, real-time status feedback, and emergency alarms, ensuring that the control system can obtain critical information in a timely manner and respond rapidly.

[0064] In this embodiment, the structured dataset is a standardized and encapsulated collection of analytically optimized data with a unified format specification and metadata description. These datasets are organized according to a predefined data model and contain complete data schema, timestamp sequences, data quality identifiers, and other information, facilitating batch processing and analysis by the big data platform in the cloud center.

[0065] In this embodiment, the data channel is an asynchronous communication channel specifically designed for transmitting large volumes of analytical data, featuring high bandwidth and tolerable latency. This channel is responsible for uploading structured datasets generated at various execution locations to the cloud center in batches, supporting features such as resume upload and data compression, ensuring that massive amounts of analytical data can be reliably and efficiently aggregated to the data analysis platform.

[0066] The working principle and beneficial effects of the above technical solution are as follows: After the operation and maintenance management center unloads the operation and maintenance tasks to the target execution locations, each target execution location collaboratively calculates key results based on the task execution map and classifies them into real-time control data and analysis and optimization data. These data are then synchronously fed back to the management center through the control channel and asynchronously uploaded to the cloud center through the data channel, respectively. This achieves intelligent hierarchical processing of task execution results, which not only ensures the efficiency of real-time feedback of key data, but also reduces the system load through asynchronous transmission, providing parallel data support for real-time decision-making and in-depth analysis. Example 6:

[0067] This invention provides a method for operation and maintenance management of communication equipment, which generates an operation and maintenance insight report and strategy optimization suggestions based on asynchronous upload results, and performs closed-loop feedback and model optimization on the operation and maintenance insight report and strategy optimization suggestions, including: Draw a global data view based on the asynchronous upload result set; Based on the predefined operation and maintenance meta-model, the entities and their corresponding relationships in the global data view are instantiated to construct and dynamically update the global operation and maintenance knowledge graph. Calculate the correlation strength between different data dimensions based on the global operation and maintenance knowledge graph, and identify potential causal relationships; When an abnormality in service quality is detected, starting from the abnormal node, the root cause analysis algorithm is executed in reverse on the operation and maintenance knowledge graph based on the potential causal relationship to calculate the probability that each node in the global operation and maintenance knowledge graph is the root cause and to determine the possible fault propagation path. Based on the possible fault propagation paths, the abnormal locations are located, and the trends of the abnormal locations are predicted. The results of the abnormal locations and the trend predictions are combined to form a complete insight. Integrate all complete insights to generate an operations and maintenance insight report; Based on the operation and maintenance insight report, strategy optimization suggestions are generated, and the operation and maintenance insight report and strategy optimization suggestions are sent to the operation and maintenance management center.

[0068] In this embodiment, the global data view is a unified data perspective formed by integrating and standardizing all operation and maintenance data in the distributed system. By establishing a unified data model and time-series index, data from different nodes and time periods are correlated and mapped to form a panoramic data map containing multi-dimensional information such as device status, business indicators, and resource usage, providing a complete data foundation for subsequent in-depth analysis.

[0069] In this embodiment, the predefined operations and maintenance meta-model is a predefined abstract framework describing concepts and relationships within the operations and maintenance domain. It specifies the various entity types involved in the operations and maintenance environment and their interrelationships, providing structured modeling specifications for knowledge graph construction and ensuring that data from different sources can be organized and managed according to unified standards.

[0070] In this embodiment, the instantiation of entities and their corresponding relationships is the process of transforming an abstract operational meta-model into a concrete knowledge graph. By mapping actual operational data to entity nodes and relational edges in the knowledge graph, a semantic network reflecting the real operational environment is constructed, enabling organic connections between data.

[0071] In this embodiment, potential causal relationships are those that may have causal influence, discovered through data analysis. These relationships are identified using methods such as statistical correlation analysis and time-series pattern matching. They manifest as a situation where a change in the state of one entity may trigger a change in the state of another entity. For example, a surge in database connections may lead to increased application service response latency. While such relationships require further verification, they have significant early warning value.

[0072] In this embodiment, correlation strength is an indicator that quantifies the degree of correlation between data dimensions. It is calculated using statistical methods such as correlation coefficient and mutual information, and is used to assess the closeness of the correlation between different operation and maintenance indicators. For example, a strong positive correlation of 0.85 is found between CPU utilization and memory usage, providing a quantitative basis for causal inference.

