Agent-based operation and maintenance task execution method and device and electronic equipment
By using genetic algorithm optimization to generate operation and maintenance solutions, the problem of untimely response of intelligent agents in complex environments is solved, and the efficient and accurate execution of operation and maintenance tasks and real-time strategy optimization of intelligent agents are realized.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- CHINA MOBILE COMM LTD RES INST
- Filing Date
- 2026-01-06
- Publication Date
- 2026-06-05
Smart Images
Figure CN122154732A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of computer technology, and in particular to a method, apparatus, and electronic device for performing operation and maintenance tasks based on intelligent agents. Background Technology
[0002] Currently, with the rapid development of modern information technology, IT systems are becoming increasingly large-scale and complex, posing unprecedented challenges to operations and maintenance (O&M) work. Intelligent agents, as a specific application of artificial intelligence technology in the O&M field, have become a core component of modern IT system management due to their ability to automatically perform O&M tasks such as resource allocation, fault detection and repair, and performance monitoring. This effectively reduces the workload of O&M personnel and significantly improves system management efficiency.
[0003] However, existing fault diagnosis agents exhibit numerous problems in practical applications, failing to meet the complex and ever-changing operational needs. On one hand, the problem of rigid rules is prominent. Existing agents primarily rely on pre-defined operational case specifications in the operational knowledge base to construct operational decisions. This static decision-making approach has extremely limited adaptability to dynamically evolving operational environments. For example, during peak business periods, performance indicators such as data traffic, CPU utilization, memory usage, and network throughput exhibit non-linear fluctuations. Agents employing static operational strategies cannot respond to such dynamic changes in a timely manner, easily leading to imbalances in server resource scheduling, and consequently causing serious operational incidents such as service delays or even outages. On the other hand, the phenomenon of knowledge silos is severe. Although the operational knowledge base has accumulated massive amounts of historical operational data, fault cases, and technical documents, knowledge updates still rely on manual intervention. This model cannot capture subtle changes in the operational environment in a timely manner and cannot provide real-time and effective knowledge support for intelligent decision-making, significantly reducing the accuracy and timeliness of operational execution.
[0004] Therefore, how to enable intelligent agents to perform operation and maintenance tasks efficiently and accurately in complex operation and maintenance environments is a technical problem that urgently needs to be solved. Summary of the Invention
[0005] This application provides a method, apparatus, and electronic device for executing operation and maintenance tasks based on intelligent agents, in order to solve the above-mentioned technical problems.
[0006] In a first aspect, embodiments of this application provide a method for executing operation and maintenance tasks based on intelligent agents, including: The system receives an operation and maintenance task for a target operating system; collects performance index data of the target operating system at the current moment; retrieves historical operation and maintenance cases corresponding to the operation and maintenance task from the operation and maintenance case database; uses a preset optimization algorithm to obtain a target operation and maintenance plan for the operation and maintenance task based on the performance index data and the historical operation and maintenance cases; and executes the operation and maintenance task based on the target operation and maintenance plan.
[0007] In one embodiment, the preset optimization algorithm is a genetic algorithm; the step of using the preset optimization algorithm to obtain a target operation and maintenance solution for the operation and maintenance task based on the performance index data and the historical operation and maintenance cases includes: using a fitness evaluation algorithm to select multiple operation and maintenance cases from the historical operation and maintenance cases as first parent individuals based on the performance index data; wherein, the operation and maintenance case includes multiple operation and maintenance execution steps; and using a genetic algorithm based on the first parent individuals to obtain a target operation and maintenance solution for the operation and maintenance task.
[0008] In one embodiment, obtaining the target operation and maintenance scheme for the operation and maintenance task based on the first parent individual using a genetic algorithm includes: iteratively executing the following operations based on the first parent individual until the iteration termination condition is met to obtain the target operation and maintenance scheme: randomly selecting two parent individuals from the first parent individuals as the first and second parent individuals to be crossovered; performing a crossover operation on the operation and maintenance execution steps of the first and second parent individuals to be crossovered to obtain the first and second child individuals; randomly selecting operation and maintenance execution steps from the first and / or second child individuals to perform a mutation operation to obtain mutated child individuals; wherein, the mutated child individuals include at least one of a third and a fourth child individual, the third child individual being obtained by mutating the first child individual, and the fourth child individual being obtained by mutating the second child individual; using a fitness evaluation algorithm, selecting the first parent individual for the next iteration process from the first parent individuals and the mutated child individuals according to the performance index data.
[0009] In one embodiment, the step of selecting multiple operation and maintenance cases from the historical operation and maintenance cases as first parent individuals using a fitness evaluation algorithm based on the performance index data includes: randomly selecting k operation and maintenance cases from the historical operation and maintenance cases, where k is a positive integer; determining the fitness value of each of the k operation and maintenance cases using the fitness evaluation algorithm based on the performance index data; selecting the operation and maintenance case with the largest fitness value from the k operation and maintenance cases as a first parent individual; removing the first parent individual from the historical operation and maintenance cases, and continuing to execute the step of randomly selecting k operation and maintenance cases from the historical operation and maintenance cases and subsequent steps, until a preset number of first parent individuals are obtained.
[0010] In one embodiment, determining the fitness value of each of the k operation and maintenance cases based on the performance index data using a fitness evaluation algorithm includes: determining the execution effect evaluation index of each operation and maintenance case using a simulation engine corresponding to the target operation system based on the performance index data; and determining the fitness value of each operation and maintenance case based on the execution effect evaluation index of each operation and maintenance case.
[0011] In one embodiment, the performance evaluation metrics include the risk level of the operation and maintenance task, and at least one of the operation and maintenance case success rate, operation and maintenance case execution time, and operation and maintenance case complexity; determining the fitness value of each operation and maintenance case based on the performance evaluation metrics of each operation and maintenance case includes: determining the fitness value of each operation and maintenance case based on the weighted sum of each metric in the performance evaluation metrics of each operation and maintenance case.
