Artificial intelligence operation and maintenance method and system based on large model

Abstract

The invention discloses an artificial intelligence operation and maintenance method and system based on a large model. The method comprises the steps of obtaining operation data of a target platform in multiple time periods; based on a preset operation and maintenance instruction set and the operation data, first operation and maintenance instructions in multiple time periods are determined by using an operation and maintenance large model, the preset operation and maintenance instruction set is associated with the instruction features, and the preset operation and maintenance instruction set comprises at least one operation and maintenance instruction; the first operation and maintenance instruction is determined by at least part of operation and maintenance instructions in a preset operation and maintenance instruction set associated with instruction features matched with the operation data of the corresponding time period; acquiring simulation operation and maintenance information of a plurality of time periods based on the first operation and maintenance instruction; carrying out reinforcement learning on the operation and maintenance large model based on the simulation operation and maintenance information to obtain systematic deviations of a plurality of time periods; and a second operation and maintenance instruction of the target platform is generated by using the target large model based on the systematic deviation and the operation data of the multiple time periods, and the second operation and maintenance instruction is sent to the target platform, so that the automation efficiency of operation and maintenance and the rationality of an operation and maintenance scheme can be improved.

Description

Technical Field

[0001] The present application relates to the field of platform operation and maintenance technology, and in particular to an artificial intelligence operation and maintenance method and system based on a large model. Background Art

[0002] Among the related technologies, the operation and maintenance solutions for platforms (such as e-commerce platforms, financial systems, or IoT devices) can mainly include: traditional manual operation and maintenance solutions, DevOps operation and maintenance solutions, containerized operation and maintenance solutions, etc.

[0003] However, most of the above-mentioned operation and maintenance methods rely on human participation during the application process, such as manual operation and management, or manual maintenance and updating of automated tools. As a result, they are highly dependent on human labor, resulting in low efficiency, and the rationality of the operation and maintenance plan is limited by the experience of the personnel. Summary of the Invention

[0004] In order to solve the above technical problems, the embodiment of the present application proposes an artificial intelligence operation and maintenance method and system based on a large model, which can improve the automation efficiency of operation and maintenance and the rationality of operation and maintenance solutions.

[0005] In a first aspect, an embodiment of the present application provides an artificial intelligence operation and maintenance method based on a large model, comprising:

[0006] Obtain the target platform's operating data over multiple time periods;

[0007] Based on at least one preset operation and maintenance instruction set and the operation data, determining a first operation and maintenance instruction corresponding to each of the multiple time periods using the operation and maintenance macro model, wherein the at least one preset operation and maintenance instruction set is associated with at least one instruction characteristic, the preset operation and maintenance instruction set includes at least one operation and maintenance instruction, the operation and maintenance instruction being suitable for indicating a platform operation and maintenance policy, and the first operation and maintenance instruction being determined by at least some of the operation and maintenance instructions in the preset operation and maintenance instruction set associated with the instruction characteristic that matches the operation data for the corresponding time period;

[0008] Based on the first operation and maintenance instruction, obtaining simulated operation and maintenance information for each of the multiple time periods;

[0009] Reinforcement learning is performed on the large operation and maintenance model based on the simulated operation and maintenance information to obtain systematic deviations corresponding to each of the multiple time periods, wherein the systematic deviation corresponding to each time period represents a systematic error of the target platform with respect to the operation and maintenance evaluation indicator under the action of the first operation and maintenance instruction corresponding to the time period;

[0010] Based on the systematic deviations and operation data corresponding to each of the multiple time periods, a second operation and maintenance instruction for the target platform is generated using the target large model, and the second operation and maintenance instruction is sent to the target platform.

[0011] Optionally, the performing reinforcement learning on the large operation and maintenance model based on the simulated operation and maintenance information includes:

[0012] Selecting at least part of the time periods from the plurality of time periods as enhanced time periods;

[0013] For each enhancement period, performing enhancement processing on the simulated operation and maintenance information of the enhancement period to obtain enhanced operation and maintenance information of the enhancement period;

[0014] Based on the enhanced operation and maintenance information and the simulated operation and maintenance information of each of the multiple time periods, reinforcement learning is performed on the operation and maintenance large model.

[0015] Optionally, the performing enhancement processing on the simulated operation and maintenance information of the enhancement period to obtain enhanced operation and maintenance information of the enhancement period includes:

[0016] Performing disturbance processing on the simulated operation and maintenance information during the enhancement period to obtain disturbed operation and maintenance information;

[0017] Randomly adding noise to the disturbed operation and maintenance information to obtain noisy operation and maintenance information;

[0018] Determining a first probability distribution of the noisy operation and maintenance information, and determining a second probability distribution of the simulated operation and maintenance information during the enhancement period;

[0019] Based on the first probability distribution and the second probability distribution, the noisy operation and maintenance information is converted into enhanced operation and maintenance information for the enhancement period.

[0020] Optionally, converting the noisy operation and maintenance information into enhanced operation and maintenance information for the enhancement period based on the first probability distribution and the second probability distribution includes:

[0021] Determining a target noise addition strategy by analyzing a difference between the first probability distribution and the second probability distribution;

[0022] Noise processing is performed on the noisy operation and maintenance information according to the target noising strategy to obtain enhanced operation and maintenance information for the enhancement period.