[0073] In this embodiment, the potential fault propagation path is the route of fault spread in the system based on knowledge graph reasoning. By analyzing the dependencies between entities through graph traversal algorithms, the transmission chain from the root cause node to the affected nodes is found, such as the hierarchical impact path starting from a disk failure, through storage services, database services, and application services, which helps to understand the scope of the fault's impact.

[0074] In this embodiment, complete insight is a comprehensive analytical conclusion formed by combining anomaly localization and trend prediction. It not only points out the specific location and cause of the current anomaly, but also predicts the development trend and potential impact of the anomaly. For example, it predicts that the current database connection pool utilization rate has reached 90% and is on the rise, and a connection timeout failure is expected in 2 hours, providing a comprehensive situational awareness.

[0075] In this embodiment, the root cause analysis algorithm is an intelligent algorithm for locating the root cause of a problem on a knowledge graph. Using techniques such as random walks and graph neural networks, it calculates the probability score of each node in the graph as a root cause, sorts them, and identifies the most likely sources of the anomaly. For example, the algorithm might determine that a configuration change on a certain switch is the root cause of network latency.

[0076] In this embodiment, the operations and maintenance insight report is a structured diagnostic document based on in-depth analysis. This report systematically presents the root causes of problems, the scope of impact, development trends, and handling recommendations, including visual charts and data support, providing comprehensive reference for operations and maintenance decisions. Examples include analysis reports containing elements such as fault propagation path diagrams and performance trend prediction curves.

[0077] In this embodiment, the strategy optimization suggestions are targeted improvement plans generated based on the operation and maintenance insight report. These suggestions are specific and feasible, involving multiple aspects such as parameter adjustment, rule optimization, and resource configuration. For example, it is suggested to adjust the maximum number of database connection pool connections from 100 to 150 to cope with the current business growth pressure and provide clear guidance for system optimization.

[0078] The working principle and beneficial effects of the above technical solution are as follows: Based on the asynchronous upload results, a full-domain data view and operation and maintenance knowledge graph are constructed. Potential causal relationships are identified through correlation analysis. Root cause location and trend prediction are performed when business anomalies occur. Finally, operation and maintenance insight reports and strategy optimization suggestions are generated and sent to the management center, realizing the transformation from massive operation and maintenance data to intelligent decision-making. Through data-driven problem location and trend prediction, the efficiency of fault handling and the predictive maintenance capability of the system are significantly improved, providing accurate decision support for operation and maintenance management. Example 7:

[0079] This invention provides a communication equipment operation and maintenance management method, which generates strategy optimization suggestions based on the operation and maintenance insight report, and performs closed-loop feedback and model optimization on the operation and maintenance insight report and strategy optimization suggestions, including: Based on the diagnostic conclusions and prediction results of the aforementioned operation and maintenance insight report, executable strategy code is automatically generated according to a predefined strategy logic template, serving as a strategy optimization suggestion. While generating executable strategy code, a metadata link is created and stored between the executable strategy code and the corresponding operation and maintenance insight report to form a decision knowledge graph. Based on the target execution location, the strategy optimization suggestions and executable strategy code are layered to form different hierarchical strategy packages; The hierarchical strategy package is sent to the corresponding target execution location via a secure channel; The effectiveness of the proposed strategy optimization suggestions is tracked, and closed-loop feedback and model optimization are performed in conjunction with the decision knowledge graph.

[0080] In this embodiment, diagnostic conclusions and prediction results are the core analytical outputs of the operation and maintenance insight report. Diagnostic conclusions accurately point out the root cause and current state of system anomalies, such as determining that the exhaustion of the database connection pool is the root cause of slow service response. Prediction results predict future system behavior based on historical data and trend analysis, such as predicting that storage space will be exhausted in 72 hours under the current growth trend. The two together constitute the decision-making basis for strategy formulation.

[0081] In this embodiment, the predefined policy logic templates are pre-encapsulated policy rule frameworks that provide structured templates for automatically generating policy code. These templates include standardized modules such as condition judgments, execution actions, and parameter configurations. For example, there is a template that automatically triggers elastic scaling when CPU utilization exceeds a threshold, ensuring that the generated policy code conforms to system execution specifications and is readable.

[0082] In this embodiment, the executable policy code is machine instructions automatically generated by the system and can be directly parsed and executed by the operation and maintenance system. This code is written in a standardized language and contains complete conditional judgment logic and operation instructions, such as Python scripts or YAML configurations that implement dynamic adjustment of load balancing weights, and can be directly loaded and run by the target execution environment.