[0012] In one embodiment, the step of performing cross-operation on the operation and maintenance execution steps of the first parent entity to be cross-operated and the second parent entity to be cross-operated to obtain the first child entity and the second child entity includes: exchanging the first operation and maintenance execution step of the first parent entity to be cross-operated and the second operation and maintenance execution step of the second parent entity to be cross-operated to obtain the first child entity and the second child entity; wherein the first operation and maintenance execution step is located at the first position of the first parent entity to be cross-operated, and the second operation and maintenance execution step is located at the second position of the second parent entity to be cross-operated.
[0013] In one embodiment, the cross-operation of the operation and maintenance execution steps for the first parent individual to be cross-operated and the second parent individual to be cross-operated, to obtain the first child individual and the second child individual, includes: the first child individual inheriting the operation and maintenance execution steps outside the first position interval of the first parent individual to be cross-operated and the operation and maintenance execution steps within the second position interval of the second parent individual to be cross-operated; the second child individual inheriting the operation and maintenance execution steps outside the second position interval of the second parent individual to be cross-operated and the operation and maintenance execution steps within the first position interval of the first parent individual to be cross-operated.
[0014] In one embodiment, the step of randomly selecting an operation and maintenance execution step from the first and / or second child individuals for mutation to obtain a mutated child individual includes: replacing the third operation and maintenance execution step in the first child individual with a fifth operation and maintenance execution step to obtain the third child individual; wherein the third operation and maintenance execution step is an operation and maintenance execution step at a random position in the first child individual, and the fifth operation and maintenance execution step is an operation and maintenance execution step in the historical operation and maintenance case that is equivalent to the third operation and maintenance execution step; and / or, replacing the fourth operation and maintenance execution step in the second child individual with a sixth operation and maintenance execution step to obtain the fourth child individual; wherein the fourth operation and maintenance execution step is an operation and maintenance execution step at a random position in the second child individual, and the sixth operation and maintenance execution step is an operation and maintenance execution step in the historical operation and maintenance case that is equivalent to the fourth operation and maintenance execution step.
[0015] In one embodiment, the step of randomly selecting maintenance execution steps from the first and / or second child individuals to perform mutation operations to obtain mutated child individuals includes: swapping the maintenance execution steps located at the third and fourth positions in the first child individual to obtain the third child individual; and / or swapping the maintenance execution steps located at the fifth and sixth positions in the second child individual to obtain the fourth child individual.
[0016] In one embodiment, the step of selecting the first parent individual for the next iteration from the first parent individual and the mutated child individuals using a fitness evaluation algorithm based on the performance index data includes: randomly selecting n operation and maintenance cases from the first parent individual and the mutated child individuals; where n is a positive integer; determining the fitness value of each of the n operation and maintenance cases based on the performance index data using a fitness evaluation algorithm; selecting the operation and maintenance case with the largest fitness value from the n operation and maintenance cases as a first parent individual for the next iteration; after removing the selected first parent individual for the next iteration from the first parent individual and the mutated child individuals, continuing to execute the step of randomly selecting n operation and maintenance cases from the first parent individual and the mutated child individuals and subsequent steps, until a preset number of first parent individuals for the next iteration are obtained.
[0017] In one embodiment, determining the fitness value of each of the n operation and maintenance cases based on the performance index data using a fitness evaluation algorithm includes: determining the execution effect evaluation index of each operation and maintenance case using a simulation engine corresponding to the target operation system based on the performance index data; and determining the fitness value of each operation and maintenance case based on the execution effect evaluation index of each operation and maintenance case.
[0018] In one embodiment, the performance evaluation metrics include the risk level of the operation and maintenance task, and at least one of the operation and maintenance case success rate, operation and maintenance case execution time, and operation and maintenance case complexity; determining the fitness value of each operation and maintenance case based on the performance evaluation metrics of each operation and maintenance case includes: determining the fitness value of each operation and maintenance case based on the weighted sum of each metric in the performance evaluation metrics of each operation and maintenance case.
[0019] In one embodiment, after executing the operation and maintenance task based on the target operation and maintenance plan, the method further includes: determining the execution result of the operation and maintenance task; and updating the operation and maintenance case database according to the target operation and maintenance plan and the execution result.
[0020] Secondly, embodiments of this application provide an agent-based operation and maintenance task execution device, comprising: The receiving module is configured to: receive an operation and maintenance task for a target operating system; the acquisition module is configured to: acquire performance indicator data of the target operating system at the current moment; the first acquisition module is configured to: acquire historical operation and maintenance cases corresponding to the operation and maintenance task from the operation and maintenance case database; the second acquisition module is configured to: obtain a target operation and maintenance plan for the operation and maintenance task based on the performance indicator data and the historical operation and maintenance cases using a preset optimization algorithm; and the execution module is configured to: execute the operation and maintenance task based on the target operation and maintenance plan.
[0021] Thirdly, embodiments of this application provide an electronic device, including a processor and a memory storing a computer program, wherein the processor executes the program to implement the steps of the agent-based operation and maintenance task execution method described in the first aspect.
[0022] Fourthly, embodiments of this application provide a non-transitory computer-readable storage medium storing a computer program thereon, wherein the computer program, when executed by a processor, implements the steps of the agent-based operation and maintenance task execution method described in the first aspect.
[0023] Fifthly, embodiments of this application provide a computer program product, including a computer program, which, when executed by a processor, implements the steps of the agent-based operation and maintenance task execution method described in the first aspect.
[0024] The intelligent agent-based operation and maintenance task execution method, apparatus, and electronic device provided in this application, upon receiving an operation and maintenance task for a target operating system, collects performance index data of the target operating system at the current moment and retrieves historical operation and maintenance cases corresponding to the operation and maintenance task from an operation and maintenance case database. Then, using a preset optimization algorithm, a target operation and maintenance plan is obtained based on the performance index data and historical operation and maintenance cases, and the operation and maintenance task is executed based on the target operation and maintenance plan. Through the synergistic effect of the preset optimization algorithm and the operation and maintenance case database, the parameter strategies in the operation and maintenance plan can be continuously optimized, thereby enabling the intelligent agent to execute operation and maintenance tasks efficiently and accurately in complex operation and maintenance environments. Attached Figure Description
[0025] To more clearly illustrate the technical solutions in this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.