[0023] Optionally, the target large model is obtained by fine-tuning the operation and maintenance large model through reinforcement learning.

[0024] Optionally, obtaining the simulated operation and maintenance information of each of the multiple time periods based on the first operation and maintenance instruction includes:

[0025] For the first operation and maintenance instruction corresponding to each time period, simulation is performed based on the first operation and maintenance instruction to simulate the operation status information of the target platform under the action of the first operation and maintenance instruction, and simulated operation and maintenance information of the time period is generated based on the operation status information.

[0026] Optionally, during the reinforcement learning process, the learning weight of the operation and maintenance model for the simulated operation and maintenance information corresponding to each time period is determined by the operating status information corresponding to the time period.

[0027] Optionally, the determining, based on at least one preset operation and maintenance instruction set and the operation data, of the first operation and maintenance instructions corresponding to each of the plurality of time periods using the operation and maintenance macro model includes:

[0028] Inputting the at least one preset operation and maintenance instruction set and the operation data of the multiple time periods into the operation and maintenance macro model;

[0029] For each time period, the operation and maintenance big model generates a corresponding first operation and maintenance instruction based on the operation data of the time period, under the prompt of at least part of the operation and maintenance instructions in the preset operation and maintenance instruction set associated with the instruction characteristics matching the operation data of the time period.

[0030] Optionally, generating the second operation and maintenance instruction of the target platform using the target macro model based on the systematic deviations and operation data corresponding to each of the multiple time periods includes:

[0031] Inputting the systematic deviations and operating data corresponding to each of the plurality of time periods into the target macro model;

[0032] The target large model determines the time period weight corresponding to each time period according to the systematic deviation corresponding to each time period;

[0033] The target large model generates the second operation and maintenance instruction based on the time period weights and operation data corresponding to each of the multiple time periods.

[0034] In a second aspect, an embodiment of the present application provides an artificial intelligence operation and maintenance system based on a large model, including:

[0035] The data acquisition module is used to obtain the operating data of the target platform in multiple time periods;

[0036] a first operation and maintenance instruction module, configured to determine, based on at least one preset operation and maintenance instruction set and the operation data, a first operation and maintenance instruction corresponding to each of the plurality of time periods using the operation and maintenance macro model, wherein the at least one preset operation and maintenance instruction set is associated with at least one instruction characteristic, the preset operation and maintenance instruction set includes at least one operation and maintenance instruction, the operation and maintenance instruction being suitable for indicating a platform operation and maintenance strategy, and the first operation and maintenance instruction being determined by at least some of the operation and maintenance instructions in the preset operation and maintenance instruction set that are associated with the instruction characteristic that matches the operation data for the corresponding time period;

[0037] a simulation module, configured to obtain simulated operation and maintenance information of each of the plurality of time periods based on the first operation and maintenance instruction;

[0038] a reinforcement learning module, configured to perform reinforcement learning on the large operation and maintenance model based on the simulated operation and maintenance information to obtain systematic deviations corresponding to each of the multiple time periods, wherein the systematic deviation corresponding to each time period represents a systematic error of the target platform with respect to the operation and maintenance evaluation indicator under the action of the first operation and maintenance instruction corresponding to the time period;

[0039] The second operation and maintenance instruction module is used to generate a second operation and maintenance instruction for the target platform using the target large model based on the systematic deviations and operation data corresponding to each of the multiple time periods, and send the second operation and maintenance instruction to the target platform.

[0040] In summary, the embodiments of the present application have at least the following beneficial effects:

[0041] According to an embodiment of the present application, operating data of a target platform in multiple time periods is obtained; based on at least one preset operation and maintenance instruction set and the operation data, a first operation and maintenance instruction corresponding to each of the multiple time periods is determined using an operation and maintenance large model, wherein the at least one preset operation and maintenance instruction set is associated with at least one instruction feature, the preset operation and maintenance instruction set includes at least one operation and maintenance instruction, the operation and maintenance instruction is suitable for indicating a platform operation and maintenance strategy, and the first operation and maintenance instruction is determined by at least part of the operation and maintenance instructions in the preset operation and maintenance instruction set associated with the instruction feature that matches the operation data of the corresponding time period; based on the first operation and maintenance instruction, simulated operation and maintenance information for each of the multiple time periods is obtained; reinforcement learning is performed on the operation and maintenance large model based on the simulated operation and maintenance information to obtain systematic deviations corresponding to each of the multiple time periods, wherein the systematic deviation corresponding to each time period represents the systematic error of the target platform with respect to the operation and maintenance evaluation indicator under the action of the first operation and maintenance instruction corresponding to the time period; based on the systematic deviations and operation data corresponding to each of the multiple time periods, a second operation and maintenance instruction for the target platform is generated using the target large model, and the second operation and maintenance instruction is sent to the target platform, thereby improving the automation efficiency of operation and maintenance and the rationality of the operation and maintenance plan. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] Figure 1 This is a flowchart of the artificial intelligence operation and maintenance method based on a large model provided in an embodiment of the present application;

[0043] Figure 2 This is a schematic diagram of the structure of the artificial intelligence operation and maintenance system based on the large model provided in the embodiment of the present application;

[0044] Figure 3 It is a structural diagram of the computer device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0045] The following will be combined with the drawings in the embodiments of this application to clearly and completely describe the technical solutions in the embodiments of this application. Obviously, the embodiments described are only part of the embodiments of this application, not all of the embodiments. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of this application.