[0083] In this embodiment, the strategy optimization recommendations are guidance documents for improvement schemes generated based on the operation and maintenance insight report. These recommendations include both specific technical implementation schemes and clarifies the business objectives that the strategy is expected to achieve. For example, it is recommended to migrate the computing nodes for video analytics tasks from the cloud to edge nodes to reduce transmission latency, while estimating that this adjustment can improve processing efficiency by 30%.

[0084] In this embodiment, the metadata link is a structured association record that establishes a traceable relationship between the policy code and its generation basis. A unique identifier is used to create a bidirectional index between the policy code and its corresponding operation and maintenance insight report, generation time, version number, and other information, forming a complete and traceable chain of evidence to ensure that every policy decision is verifiable.

[0085] In this embodiment, the decision knowledge graph is a semantic network that stores knowledge of the entire strategy decision-making process. This graph visually represents the complex relationships between strategy codes, operational insights, and system entities, recording the strategy generation logic and expected results, and providing experience references and learning samples for subsequent strategy optimization.

[0086] In this embodiment, the hierarchical policy package is a policy deployment unit categorized and encapsulated according to the characteristics of the target execution location. Taking into account the characteristics of different execution environments such as cloud, edge, and terminal, the policy code, configuration parameters, and dependent resources are packaged into deployment packages that adapt to the needs of each level, ensuring that the policy can be executed accurately in different environments.

[0087] In this embodiment, the secure channel is a dedicated communication channel that ensures the confidentiality and integrity of the policy transmission process. Security mechanisms such as encrypted transmission, authentication, and anti-tampering are employed to ensure the policy package is secure and controllable throughout its generation and deployment, preventing the policy from being stolen or tampered with during transmission.

[0088] In this embodiment, effect tracking is a systematic monitoring and evaluation process for the effectiveness of deployed strategies. By collecting data on changes in system metrics and business impact after strategy implementation, the degree of alignment between the actual effect of the strategy and the expected goals is quantitatively analyzed, providing empirical evidence for strategy optimization.

[0089] In this embodiment, closed-loop feedback and model optimization are an iterative process that continuously improves the decision-making model based on the policy execution effect. Policy execution effect data is fed back to the policy generation model, and machine learning methods are used to adjust model parameters, continuously improving the accuracy and effectiveness of policy generation, thus forming a self-improving intelligent decision-making loop.

[0090] The working principle and beneficial effects of the above technical solution are as follows: versioned executable policy code is generated based on operation and maintenance insight reports and policy logic templates, a decision knowledge graph is formed by establishing a metadata link with the report, policy deployment is achieved through layered encapsulation and secure distribution, and closed-loop optimization is completed based on effect tracking, realizing the automated transformation from operation and maintenance insight to policy execution, ensuring accurate policy deployment through version management and layered distribution, continuously improving policy effectiveness through closed-loop feedback, and building an autonomous and self-optimizing intelligent operation and maintenance system. Example 8:

[0091] This invention provides a communication equipment operation and maintenance management system, such as... Figure 2 As shown, it includes: Map generation module: The operation and maintenance management center receives operation and maintenance tasks, performs modeling and analysis on the operation and maintenance tasks, and generates a task execution map; Location determination module: Based on the task execution map, obtain the resource requirement attributes of the operation and maintenance task, and based on the resource requirement attributes and the real-time resource status information, dynamically determine the target execution location of the operation and maintenance task through an intelligent scheduling algorithm; Asynchronous upload module: Unloads the operation and maintenance task to the corresponding target execution location, and after execution, obtains key results, classifies the key results, and uploads the classification results asynchronously; Distribution module: Generates operation and maintenance insight reports and strategy optimization suggestions based on asynchronous upload results, and performs closed-loop feedback and model optimization on the operation and maintenance insight reports and strategy optimization suggestions.

[0092] The working principle and beneficial effects of the above technical solution are as follows: The operation and maintenance management center generates a task execution map through modeling and analysis. Based on the resource demand attributes of the task execution map and the real-time acquired resource status information, it intelligently schedules operation and maintenance tasks to the target execution location. After execution, key results are classified and uploaded asynchronously, generating operation and maintenance insight reports and strategy optimization suggestions. Finally, model optimization is achieved through closed-loop feedback. This enables intelligent scheduling of computing resources across regions and levels, solves the problem of computing power mismatch by dynamically sensing task requirements and resource status, improves resource utilization, meets the low latency requirements of 5G / 6G edge applications, and constructs a new paradigm for computing power network operation and maintenance.