[0026] Figure 1 This is a flowchart illustrating the agent-based operation and maintenance task execution method provided in an embodiment of this application; Figure 2 This is a flowchart illustrating the method for determining a target operation and maintenance solution for an operation and maintenance task, as provided in an embodiment of this application. Figure 3 This is a schematic diagram illustrating the execution process of an agent-based operation and maintenance task provided in an embodiment of this application; Figure 4 This is a schematic diagram of the structure of the agent-based operation and maintenance task execution device provided in the embodiments of this application; Figure 5 This is a schematic diagram of the structure of the electronic device provided in the embodiments of this application. Detailed Implementation
[0027] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.
[0028] Figure 1 This is a flowchart illustrating the agent-based operation and maintenance task execution method provided in an embodiment of this application. (Refer to...) Figure 1 This application provides an agent-based operation and maintenance task execution method, executed by an operation and maintenance agent, which may include: Step 101: Receive the operation and maintenance task for the target operating system.
[0029] A target operation system is a highly integrated, stable, and functionally complete system built to achieve predetermined business operation goals. It may include, but is not limited to, IT systems (such as ERP, CRM, and SCM systems), communication systems, manufacturing systems, and office systems.
[0030] Operation and maintenance (O&M) tasks are a series of operational activities performed to ensure the stable operation of a target system, optimize system performance, or resolve system faults. For example, an O&M task could be to fix a "database connection pool exhausted" fault.
[0031] In practice, the operation and maintenance intelligent agent can receive operation and maintenance task requests for specific target operating systems from the operation and maintenance management platform, user submissions, or automatic system triggers through preset interfaces or message queues.
[0032] An operations and maintenance intelligent agent is an intelligent entity built on artificial intelligence and automation technologies. It is designed to simulate the decision-making and operational capabilities of human operations and maintenance experts and automatically execute the operations and maintenance tasks of the target operating system. It integrates functional modules such as data collection, fault diagnosis, solution generation and execution optimization.
[0033] Step 102: Collect the performance index data of the target operating system at the current moment.
[0034] Performance metrics are a series of data indicators used to quantitatively evaluate the operational status and performance of a target system. Taking IT systems as an example, performance metrics may include the following data: CPU metrics include CPU utilization, CPU load, and context switch counts. Data is collected every 5 seconds under high load and every 30 seconds under low load. For example, the normal range for utilization is less than 80% (unit: %).
[0035] Memory metrics include utilization, swap space, and cache usage. Data is collected every 5 seconds under high load and every 30 seconds under low load. For example, the normal range for utilization is less than 85% (unit: %).
[0036] Disk metrics include I / O read / write speed, latency, and error rate, collected every 10 seconds. For example, the normal range for error rate is less than 5% (unit: %).
[0037] Network metrics include bandwidth utilization, latency, and packet loss rate, collected every 5 seconds. For example, the normal range for bandwidth utilization is less than 85% (unit: %).
[0038] Service metrics include response time, error rate, and concurrency, collected every 10 seconds. For example, response time is within the normal range of less than 500ms (unit: milliseconds).
[0039] Log metrics: Error log quantity, level distribution, etc., monitored in real time. Taking call exceptions as an example, the normal range is that the preset error logs (such as Service Unavailable) do not appear.
[0040] In the specific implementation process, the monitoring agent program deployed in the target operating system or the monitoring tools built into the system can be used to collect various performance index data of the target operating system in real time; the collected data is packaged according to a predetermined format; and sent to the designated data storage or processing server through network transmission protocol.
[0041] Step 103: Retrieve historical operation and maintenance cases corresponding to the operation and maintenance tasks from the operation and maintenance case database.
[0042] The Operations and Maintenance Case Database is used to store and manage historical operations and maintenance cases. Each operation and maintenance case in the database records multiple aspects of information: diagnostic process information, such as the sequence of diagnostic steps, the time taken for each step, and the diagnostic accuracy rate; repair operation records: repair plan, repair success rate, repair time, and post-repair system stability; fault classification and severity: fault classification includes hardware faults, software faults, network faults, human error, etc., and fault severity is divided into critical (the system is completely unavailable, and core business is interrupted), high (the system performance is severely degraded, and critical business is affected), medium (some system functions are abnormal, and non-critical business is affected), and low (the system has minor abnormalities, and business is basically unaffected).
[0043] In the specific implementation process, query conditions can be constructed based on the key information in the received operation and maintenance tasks; the query conditions can be used to search the operation and maintenance case database; and historical operation and maintenance cases with different complexity rates and success rates that match the current operation and maintenance task can be selected from the search results.
[0044] Step 104: Using a preset optimization algorithm, based on performance index data and historical operation and maintenance cases, obtain the target operation and maintenance solution for the operation and maintenance task.
[0045] The preset optimization algorithm is an algorithm used to generate the optimal operation and maintenance solution through specific calculation and reasoning processes based on performance index data and historical operation and maintenance cases. Examples include genetic algorithms and simulated annealing algorithms.
[0046] The target operation and maintenance solution is the optimal operation and maintenance case for operation and maintenance tasks, which aims to guide the operation and maintenance agent to execute operation and maintenance tasks efficiently and accurately.
[0047] In some embodiments, the preset optimization algorithm is a genetic algorithm. For an example of using a genetic algorithm to derive a target operation and maintenance solution for an operation and maintenance task based on performance index data and historical operation and maintenance cases, see [link to example]. Figure 2 The relevant content will not be repeated here.
[0048] Step 105: Execute maintenance tasks based on the target maintenance plan.