[0046] In the description of this application, the terms "first", "second", "third", etc. are used for descriptive purposes only and are not to be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Thus, features defined as "first", "second", "third", etc. may explicitly or implicitly include one or more of such features. In the description of this application, unless otherwise specified, "multiple" means two or more. In the description of this application, the term "including" and its variations are open inclusions, i.e., "including but not limited to". The term "based on" means "at least partially based on". The term "according to" means "at least partially based on". The term "one embodiment" means "at least one embodiment"; the term "another embodiment" means "at least one other embodiment"; the term "some embodiments" means "at least some embodiments".

[0047] In the description of this application, it should be noted that, unless otherwise expressly specified or limited, the terms "mounted," "connected," and "connected" should be understood in a broad sense. For example, they can refer to fixed connections, detachable connections, or integral connections; mechanical connections or electrical connections; direct connections or indirect connections through an intermediate medium; and internal connections between two components. Those skilled in the art will understand the specific meanings of the above terms in this application based on the specific circumstances.

[0048] In the description of this application, it should be noted that, unless otherwise defined, all technical and scientific terms used in this application have the same meanings as those commonly understood by those skilled in the art. The terms used in this specification are only for the purpose of describing specific embodiments and are not intended to limit this application. Those of ordinary skill in the art will understand the specific meanings of the above terms in this application in specific circumstances.

[0049] First, see Figure 1 , shows a flow chart of an artificial intelligence operation and maintenance method based on a large model provided by an embodiment of the present application, the method includes steps S101-S105, which are specifically as follows:

[0050] S101, obtaining the operating data of the target platform in multiple time periods.

[0051] In one example, the multiple time periods may be multiple consecutive time periods.

[0052] In one example, obtaining the target platform's operating data for multiple time periods may include: obtaining continuous operating data of the target platform over a preset period of time in the past according to a target period, dividing the continuous operating data into multiple sub-data in a time sequence, and selecting the operating data for the multiple time periods from the divided sub-data. The target period may be preset or dynamically adjusted based on the last second operation and maintenance instruction.

[0053] S102, based on at least one preset operation and maintenance instruction set and the operation data, use the operation and maintenance big model to determine the first operation and maintenance instruction corresponding to each of the multiple time periods, wherein the at least one preset operation and maintenance instruction set is associated with at least one instruction feature, the preset operation and maintenance instruction set includes at least one operation and maintenance instruction, the operation and maintenance instruction is suitable for indicating the platform operation and maintenance strategy, and the first operation and maintenance instruction is determined by at least part of the operation and maintenance instructions in the preset operation and maintenance instruction set associated with the instruction feature that matches the operation data of the corresponding time period.

[0054] In one example, the aforementioned large operation and maintenance model may be a pre-trained large language model, such as a Deepseek model.

[0055] In one example, the above operation and maintenance model may include:

[0056] The input layer is used to receive sample operation data from the platform, including but not limited to system performance indicators (such as CPU utilization, memory usage, etc.), log files, network traffic, etc., and use this data as input features of the model;

[0057] The feature extraction layer is used to process the time series data in the sample running data using convolutional neural networks, recurrent neural networks, long short-term memory networks, or transformer architectures in deep learning to extract key features from it so that the model can understand complex patterns and dependencies;

[0058] The decision layer is used to determine and output appropriate operation and maintenance instructions based on the features extracted by the feature extraction layer using a common supervised learning algorithm.

[0059] The output layer is used to generate a first operation and maintenance instruction for a specific time period based on the output of the decision layer. The first operation and maintenance instruction can be used to indicate at least one of the following: resource allocation adjustment, service restart, load balancing strategy change, etc.

[0060] In one example, for each preset operation and maintenance instruction set, instructions that meet the screening conditions in the second operation and maintenance instructions corresponding to the instruction characteristics associated with the preset operation and maintenance instruction set, determined in the process of generating the second operation and maintenance instructions for the target platform, can be added to the preset operation and maintenance instruction set to complete the update of the preset operation and maintenance instruction set, and the updated preset operation and maintenance instruction set is used as the current preset operation and maintenance instruction set. The systematic deviation corresponding to the second operation and maintenance instruction can be determined by the embodiment of the present application related to determining the corresponding systematic deviation using the first operation and maintenance instruction. In this way, the screening condition can include that the systematic deviation corresponding to the second operation and maintenance instruction is less than the deviation threshold.

[0061] In one example, the above instruction features can be obtained by clustering the features of the operation and maintenance instructions included in the corresponding preset operation and maintenance instruction set.

[0062] In one example, at least one preset operation and maintenance instruction set is associated one-to-one with at least one instruction feature.