[0093] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for operation and maintenance management of communication equipment, characterized in that, include: Step 1: The operation and maintenance management center receives the operation and maintenance tasks, models and analyzes the operation and maintenance tasks, and generates a task execution graph; Step 2: Obtain the resource requirement attributes of the operation and maintenance task according to the task execution map. Based on the resource requirement attributes and the real-time resource status information, dynamically determine the target execution location of the operation and maintenance task through an intelligent scheduling algorithm. Step 3: Unload the operation and maintenance task to the corresponding target execution location. After execution, obtain the key results, classify the key results, and upload the classification results asynchronously. Step 4: Generate an operation and maintenance insight report and strategy optimization suggestions based on the asynchronous upload results, and perform closed-loop feedback and model optimization on the operation and maintenance insight report and strategy optimization suggestions; In step 1, the operation and maintenance management center receives the operation and maintenance task, parses the high-level business intent of the operation and maintenance task, and identifies the core objectives and task constraints of the operation and maintenance task. The high-level business intent is decomposed into a set of atomic task sequences that can be directly executed by the system, and the logical dependencies between the atomic task sequences are defined. Based on the core objectives and task constraints, determine the resource requirement attributes corresponding to each atomic task sequence, and form a resource requirement model for each atomic task sequence based on the resource requirement attributes. Construct a task execution graph based on the atomic task sequence, the corresponding resource requirement model, and logical dependencies; Step 2 involves obtaining the resource requirement attributes of the operation and maintenance task based on the task execution graph, and dynamically determining the target execution location of the operation and maintenance task based on the resource requirement attributes and real-time acquired resource status information, using an intelligent scheduling algorithm. This includes: The task execution graph is analyzed to extract the sub-requirements of each atomic task sequence. The sub-requirements are aggregated and calculated in combination with logical dependencies to obtain the resource requirement attributes of the operation and maintenance task. The resource requirement attributes are compared with the resource status information obtained in real time from the terminal side, edge side and cloud side to generate a set of candidate execution locations; Using the latency sensitivity and task priority in the resource demand attributes as the core optimization objectives and other resource demands as demand constraints, the resource matching matrix is ​​evaluated using an intelligent scheduling algorithm to solve for the Pareto optimal solution set. The resource matching matrix is ​​derived by using the candidate execution positions in the candidate execution position set as rows and the quantified resource requirement attributes and the real-time acquired resource status information as columns. From the Pareto optimal solution set, a final target execution location is selected according to a preset scheduling strategy.

2. The communication equipment operation and maintenance management method according to claim 1, characterized in that, From the Pareto optimal solution set, a final target execution location is selected according to a preset scheduling strategy, including: Based on the metadata of the operation and maintenance task, identify the tenant identity to which the operation and maintenance task belongs, and query the customized scheduling policy bound to the corresponding tenant identity. Using the customized scheduling strategy as the evaluation criterion, the priority of each candidate execution position in the Pareto optimal solution set is ranked. Select the candidate execution position that has the highest priority and use it as the target execution position.

3. The communication equipment operation and maintenance management method according to claim 1, characterized in that, The operation and maintenance task is unloaded to the corresponding target execution location. After execution, key results are obtained, these key results are categorized, and the categorization results are uploaded asynchronously, including: Based on the scheduling decision, the operation and maintenance management center generates task execution instructions and unloads the operation and maintenance tasks and atomic task sequences to the corresponding target execution locations via the control channel; After receiving instructions, each target execution location loads the operation and maintenance task context and performs collaborative calculations based on the logical dependencies in the task execution graph to obtain key results. The key results obtained at each target execution location are divided into two categories: real-time control data and analysis and optimization data. The encapsulated real-time control data is synchronously fed back to the operation and maintenance management center through the control channel, while the encapsulated analysis and optimization data is uploaded to the cloud center asynchronously through the data channel as a structured dataset.