[0049] In practice, the operations and maintenance (O&M) intelligent agent can execute O&M tasks based on the generated target O&M plan. (This is just an example.) Figure 3 As shown, the operations and maintenance intelligent agent can perform the following operations: Tool selection: Based on the target operation and maintenance solution, select the appropriate tool from a variety of operation and maintenance tools (such as inspection tools, restart tools, capacity expansion tools, etc.).
[0050] Parameter extraction: Based on the target operation and maintenance plan, extract the parameters required to execute the operation and maintenance tasks from relevant data sources, such as the server's IP address, port number, configuration parameters, etc.
[0051] API execution: Based on the target operation and maintenance plan, specific operation and maintenance operations are performed by calling the corresponding application programming interface (API), using the selected tools and passing in the extracted parameters.
[0052] Diagnostic decision-making: During execution, the system status and operation results are monitored in real time, and diagnostic decisions are made based on preset rules and models to determine whether the task execution has achieved the expected results.
[0053] Figure 2 This is a flowchart illustrating the method for determining a target operation and maintenance solution for an operation and maintenance task, as provided in an embodiment of this application. (Refer to...) Figure 2 This application provides a method for determining a target operation and maintenance solution for an operation and maintenance task, which may include: Step 201: Using the fitness evaluation algorithm, select multiple operation and maintenance cases from historical operation and maintenance cases based on performance index data, and use them as the first parent individuals.
[0054] An operations and maintenance case study contains multiple operations and maintenance execution steps, representing a possible SOP (Standard Operating Procedure) operation process.
[0055] The first parent individual is the parent individual that serves as the initial population.
[0056] Fitness evaluation algorithms are used to quantify the performance of an individual (i.e., an operation and maintenance case) in performing an operation and maintenance task. By calculating a fitness value, they measure the individual's performance in solving the operation and maintenance task.
[0057] In practice, various methods can be used, such as fitness evaluation algorithms, to select multiple maintenance cases from historical maintenance cases based on performance metrics data, without being limited by the description in this manual. For example, tournament algorithms, roulette wheel selection algorithms, etc., can be used to select multiple maintenance cases from historical maintenance cases based on performance metrics data using fitness evaluation algorithms.
[0058] In some embodiments, k operation and maintenance cases can be randomly selected from historical operation and maintenance cases, where k is a positive integer; using a fitness evaluation algorithm, the fitness value of each operation and maintenance case in the k operation and maintenance cases is determined based on performance index data; from the k operation and maintenance cases, the operation and maintenance case with the largest fitness value is selected as a first parent individual; after removing the first parent individual from the historical operation and maintenance cases, the steps of randomly selecting k operation and maintenance cases from the historical operation and maintenance cases and subsequent steps are continued until a preset number of first parent individuals are obtained.
[0059] In the specific implementation process, the performance evaluation index for each operation and maintenance case can be determined based on the performance index data and the simulation engine corresponding to the target operation system; and the fitness value of each operation and maintenance case can be determined based on the performance evaluation index for each operation and maintenance case.
[0060] Execution effectiveness evaluation metrics are a series of quantitative or qualitative standards used to measure the effectiveness of operation and maintenance cases in solving operation and maintenance tasks.
[0061] In some embodiments, the performance evaluation metrics include the risk level of the operation and maintenance task, and at least one of the following: operation and maintenance case success rate, operation and maintenance case execution time, and operation and maintenance case complexity.
[0062] As an example, the performance evaluation metrics are shown in Table 1 below: Table 1
[0063] A simulation engine is a specialized software tool or simulation environment built based on the characteristics of a target operating system. Its core function is to provide quantitative evidence for evaluating the effectiveness of operation and maintenance solutions by simulating the operation of the target system under specific conditions.
[0064] In the specific implementation process, the collected performance index data of the target operating system can be input into the simulation engine; the simulation engine is then started, and the operation and maintenance operations are simulated in the simulation environment according to the steps and sequence specified in the operation and maintenance case. Performance index data of the simulation engine is collected after the operation and maintenance case is executed. Based on the changes in performance index data, evaluation indicators for execution effectiveness are determined. For example, for the success rate of the operation and maintenance case, the number of times the preset performance target is reached during multiple simulations is counted; for the execution time of the operation and maintenance case, the time elapsed from the start of the operation and maintenance operation until the system reaches a stable state (where various performance indicators no longer show significant changes) is recorded; for the complexity of the operation and maintenance case, information such as the number of operation steps and the technical difficulty involved is recorded.
[0065] In some embodiments, the fitness value of each operation and maintenance case can be determined based on the weighted sum of each indicator in the performance evaluation metrics for each operation and maintenance case. The calculation formula is as follows, for example only.
[0066] ; Where SuccessRate is the success rate of the operation and maintenance case; ExecutionTime is the average execution time (minutes); Complexity is the complexity of the operation and maintenance case (number of steps); Risk is the risk level of the operation and maintenance task. - These are the weighting coefficients, and their sum is 1.
[0067] In practice, the weighting coefficients can be dynamically adjusted based on the risk level of the maintenance task (for example, the weighting coefficients can be fine-tuned according to the specific fault classification scenario): emergency faults, =0.5, =0.3, =0.1, =0.1; Advanced fault, =0.4, =0.3, =0.2, =0.1; Intermediate fault, =0.3, =0.2, =0.3, =0.2; Low-level fault, =0.2, =0.2, =0.2, =0.4.
[0068] Based on the first parent individual obtained above, the following steps are executed iteratively until the iteration termination condition is met, and the target operation and maintenance solution is obtained.
[0069] Step 202: Randomly select two parent individuals from the first parent individuals as the first and second parent individuals to be crossed; perform crossover operations on the operation and maintenance execution steps of the first and second parent individuals to be crossed to obtain the first and second child individuals.
[0070] Crossover is an operation in genetic algorithms used to generate new offspring by exchanging some of the operational execution steps of two parent individuals, thereby increasing the diversity of solutions. For example, in two operational cases addressing server performance issues, exchanging their hardware check steps generates two new operational cases.