[0063] In one example, the platform operation and maintenance strategy may include at least one of the following:

[0064] Performance optimization strategies, such as adjusting resource configuration (such as CPU and memory allocation), database optimization, and cache strategy adjustments, can improve the overall performance of the platform.

[0065] Security protection strategies, such as firewall configuration modification and intrusion detection settings, can enhance the security of the platform system and prevent potential security threats.

[0066] Data backup and recovery strategies can, for example, instruct immediate data backup and provide a set of data recovery instructions when necessary to protect important data from loss.

[0067] Resource management strategies, such as monitoring and managing the usage of platform computing resources, ensure the effective use of resources while avoiding service quality degradation caused by excessive resource consumption.

[0068] Version upgrade and maintenance strategies, such as developing software and hardware upgrade plans to ensure that the platform technology stack remains up to date and can introduce the latest features.

[0069] Log management policies, such as how to collect, store, and analyze platform system logs, to facilitate tracing the root causes of problems and improving system stability.

[0070] It is understandable that, through the above platform operation and maintenance strategy, the operation and maintenance big model and / or the target big model can be prompted to match the corresponding strategy according to the operation data, so as to improve the reliability and rationality of the corresponding operation and maintenance instructions generated.

[0071] S103: Based on the first operation and maintenance instruction, obtain simulated operation and maintenance information of each of the multiple time periods.

[0072] In one example, the operation and maintenance simulation software that has been pre-set according to the relevant parameters of the target platform can be used to simulate the first operation and maintenance instructions of each time period to obtain corresponding simulated operation and maintenance information. The simulated operation and maintenance information can be used to indicate at least one of the following items of the target platform under the action of the corresponding first operation and maintenance instructions: platform system performance, resource utilization, operation and maintenance costs (such as time and / or resource consumption, etc.). Among them, the operation and maintenance simulation software may include at least one of the following items: Apache JMeter for e-commerce platforms (which can be used for load testing and performance measurement to help evaluate the performance of e-commerce platforms under high concurrency conditions), Simulacron for financial systems (a simulator of a Cassandra database that can be used to test database operations in financial systems), and IoT-LAB for IoT devices (which can be used to provide real hardware resources and simulation tools that can be used to test IoT applications and services, including protocol stacks, energy consumption analysis, etc.).

[0073] S104, performing reinforcement learning on the operation and maintenance large model based on the simulated operation and maintenance information to obtain the systematic deviations corresponding to each of the multiple time periods, wherein the systematic deviation corresponding to each time period represents the systematic error of the target platform regarding the operation and maintenance evaluation index under the action of the first operation and maintenance instruction corresponding to the time period.

[0074] In one example, under the reinforcement learning framework, the operation and maintenance large model can adjust its own model parameters based on the simulated operation and maintenance information. The adjustment direction / adjustment purpose can be to minimize the difference between the operation and maintenance evaluation indicators and the expected goals. In this process, the operation and maintenance large model can try to find the optimal operation and maintenance strategy so that the actual system performance of the target platform is as close to the ideal state as possible. For each time period, the systematic deviation corresponding to the time period can be calculated based on the simulated operation and maintenance information and the learning results of the operation and maintenance large model. Here, the systematic deviation can characterize the systematic error of the target platform regarding the operation and maintenance evaluation indicators under the action of the first operation and maintenance instruction corresponding to the time period. In other words, the systematic deviation can be a measure of the gap between the actual operation and maintenance effect and the expected value. At this time, the systematic deviation can be obtained by calculating the measurement value used to measure the gap (such as similarity, norm, etc.).

[0075] In one example, for reinforcement learning algorithms, policy gradient methods, DQN (Deep Q-Networks), and other technologies can be used to optimize long-term rewards. This reward can then be used to find operational instructions that match operational evaluation metrics (such as maximizing system performance and / or minimizing failure rate, which can be achieved by treating different objectives as different sub-goals of the optimization algorithm in the "and" relationship). During the reinforcement learning phase, the operational model can use simulated operational information to observe feedback and adjust its strategies based on the observed feedback (such as system stability, performance improvement, etc.) to achieve better long-term results.

[0076] S105 , based on the systematic deviations and operation data corresponding to each of the multiple time periods, generate a second operation and maintenance instruction for the target platform using the target macro model, and send the second operation and maintenance instruction to the target platform.

[0077] In one example, the target large model can be a pre-trained large language model. Its related structure can refer to the embodiments related to the operation and maintenance large model in this application, and will not be repeated here. However, it should be understood that the target large model can be an independent large language model.