4. The communication equipment operation and maintenance management method according to claim 3, characterized in that, Based on the asynchronous upload results, an operation and maintenance insight report and strategy optimization suggestions are generated. These reports and suggestions are then subjected to closed-loop feedback and model optimization, including: Draw a global data view based on the asynchronous upload result set; Based on the predefined operation and maintenance meta-model, the entities and their corresponding relationships in the global data view are instantiated to construct and dynamically update the global operation and maintenance knowledge graph. Calculate the correlation strength between different data dimensions based on the global operation and maintenance knowledge graph, and identify potential causal relationships; When an abnormality in service quality is detected, starting from the abnormal node, the root cause analysis algorithm is executed in reverse on the operation and maintenance knowledge graph based on the potential causal relationship to calculate the probability that each node in the global operation and maintenance knowledge graph is the root cause and to determine the possible fault propagation path. Based on the possible fault propagation paths, the abnormal locations are located, and the trends of the abnormal locations are predicted. The results of the abnormal locations and the trend predictions are combined to form a complete insight. Integrate all complete insights to generate an operations and maintenance insight report; Based on the operation and maintenance insight report, strategy optimization suggestions are generated, and the operation and maintenance insight report and strategy optimization suggestions are sent to the operation and maintenance management center.

5. The communication equipment operation and maintenance management method according to claim 4, characterized in that, Based on the aforementioned operation and maintenance insight report, strategy optimization suggestions are generated, and the operation and maintenance insight report and strategy optimization suggestions are subjected to closed-loop feedback and model optimization, including: Based on the diagnostic conclusions and prediction results of the aforementioned operation and maintenance insight report, executable strategy code is automatically generated according to a predefined strategy logic template, serving as a strategy optimization suggestion. While generating executable strategy code, a metadata link is created and stored between the executable strategy code and the corresponding operation and maintenance insight report to form a decision knowledge graph. Based on the target execution location, the strategy optimization suggestions and executable strategy code are layered to form different hierarchical strategy packages; The hierarchical strategy package is sent to the corresponding target execution location via a secure channel; The effectiveness of the proposed strategy optimization suggestions is tracked, and closed-loop feedback and model optimization are performed in conjunction with the decision knowledge graph.

6. A communication equipment operation and maintenance management system, characterized in that, include: Map generation module: The operation and maintenance management center receives operation and maintenance tasks, performs modeling and analysis on the operation and maintenance tasks, and generates a task execution map; Location determination module: Based on the task execution map, obtain the resource requirement attributes of the operation and maintenance task, and based on the resource requirement attributes and the real-time resource status information, dynamically determine the target execution location of the operation and maintenance task through an intelligent scheduling algorithm; Asynchronous upload module: Unloads the operation and maintenance task to the corresponding target execution location, and after execution, obtains key results, classifies the key results, and uploads the classification results asynchronously; The distribution module generates an operation and maintenance insight report and strategy optimization suggestions based on the asynchronous upload results, and performs closed-loop feedback and model optimization on the operation and maintenance insight report and strategy optimization suggestions; The graph generation module is used by the operations and maintenance management center to receive operations and maintenance tasks, parse the high-level business intent of the operations and maintenance tasks, and identify the core objectives and task constraints of the operations and maintenance tasks. The high-level business intent is decomposed into a set of atomic task sequences that can be directly executed by the system, and the logical dependencies between the atomic task sequences are defined. Based on the core objectives and task constraints, determine the resource requirement attributes corresponding to each atomic task sequence, and form a resource requirement model for each atomic task sequence based on the resource requirement attributes. Construct a task execution graph based on the atomic task sequence, the corresponding resource requirement model, and logical dependencies; The location determination module obtains the resource requirement attributes of the operation and maintenance task based on the task execution map, and dynamically determines the target execution location of the operation and maintenance task based on the resource requirement attributes and real-time acquired resource status information, using an intelligent scheduling algorithm, including: The task execution graph is analyzed to extract the sub-requirements of each atomic task sequence. The sub-requirements are aggregated and calculated in combination with logical dependencies to obtain the resource requirement attributes of the operation and maintenance task. The resource requirement attributes are compared with the resource status information obtained in real time from the terminal side, edge side and cloud side to generate a set of candidate execution locations; Using the latency sensitivity and task priority in the resource demand attributes as the core optimization objectives and other resource demands as demand constraints, the resource matching matrix is ​​evaluated using an intelligent scheduling algorithm to solve for the Pareto optimal solution set. The resource matching matrix is ​​derived by using the candidate execution positions in the candidate execution position set as rows and the quantified resource requirement attributes and the real-time acquired resource status information as columns. From the Pareto optimal solution set, a final target execution location is selected according to a preset scheduling strategy.

Images