[0071] In some embodiments, the first operation and maintenance execution steps of the first parent individual to be cross-referenced and the second operation and maintenance execution steps of the second parent individual to be cross-referenced can be interchanged to obtain a first child individual and a second child individual; wherein, the first operation and maintenance execution step is located at the first position of the first parent individual to be cross-referenced, and the second operation and maintenance execution step is located at the second position of the second parent individual to be cross-referenced. In specific implementation, the first position and the second position can be the same or different.
[0072] For example, the first parent entity to be crossed contains 5 operation and maintenance execution steps, and the second parent entity to be crossed also contains 5 operation and maintenance execution steps. The second step of the first parent entity to be crossed and the second parent entity to be crossed can be swapped to obtain the first child entity and the second child entity.
[0073] In some embodiments, the first child individual may inherit the operation and maintenance execution steps outside the first position interval of the first parent individual to be crossed and the operation and maintenance execution steps within the second position interval of the second parent individual to be crossed; the second child individual may inherit the operation and maintenance execution steps outside the second position interval of the second parent individual to be crossed and the operation and maintenance execution steps within the first position interval of the first parent individual to be crossed.
[0074] For example, the first parent entity to be cross-referenced contains 5 operation and maintenance execution steps, and the second parent entity to be cross-referenced also contains 5 operation and maintenance execution steps. The first position interval and the second position interval are both the second to the third operation and maintenance execution steps. Then, the first child entity inherits the first, fourth, and fifth operation and maintenance execution steps of the first parent entity to be cross-referenced and the second and third operation and maintenance execution steps of the second parent entity to be cross-referenced. The second child entity inherits the first, fourth, and fifth operation and maintenance execution steps of the second parent entity to be cross-referenced and the second and third operation and maintenance execution steps of the first parent entity to be cross-referenced.
[0075] Step 203: Randomly select an operation and maintenance execution step from the first and / or second child individuals to perform a mutation operation to obtain mutated child individuals; wherein, the mutated child individuals include at least one of the third and fourth child individuals, the third child individual is obtained by performing a mutation operation on the first child individual, and the fourth child individual is obtained by performing a mutation operation on the second child individual.
[0076] Mutation is an operation in genetic algorithms used to introduce new genetic information by randomly altering some operational steps in offspring individuals, thus preventing the algorithm from getting trapped in local optima. For example, in a network fault resolution case, a hardware reboot step might be replaced with a software reset step.
[0077] In the specific implementation process, various methods can be used to randomly select maintenance execution steps from the first and / or second offspring individuals to perform mutation operations, thereby obtaining mutated offspring individuals, which is not limited to the description in this manual.
[0078] In some embodiments, the third operation and maintenance execution step in the first child individual can be replaced with the fifth operation and maintenance execution step to obtain the third child individual; wherein, the third operation and maintenance execution step is the operation and maintenance execution step at a random position in the first child individual, and the fifth operation and maintenance execution step is the operation and maintenance execution step that is equivalent to the third operation and maintenance execution step in historical operation and maintenance cases.
[0079] The fifth and third maintenance execution steps share similar or identical functions and roles with the third step in achieving the objectives, producing results, impacting the target operating system's state, and meeting maintenance needs. They can be substituted for each other without significantly altering the overall execution logic and final expected outcomes of the maintenance case. For example, in server performance optimization, the third maintenance execution step is "increasing server memory capacity to alleviate insufficient memory issues," while the fifth step is "optimizing the application's memory usage algorithm to reduce memory consumption." Although the methods differ, both effectively improve the server's memory constraints, making the system run more smoothly.
[0080] In some embodiments, the fourth operation and maintenance execution step in the second child individual can be replaced with the sixth operation and maintenance execution step to obtain the fourth child individual; wherein, the fourth operation and maintenance execution step is the operation and maintenance execution step at a random position in the second child individual, and the sixth operation and maintenance execution step is the operation and maintenance execution step that is equivalent to the fourth operation and maintenance execution step in historical operation and maintenance cases.
[0081] The sixth maintenance execution step and the fourth maintenance execution step have the same or similar functions and roles in key aspects such as achieving the goal, producing results, impacting the state of the target operating system, and meeting maintenance requirements. They can be substituted for each other without significantly changing the overall execution logic and final expected results of the maintenance case. For example, in security maintenance, the fourth maintenance execution step is "installing security patches of version A to fix known vulnerabilities," while the sixth maintenance execution step is "using security hardening tool B to fix vulnerabilities in the system." Both can meet the maintenance requirements of ensuring system security.
[0082] In some embodiments, the operation and maintenance execution steps located in the third and fourth positions in the first child individual can be swapped to obtain the third child individual.
[0083] For example, if the first generation individual has 5 operation and maintenance execution steps, the 3rd and 4th operation and maintenance execution steps can be swapped to obtain the third generation individual.
[0084] In some embodiments, the operation and maintenance execution steps located at the fifth and sixth positions in the second child individual can be swapped to obtain the fourth child individual.
[0085] For example, a second-generation individual has 6 operation and maintenance execution steps. The 5th and 6th operation and maintenance execution steps can be swapped to obtain a third-generation individual.
[0086] Step 204: Using the fitness evaluation algorithm, select the first parent individual for the next iteration from the first parent individual and the mutated offspring individuals based on the performance index data.
[0087] In some embodiments, n operation and maintenance cases can be randomly selected from the first parent individual and the mutated child individuals, where n is a positive integer; using a fitness evaluation algorithm, the fitness value of each operation and maintenance case among the n operation and maintenance cases is determined based on performance index data; from the n operation and maintenance cases, the operation and maintenance case with the largest fitness value is selected as the first parent individual in the next iteration process; after removing the selected first parent individuals in the next iteration process from the first parent individual and the mutated child individuals, the step of randomly selecting n operation and maintenance cases from the first parent individual and the mutated child individuals and the subsequent steps are continued until a preset number of first parent individuals in the next iteration process are obtained.
[0088] In some embodiments, the performance evaluation index for each operation and maintenance case can be determined based on the performance index data and using a simulation engine corresponding to the target operation system; and the fitness value for each operation and maintenance case can be determined based on the performance evaluation index for each operation and maintenance case.