[0078] It can be understood that the embodiment of the present application first obtains the operating data of the target platform in multiple time periods to ensure a comprehensive understanding of the actual status of the platform system in the target time period (for example, in a certain historical period in the past, which can be used to assist in analyzing the most appropriate second operation and maintenance instructions in a subsequent period of time); through the operation and maintenance big model, the first operation and maintenance instruction is determined in combination with the preset operation and maintenance instruction set and the operation data, so that the complex / typical factors represented by the preset operation and maintenance instruction set can be taken into account, so as to generate a more reasonable and effective operation and maintenance strategy; then, based on the first operation and maintenance instruction, simulated operation and maintenance information is obtained, and based on this, the operation and maintenance big model is reinforced to obtain systematic deviations to evaluate the effectiveness of the decision corresponding to each first operation and maintenance instruction, and identify systematic deviations, which can be used to represent potential problems; finally, based on the systematic deviations and operation data, the second operation and maintenance instruction is generated using the target big model, so that the target big model can optimize the operation and maintenance strategy on the basis of fully considering the systematic deviations (such as the potential problems it represents) to reduce the possibility of wrong decisions, thereby improving the rationality of the operation and maintenance plan, and the embodiment of the present application can also reduce manual participation in the operation and maintenance process to improve automation efficiency, therefore, AIOps (Artificial Intelligence) can be realized. forIT Operations, artificial intelligence operation and maintenance) functions.

[0079] In an optional embodiment, the performing reinforcement learning on the large operation and maintenance model based on the simulated operation and maintenance information includes:

[0080] Selecting at least part of the time periods from the plurality of time periods as enhanced time periods;

[0081] For each enhancement period, performing enhancement processing on the simulated operation and maintenance information of the enhancement period to obtain enhanced operation and maintenance information of the enhancement period;

[0082] Based on the enhanced operation and maintenance information and the simulated operation and maintenance information of each of the multiple time periods, reinforcement learning is performed on the operation and maintenance large model.

[0083] In this embodiment, enhancement processing may be performed on the simulated operation and maintenance information of the enhancement period to increase the amount of operation and maintenance information of the enhancement period, thereby improving the reliability and accuracy of the systematic deviation corresponding to the relevant enhancement period obtained through reinforcement learning.

[0084] In one example, the above-mentioned enhancement processing may refer to expanding the amount of information of the simulated operation and maintenance information, for example, using multiple simulation methods to generate corresponding simulated operation and maintenance information according to the first operation and maintenance instructions corresponding to the enhancement period, and integrating the simulated operation and maintenance information generated by different simulation methods to obtain enhanced operation and maintenance information for the enhancement period.

[0085] In an optional implementation, performing enhancement processing on the simulated operation and maintenance information of the enhancement period to obtain enhanced operation and maintenance information of the enhancement period includes:

[0086] Performing disturbance processing on the simulated operation and maintenance information during the enhancement period to obtain disturbed operation and maintenance information;

[0087] Randomly adding noise to the disturbed operation and maintenance information to obtain noisy operation and maintenance information;

[0088] Determining a first probability distribution of the noisy operation and maintenance information, and determining a second probability distribution of the simulated operation and maintenance information during the enhancement period;

[0089] Based on the first probability distribution and the second probability distribution, the noisy operation and maintenance information is converted into enhanced operation and maintenance information for the enhancement period.

[0090] In this embodiment, the simulated operation and maintenance information can be first disturbed and randomly noised in sequence so that the simulated operation and maintenance information has a large amount of interference and noise, so as to increase the amount of information; then, the noisy operation and maintenance information can be converted into enhanced operation and maintenance information of the enhanced period through the comparison results / difference between the first probability distribution and the second probability distribution.

[0091] In one example, the first probability distribution may include a first cross-attention distribution, and the second probability distribution may include a second cross-attention distribution.

[0092] In an example, if the noise added by the random noising is Gaussian noise, the first probability distribution may include a first Gaussian distribution, and the second probability distribution may include a second Gaussian distribution.

[0093] In one example, both the first probability distribution and the second probability distribution can be obtained using kernel density estimation.

[0094] In one example, based on the first probability distribution and the second probability distribution, converting the noisy operation and maintenance information into the enhanced operation and maintenance information of the enhanced time period may include: inputting the noisy operation and maintenance information, the first probability distribution and the second probability distribution into the target large model, so that the target large model can convert the noisy operation and maintenance information into the enhanced operation and maintenance information of the enhanced time period under the prompt of the difference between the first probability distribution and the second probability distribution. Among them, the relevant description of the target large model can refer to the corresponding embodiment of the present application and will not be repeated here. In addition, it should be noted that it is not difficult to understand in this field that if the target large model wants to realize the functions described in this embodiment, it is only necessary to perform corresponding fine-tuning on the pre-trained large model.

[0095] In an optional implementation, converting the noisy operation and maintenance information into enhanced operation and maintenance information for the enhancement period based on the first probability distribution and the second probability distribution includes:

[0096] Determining a target noise addition strategy by analyzing a difference between the first probability distribution and the second probability distribution;

[0097] Noise processing is performed on the noisy operation and maintenance information according to the target noising strategy to obtain enhanced operation and maintenance information for the enhancement period.

[0098] In this embodiment, the target noise addition strategy can adapt to the difference between the first probability distribution and the second probability distribution, so that the subsequent noise addition processing can be targeted based on the difference. Noise addition can complete the amplification of the amount of information by adding noise, and the added noise can adapt to the difference, so that the probability distribution of the enhanced operation and maintenance information obtained by noise addition can also converge to the second probability distribution of the simulated operation and maintenance information (that is, the target noise addition strategy can be used to indicate that the probability distribution of the enhanced operation and maintenance information converges to the second probability distribution of the simulated operation and maintenance information). In this way, in addition to the amplification of the amount of information, this embodiment can also improve the adaptability of the large operation and maintenance model after reinforcement learning to interference and / or noise, so that when it is used next time, the large operation and maintenance model with stronger adaptability to interference and / or noise can be called.

[0099] In one example, the above-mentioned target noise addition strategy can be used to add target noise to the noisy operation and maintenance information based on the difference between the first probability distribution and the second probability distribution, so that the probability distribution of the enhanced operation and maintenance information converges to the second probability distribution of the simulated operation and maintenance information. For example, for a certain information area, if the difference characterizes that the gap between the two probability distributions in the information area is small, then the added target noise needs to be added to the information area. More noise can be added. Correspondingly, in another information area, if the difference characterizes that the gap between the two probability distributions in the information area is large, then the added target noise needs to be added to the information area. Less noise can be added. In this way, the gap between the two probability distributions in different information areas can be brought to a similar level. Since the probability distribution is a relative quantity that represents the probability of each place, the probability distribution of the enhanced operation and maintenance information can actually converge to the second probability distribution of the simulated operation and maintenance information at this time.

[0100] In an optional embodiment, the target large model is obtained by fine-tuning the operation and maintenance large model through reinforcement learning.

[0101] In an optional implementation, obtaining the simulated operation and maintenance information of each of the multiple time periods based on the first operation and maintenance instruction includes:

[0102] For the first operation and maintenance instruction corresponding to each time period, simulation is performed based on the first operation and maintenance instruction to simulate the operation status information of the target platform under the action of the first operation and maintenance instruction, and simulated operation and maintenance information of the time period is generated based on the operation status information.

[0103] In one example, simulation can be performed using operation and maintenance simulation software that has been pre-configured according to the relevant parameters of the target platform. The operation and maintenance simulation software may include at least one of the following: Apache JMeter for e-commerce platforms (which can be used for load testing and performance measurement to help evaluate the performance of e-commerce platforms under high concurrency conditions), Simulacron for financial systems (a Cassandra database simulator that can be used to test database operations in financial systems), and IoT-LAB for IoT devices (which can be used to provide real hardware resources and simulation tools for testing IoT applications and services, including protocol stacks and energy consumption analysis).

[0104] In an optional embodiment, during the reinforcement learning process, the learning weight of the operation and maintenance model for the simulated operation and maintenance information corresponding to each time period is determined by the operating status information corresponding to the time period.

[0105] In one example, the learning weight may include at least one of the following: a step size (ie, a learning rate) used when updating model parameters, and an importance weight assigned to different simulated operation and maintenance information.

[0106] In one example, the operating status information may include at least one of the following:

[0107] Performance metrics, such as CPU utilization, memory usage (including physical memory usage and virtual memory usage), disk I / O (which measures the speed and frequency of read and write operations), and / or network traffic (which indicates the number and speed of data packets entering and leaving the platform system);

[0108] Resource availability, which may include, for example, service availability (indicating whether a particular service is online and accessible) and / or bandwidth utilization (indicating the ratio of the maximum capacity of a network connection to its current usage);

[0109] Application-specific metrics (which may vary depending on the platform) may include, for example, transaction success rate and / or latency for financial systems; and sensor data accuracy and / or response time for IoT devices.

[0110] In an optional embodiment, the determining, based on at least one preset operation and maintenance instruction set and the operation data, using the operation and maintenance macro model to determine the first operation and maintenance instructions corresponding to each of the multiple time periods includes:

[0111] Inputting the at least one preset operation and maintenance instruction set and the operation data of the multiple time periods into the operation and maintenance macro model;

[0112] For each time period, the operation and maintenance big model generates a corresponding first operation and maintenance instruction based on the operation data of the time period, under the prompt of at least part of the operation and maintenance instructions in the preset operation and maintenance instruction set associated with the instruction characteristics matching the operation data of the time period.

[0113] In an optional embodiment, generating the second operation and maintenance instruction of the target platform using the target macro model based on the systematic deviations and operation data corresponding to each of the multiple time periods includes:

[0114] Inputting the systematic deviations and operating data corresponding to each of the plurality of time periods into the target macro model;

[0115] The target large model determines the time period weight corresponding to each time period according to the systematic deviation corresponding to each time period;

[0116] The target large model generates the second operation and maintenance instruction based on the time period weights and operation data corresponding to each of the multiple time periods.

[0117] On the second aspect, accordingly, the embodiments of the present application also provide an artificial intelligence operation and maintenance system based on a large model, which can implement all the processes of the artificial intelligence operation and maintenance method based on a large model provided in the above embodiments.