[0089] In some embodiments, the performance evaluation metrics include the risk level of the operation and maintenance task, and at least one of the operation and maintenance case success rate, operation and maintenance case execution time, and operation and maintenance case complexity; the fitness value of each operation and maintenance case is determined based on the performance evaluation metrics of each operation and maintenance case, including: determining the fitness value of each operation and maintenance case based on the weighted sum of each metric in the performance evaluation metrics of each operation and maintenance case.
[0090] For a detailed description of the fitness evaluation algorithm, simulation engine, and execution performance evaluation metrics, please refer to the relevant content in step 201, which will not be repeated here.
[0091] In the specific implementation process, if the iteration termination condition is met, the iteration stops and the operation and maintenance case with the highest fitness value is output as the target operation and maintenance solution; if the iteration termination condition is not met, steps 202 to 204 are re-executed based on the first parent individual selected in the next round of iteration until the iteration termination condition is met.
[0092] The iteration termination conditions may include, but are not limited to, reaching the maximum number of iterations (e.g., 100 generations), the fitness value not showing a significant improvement for 20 consecutive generations, or finding the optimal solution that meets a preset threshold.
[0093] In some embodiments, after performing operation and maintenance tasks based on the target operation and maintenance plan, such as Figure 3 As shown, the execution result of the operation and maintenance task can be determined; the operation and maintenance case database is updated based on the target operation and maintenance plan and the execution result. At the same time, the operational effect of the target operation and maintenance plan can be continuously monitored to verify its stability.
[0094] In the embodiments provided in this application, genetic algorithms are deeply applied to the operation and maintenance tasks of the operation and maintenance agent, and dynamic iterative updates of the operation and maintenance case database are achieved. By utilizing the global search and adaptive adjustment characteristics of genetic algorithms, target operation and maintenance solutions that can efficiently and accurately complete operation and maintenance tasks are obtained through continuous iterative optimization. This establishes a closed-loop system of environment perception, solution generation, and case update, enabling real-time dynamic optimization of operation and maintenance strategies and effectively improving the adaptability and execution efficiency of the operation and maintenance agent in complex and ever-changing environments.
[0095] The following example, using a specific case of optimizing database connection pool failure SOPs, illustrates the technical effects of the embodiments of this application.
[0096] Fault scenario description: The application reports a "database connection pool exhausted" error, the system response time exceeds the threshold, and the number of database connections is close to the maximum value; Fault severity: High; Possible causes: Improper connection pool configuration, connection leakage, excessive database load, application code issues.
[0097] The historical SOP solutions obtained from the operations and maintenance case database are as follows: Step 1: Check the application logs to confirm the error message; Step 2: Check the number of database connections; Step 3: Check the connection pool configuration; Step 4: Restart the application service; Step 5: Monitor system metrics.
[0098] The success rate of this operation and maintenance case is 65%, the average processing time is 45 minutes, and the complexity is medium.
[0099] Using the method provided in this application's embodiments, five SOP schemes for similar faults are extracted from the operation and maintenance case database, and 45 mutation schemes are randomly generated, with an initial population size of 50. Fitness calculation: based on weights... =0.4, =0.3, =0.2, =0.1 Calculate the fitness value for each operation and maintenance case.
[0100] Evolutionary process: 100 iterations, with 20 individuals selected from each generation to enter the next generation, crossover probability 0.8, mutation probability 0.1.
[0101] The final target operation and maintenance plan is as follows: Step 1: Check the application logs to confirm the error message; Step 2: Check the number of database connections and waiting threads; Step 3: Check the connection pool configuration; Step 4: Check the connection usage in the application code; Step 5: Temporarily increase the connection pool size; Step 6: Fix connection leaks; Step 7: Optimize connection pool configuration; Step 8: Restart the application service; Step 9: Monitor system metrics.
[0102] The success rate of this operation and maintenance solution is 92%, the average processing time is 38 minutes, and the complexity is high.
[0103] The following is an example of how the target operation and maintenance plan is executed: Check the application logs to confirm the error message: "DataSource.getConnection timed out". Check database connection count: Current connection count = 490 / 500, waiting threads = 15; Check the connection pool configuration: maxTotal=500, maxWaitMillis=30000; An inspection of connection usage in the application code revealed that some connections were not closed. Temporarily increase the connection pool size: maxTotal = 800; Fix the connection leak issue: Modify the application code to ensure all connections are closed correctly; Optimize connection pool configuration: maxTotal=600, maxWaitMillis=5000; Restart the application service; Monitoring system metrics: Connection count remains stable at around 300, with no waiting threads; Execution results: Connection pool issue resolved, system response time returned to normal, connection leak issue fixed, connection pool configuration optimized.
[0104] The following describes the agent-based operation and maintenance task execution device provided in the embodiments of this application. The agent-based operation and maintenance task execution device described below can be referred to in correspondence with the agent-based operation and maintenance task execution method described above.
[0105] Figure 4 This is a schematic diagram of the structure of the agent-based operation and maintenance task execution device provided in an embodiment of this application. Figure 4 As shown, the agent-based operation and maintenance task execution device includes the following modules.
[0106] The receiving module 410 is used to receive maintenance tasks for the target operating system.
[0107] The acquisition module 420 is used to: acquire the performance index data of the target operating system at the current moment.
[0108] The first acquisition module 430 is used to: acquire historical operation and maintenance cases corresponding to the operation and maintenance task from the operation and maintenance case database.
[0109] The second acquisition module 440 is used to: obtain a target operation and maintenance plan for the operation and maintenance task based on the performance index data and the historical operation and maintenance cases using a preset optimization algorithm.
[0110] The execution module 450 is used to: execute the operation and maintenance tasks based on the target operation and maintenance plan.