[0118] See also Figure 2 , shows a schematic diagram of the structure of an artificial intelligence operation and maintenance system based on a large model provided in an embodiment of the present application, the artificial intelligence operation and maintenance system based on a large model includes:

[0119] The data acquisition module 201 is used to acquire the operation data of the target platform in multiple time periods;

[0120] A first operation and maintenance instruction module 202 is configured to determine, based on at least one preset operation and maintenance instruction set and the operation data, a first operation and maintenance instruction corresponding to each of the multiple time periods using the operation and maintenance macro model, wherein the at least one preset operation and maintenance instruction set is associated with at least one instruction characteristic, the preset operation and maintenance instruction set includes at least one operation and maintenance instruction, the operation and maintenance instruction being suitable for indicating a platform operation and maintenance strategy, and the first operation and maintenance instruction being determined by at least some of the operation and maintenance instructions in the preset operation and maintenance instruction set associated with the instruction characteristic matching the operation data of the corresponding time period;

[0121] A simulation module 203 is configured to obtain simulated operation and maintenance information for each of the plurality of time periods based on the first operation and maintenance instruction;

[0122] A reinforcement learning module 204 is configured to perform reinforcement learning on the large operation and maintenance model based on the simulated operation and maintenance information to obtain systematic deviations corresponding to each of the multiple time periods, wherein the systematic deviation corresponding to each time period represents a systematic error of the target platform with respect to the operation and maintenance evaluation indicator under the first operation and maintenance instruction corresponding to the time period;

[0123] The second operation and maintenance instruction module 205 is used to generate a second operation and maintenance instruction for the target platform using the target large model based on the systematic deviations and operation data corresponding to each of the multiple time periods, and send the second operation and maintenance instruction to the target platform.

[0124] In an optional embodiment, the performing reinforcement learning on the large operation and maintenance model based on the simulated operation and maintenance information includes:

[0125] Selecting at least part of the time periods from the plurality of time periods as enhanced time periods;

[0126] For each enhancement period, performing enhancement processing on the simulated operation and maintenance information of the enhancement period to obtain enhanced operation and maintenance information of the enhancement period;

[0127] Based on the enhanced operation and maintenance information and the simulated operation and maintenance information of each of the multiple time periods, reinforcement learning is performed on the operation and maintenance large model.

[0128] In an optional implementation, performing enhancement processing on the simulated operation and maintenance information of the enhancement period to obtain enhanced operation and maintenance information of the enhancement period includes:

[0129] Performing disturbance processing on the simulated operation and maintenance information during the enhancement period to obtain disturbed operation and maintenance information;

[0130] Randomly adding noise to the disturbed operation and maintenance information to obtain noisy operation and maintenance information;

[0131] Determining a first probability distribution of the noisy operation and maintenance information, and determining a second probability distribution of the simulated operation and maintenance information during the enhancement period;

[0132] Based on the first probability distribution and the second probability distribution, the noisy operation and maintenance information is converted into enhanced operation and maintenance information for the enhancement period.

[0133] In an optional implementation, converting the noisy operation and maintenance information into enhanced operation and maintenance information for the enhancement period based on the first probability distribution and the second probability distribution includes:

[0134] Determining a target noise addition strategy by analyzing a difference between the first probability distribution and the second probability distribution;

[0135] Noise processing is performed on the noisy operation and maintenance information according to the target noising strategy to obtain enhanced operation and maintenance information for the enhancement period.

[0136] In an optional embodiment, the target large model is obtained by fine-tuning the operation and maintenance large model through reinforcement learning.

[0137] In an optional implementation, obtaining the simulated operation and maintenance information of each of the multiple time periods based on the first operation and maintenance instruction includes:

[0138] For the first operation and maintenance instruction corresponding to each time period, simulation is performed based on the first operation and maintenance instruction to simulate the operation status information of the target platform under the action of the first operation and maintenance instruction, and simulated operation and maintenance information of the time period is generated based on the operation status information.

[0139] In an optional embodiment, during the reinforcement learning process, the learning weight of the operation and maintenance model for the simulated operation and maintenance information corresponding to each time period is determined by the operating status information corresponding to the time period.

[0140] In an optional embodiment, the determining, based on at least one preset operation and maintenance instruction set and the operation data, of the first operation and maintenance instructions corresponding to each of the plurality of time periods using the operation and maintenance macro model includes:

[0141] Inputting the at least one preset operation and maintenance instruction set and the operation data of the multiple time periods into the operation and maintenance macro model;

[0142] For each time period, the operation and maintenance big model generates a corresponding first operation and maintenance instruction based on the operation data of the time period, under the prompt of at least part of the operation and maintenance instructions in the preset operation and maintenance instruction set associated with the instruction characteristics matching the operation data of the time period.

[0143] In an optional embodiment, generating the second operation and maintenance instruction of the target platform using the target macro model based on the systematic deviations and operation data corresponding to each of the multiple time periods includes:

[0144] Inputting the systematic deviations and operating data corresponding to each of the plurality of time periods into the target macro model;

[0145] The target large model determines the time period weight corresponding to each time period according to the systematic deviation corresponding to each time period;

[0146] The target large model generates the second operation and maintenance instruction based on the time period weights and operation data corresponding to each of the multiple time periods.

[0147] In a third aspect, an embodiment of the present application provides a computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, the steps of the large model-based artificial intelligence operation and maintenance method described above are implemented.