[0111] Figure 5 An example is a schematic diagram of the physical structure of an electronic device, such as... Figure 5As shown, the electronic device may include: a processor 510, a communication interface 520, a memory 530, and a communication bus 540, wherein the processor 510, the communication interface 520, and the memory 530 communicate with each other via the communication bus 540. The processor 510 can call a computer program in the memory 530 to execute the steps of an agent-based operation and maintenance task execution method, such as including: The system receives an operation and maintenance task for a target operating system; collects performance index data of the target operating system at the current moment; retrieves historical operation and maintenance cases corresponding to the operation and maintenance task from the operation and maintenance case database; uses a preset optimization algorithm to obtain a target operation and maintenance plan for the operation and maintenance task based on the performance index data and the historical operation and maintenance cases; and executes the operation and maintenance task based on the target operation and maintenance plan.
[0112] Furthermore, the logical instructions in the aforementioned memory 530 can be implemented as software functional units and, when sold or used as independent products, 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 the prior art, or a portion 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, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.
[0113] On the other hand, this application also provides a computer program product, which includes a computer program that can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can perform the steps of the agent-based operation and maintenance task execution method provided in the above embodiments, such as including: The system receives an operation and maintenance task for a target operating system; collects performance index data of the target operating system at the current moment; retrieves historical operation and maintenance cases corresponding to the operation and maintenance task from the operation and maintenance case database; uses a preset optimization algorithm to obtain a target operation and maintenance plan for the operation and maintenance task based on the performance index data and the historical operation and maintenance cases; and executes the operation and maintenance task based on the target operation and maintenance plan.
[0114] On the other hand, embodiments of this application also provide a processor-readable storage medium storing a computer program for causing a processor to perform the steps of the methods provided in the above embodiments, such as including: The system receives an operation and maintenance task for a target operating system; collects performance index data of the target operating system at the current moment; retrieves historical operation and maintenance cases corresponding to the operation and maintenance task from the operation and maintenance case database; uses a preset optimization algorithm to obtain a target operation and maintenance plan for the operation and maintenance task based on the performance index data and the historical operation and maintenance cases; and executes the operation and maintenance task based on the target operation and maintenance plan.
[0115] The processor-readable storage medium can be any available medium or data storage device that the processor can access, including but not limited to magnetic memory (e.g., floppy disk, hard disk, magnetic tape, magneto-optical disk (MO)), optical memory (e.g., CD, DVD, BD, HVD), and semiconductor memory (e.g., ROM, EPROM, EEPROM, non-volatile memory (NAND FLASH), solid-state drive (SSD)).
[0116] The device embodiments described above are merely illustrative. 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 network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.
[0117] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.
[0118] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application 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. Such 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 this application.
Claims
1. A method for executing operation and maintenance tasks based on intelligent agents, characterized in that, Applied to an operational intelligence agent, the method includes: Received maintenance tasks for the target operating system; Collect the performance index data of the target operating system at the current moment; Retrieve historical operation and maintenance cases corresponding to the operation and maintenance task from the operation and maintenance case database; Using a preset optimization algorithm, based on the performance index data and the historical operation and maintenance cases, a target operation and maintenance solution is obtained for the operation and maintenance task. Based on the target operation and maintenance plan, execute the operation and maintenance tasks.
2. The method for executing operation and maintenance tasks based on intelligent agents according to claim 1, characterized in that, The preset optimization algorithm is a genetic algorithm; The step of using a preset optimization algorithm to obtain a target operation and maintenance solution for the operation and maintenance task based on the performance index data and the historical operation and maintenance cases includes: Using a fitness evaluation algorithm, based on the performance index data, multiple operation and maintenance cases are selected from the historical operation and maintenance cases as the first parent individuals; wherein, each operation and maintenance case contains multiple operation and maintenance execution steps; Based on the first parent individual, a target operation and maintenance plan for the operation and maintenance task is obtained using a genetic algorithm.
3. The agent-based operation and maintenance task execution method according to claim 2, characterized in that, The step of obtaining the target operation and maintenance solution for the operation and maintenance task based on the first parent individual using a genetic algorithm includes: Based on the first parent individual, the following operations are performed iteratively until the iteration termination condition is met, resulting in the target operation and maintenance solution: Two parent individuals are randomly selected from the first parent individuals to serve as the first and second parent individuals to be crossed. Cross operations are performed on the operation and maintenance execution steps of the first parent individual to be crossed and the second parent individual to be crossed to obtain the first child individual and the second child individual; Randomly select maintenance execution steps from the first and / or second child individuals to perform mutation operations to obtain mutated child individuals; wherein, the mutated child individuals include at least one of a third and a fourth child individual, the third child individual is obtained by performing mutation operations on the first child individual, and the fourth child individual is obtained by performing mutation operations on the second child individual; Using a fitness evaluation algorithm, based on the performance index data, the first parent individual for the next iteration is selected from the first parent individual and the mutated offspring individuals.
4. The method for executing operation and maintenance tasks based on intelligent agents according to claim 3, characterized in that, The fitness evaluation algorithm, based on the performance index data, selects multiple operation and maintenance cases from the historical operation and maintenance cases as the first parent individuals, including: From the historical operation and maintenance cases, k operation and maintenance cases are randomly selected; where k is a positive integer. Using a fitness evaluation algorithm, the fitness value of each of the k operation and maintenance cases is determined based on the performance index data; From the k operation and maintenance cases, select the operation and maintenance case with the largest fitness value as a first parent individual; After removing the first parent individual from the historical operation and maintenance cases, continue to execute the step of randomly selecting k operation and maintenance cases from the historical operation and maintenance cases and the subsequent steps, until a preset number of first parent individuals are obtained.
5. The agent-based operation and maintenance task execution method according to claim 4, characterized in that, The step of using a fitness evaluation algorithm to determine the fitness value of each of the k operation and maintenance cases based on the performance index data includes: Based on the performance index data, the execution effect evaluation index of each operation and maintenance case is determined using a simulation engine corresponding to the target operation system. Based on the performance evaluation metrics of each operation and maintenance case, the fitness value of each operation and maintenance case is determined.
6. The agent-based operation and maintenance task execution method according to claim 5, characterized in that, The performance evaluation metrics include the risk level of the operation and maintenance task, as well as at least one of the following: operation and maintenance case success rate, operation and maintenance case execution time, and operation and maintenance case complexity. The step of determining the fitness value of each operation and maintenance case based on the performance evaluation indicators of each operation and maintenance case includes: The fitness value of each operation and maintenance case is determined by the weighted sum of each indicator in the performance evaluation index of each operation and maintenance case.
7. The agent-based operation and maintenance task execution method according to any one of claims 3 to 6, characterized in that, The operation and maintenance execution steps for the first and second parent individuals to be cross-operated are cross-operated to obtain the first and second child individuals, including: The first operation and maintenance execution step of the first parent individual to be crossed and the second operation and maintenance execution step of the second parent individual to be crossed are swapped to obtain the first child individual and the second child individual; wherein, the first operation and maintenance execution step is located at the first position of the first parent individual to be crossed, and the second operation and maintenance execution step is located at the second position of the second parent individual to be crossed.
8. The agent-based operation and maintenance task execution method according to any one of claims 3 to 6, characterized in that, The operation and maintenance execution steps for the first and second parent individuals to be cross-operated are cross-operated to obtain the first and second child individuals, including: The first child individual inherits the operation and maintenance execution steps outside the first position interval of the first parent individual to be crossed, and the operation and maintenance execution steps within the second position interval of the second parent individual to be crossed; The second child individual inherits the operation and maintenance execution steps outside the second position interval of the second parent individual to be crossed, as well as the operation and maintenance execution steps within the first position interval of the first parent individual to be crossed.
9. The agent-based operation and maintenance task execution method according to any one of claims 3 to 6, characterized in that, The step of randomly selecting maintenance execution steps from the first and / or second offspring individuals to perform mutation operations to obtain mutated offspring individuals includes: The third maintenance execution step in the first child individual is replaced with the fifth maintenance execution step to obtain the third child individual; wherein, the third maintenance execution step is the maintenance execution step at a random position in the first child individual, and the fifth maintenance execution step is the maintenance execution step in the historical maintenance case that is equivalent to the third maintenance execution step; And / or, replace the fourth operation and maintenance execution step in the second child individual with the sixth operation and maintenance execution step to obtain the fourth child individual; wherein, the fourth operation and maintenance execution step is the operation and maintenance execution step at a random position in the second child individual, and the sixth operation and maintenance execution step is the operation and maintenance execution step in the historical operation and maintenance case that is equivalent to the fourth operation and maintenance execution step.
10. The agent-based operation and maintenance task execution method according to any one of claims 3 to 6, characterized in that, The step of randomly selecting maintenance execution steps from the first and / or second offspring individuals to perform mutation operations to obtain mutated offspring individuals includes: The operation and maintenance execution steps located at the third and fourth positions in the first child individual are swapped to obtain the third child individual; And / or, swap the operation and maintenance execution steps located at the fifth and sixth positions in the second child individual to obtain the fourth child individual.
11. The agent-based operation and maintenance task execution method according to claim 3, characterized in that, The step of using a fitness evaluation algorithm to select the first parent individual for the next iteration from the first parent individual and the mutated offspring individuals based on the performance index data includes: From the first parent individual and the mutated offspring individuals, n operation and maintenance cases are randomly selected; where n is a positive integer; Using a fitness evaluation algorithm, the fitness value of each of the n operation and maintenance cases is determined based on the performance index data; From the n operation and maintenance cases, select the operation and maintenance case with the largest fitness value as the first parent individual in the next iteration process; After removing the first parent individuals selected for the next iteration from the first parent individuals and the mutated child individuals, continue executing the step of randomly selecting n operation and maintenance cases from the first parent individuals and the mutated child individuals, and subsequent steps, until a preset number of first parent individuals for the next iteration are obtained.
12. The agent-based operation and maintenance task execution method according to claim 11, characterized in that, The step of using a fitness evaluation algorithm to determine the fitness value of each of the n operation and maintenance cases based on the performance index data includes: Based on the performance index data, the execution effect evaluation index of each operation and maintenance case is determined using a simulation engine corresponding to the target operation system. Based on the performance evaluation metrics of each operation and maintenance case, the fitness value of each operation and maintenance case is determined.
13. The agent-based operation and maintenance task execution method according to claim 12, characterized in that, The performance evaluation metrics include the risk level of the operation and maintenance task, as well as at least one of the following: operation and maintenance case success rate, operation and maintenance case execution time, and operation and maintenance case complexity. The step of determining the fitness value of each operation and maintenance case based on the performance evaluation indicators of each operation and maintenance case includes: The fitness value of each operation and maintenance case is determined by the weighted sum of each indicator in the performance evaluation index of each operation and maintenance case.
14. The method for executing operation and maintenance tasks based on intelligent agents according to claim 1, characterized in that, After executing the operation and maintenance task based on the target operation and maintenance plan, the method further includes: Determine the execution result of the operation and maintenance task; Update the operation and maintenance case database based on the target operation and maintenance plan and the execution results.
15. A device for executing operation and maintenance tasks based on intelligent agents, characterized in that, include: The receiving module is used to: receive maintenance tasks for the target operating system; The data acquisition module is used to: acquire the performance index data of the target operating system at the current moment; The first acquisition module is used to: acquire historical operation and maintenance cases corresponding to the operation and maintenance task from the operation and maintenance case database; The second acquisition module is used to: obtain a target operation and maintenance plan for the operation and maintenance task based on the performance index data and the historical operation and maintenance cases using a preset optimization algorithm; The execution module is used to execute the operation and maintenance tasks based on the target operation and maintenance plan.
16. An electronic device comprising a processor and a memory storing a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the agent-based operation and maintenance task execution method according to any one of claims 1 to 14.
17. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the agent-based operation and maintenance task execution method as described in any one of claims 1 to 14.
18. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by the processor, it implements the steps of the agent-based operation and maintenance task execution method according to any one of claims 1 to 14.