[0148] In a fourth aspect, an embodiment of the present application provides a computer program product, comprising computer instructions, which, when executed by a processor, implement the steps of any of the above-mentioned large-model-based artificial intelligence operation and maintenance methods.

[0149] In a fifth aspect, an embodiment of the present application provides a computer device comprising a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, wherein when the processor executes the computer program, the steps of the large model-based artificial intelligence operation and maintenance method described above are implemented.

[0150] See also Figure 3 The computer device of this embodiment includes: a processor 301, a memory 302, and a computer program stored in the memory 302 and executable on the processor 301, such as an artificial intelligence operation and maintenance program based on a large model. When the processor 301 executes the computer program, the steps in each of the above-mentioned embodiments of the artificial intelligence operation and maintenance method based on a large model are implemented, such as Figure 1 Steps S101-S105 are shown.

[0151] For example, the computer program may be divided into one or more modules / units, which are stored in the memory 302 and executed by the processor 301 to implement the present application. The one or more modules / units may be a series of computer program instruction segments capable of implementing specific functions, and the instruction segments are used to describe the execution process of the computer program in the computer device.

[0152] The computer device may be a desktop computer, laptop, PDA, cloud server, or other computing device. The computer device may include, but is not limited to, a processor 301 and a memory 302. Those skilled in the art will appreciate that the schematic diagram is merely an example of a computer device and does not limit the computer device. The computer device may include more or fewer components than shown, or a combination of certain components, or different components. For example, the computer device may also include input and output devices, network access devices, buses, and the like.

[0153] The processor 301 may be a central processing unit (CPU), other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA), other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor, or the processor 301 may be any conventional processor. The processor 301 is the control center of the computer device, connecting various parts of the entire computer device using various interfaces and lines.

[0154] The memory 302 can be used to store the computer programs and / or modules. The processor 301 implements various functions of the computer device by running or executing the computer programs and / or modules stored in the memory 302 and calling the data stored in the memory 302. The memory 302 can mainly include a program storage area and a data storage area. The program storage area can store an operating system and at least one application required for a function (such as a sound playback function, an image playback function, etc.); the data storage area can store data created based on the use of the mobile phone (such as audio data, a phone book, etc.). In addition, the memory 302 can include high-speed random access memory and non-volatile memory, such as a hard disk, internal memory, a plug-in hard disk, a smart media card (SMC), a secure digital (SD) card, a flash card, at least one disk storage device, a flash memory device, or other volatile solid-state storage device.

[0155] Wherein, if the module / unit integrated in the computer device is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the present application implements all or part of the process in the above-mentioned embodiment method, and can also be completed by instructing the relevant hardware through a computer program. The computer program can be stored in a computer-readable storage medium, and the computer program, when executed by the processor 301, can implement the steps of the above-mentioned various method embodiments. Wherein, the computer program includes computer program code, and the computer program code can be in source code form, object code form, executable file or some intermediate form. The computer-readable medium may include: any entity or device capable of carrying the computer program code, recording medium, U disk, mobile hard disk, magnetic disk, optical disk, computer memory, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), electrical carrier signal, telecommunication signal and software distribution medium, etc.

[0156] In summary, the embodiments of the present application have at least the following beneficial effects:

[0157] According to an embodiment of the present application, operating data of a target platform in multiple time periods is obtained; based on at least one preset operation and maintenance instruction set and the operation data, a first operation and maintenance instruction corresponding to each of the multiple time periods is determined using an operation and maintenance large model, wherein the at least one preset operation and maintenance instruction set is associated with at least one instruction feature, the preset operation and maintenance instruction set includes at least one operation and maintenance instruction, the operation and maintenance instruction is suitable for indicating a platform operation and maintenance strategy, and the first operation and maintenance instruction is determined by at least part of the operation and maintenance instructions in the preset operation and maintenance instruction set associated with the instruction feature that matches the operation data of the corresponding time period; based on the first operation and maintenance instruction, simulated operation and maintenance information for each of the multiple time periods is obtained; reinforcement learning is performed on the operation and maintenance large model based on the simulated operation and maintenance information to obtain systematic deviations corresponding to each of the multiple time periods, wherein the systematic deviation corresponding to each time period represents the systematic error of the target platform with respect to the operation and maintenance evaluation indicator under the action of the first operation and maintenance instruction corresponding to the time period; based on the systematic deviations and operation data corresponding to each of the multiple time periods, a second operation and maintenance instruction for the target platform is generated using the target large model, and the second operation and maintenance instruction is sent to the target platform, thereby improving the automation efficiency of operation and maintenance and the rationality of the operation and maintenance plan.

[0158] Through the description of the above implementation methods, those skilled in the art can clearly understand that the present application can be implemented by means of software plus the necessary hardware platform, and of course it can also be implemented entirely by hardware. Based on this understanding, all or part of the contribution of the technical solution of the present application to the background technology can be embodied in the form of a software product, and the computer software product can be stored in a storage medium, such as ROM (Read-Only Memory) / RAM (Random Access Memory), a disk, an optical disk, etc., and includes a number of instructions for enabling a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment of the present application or certain parts of the embodiments.

[0159] The above is a preferred embodiment of the present application. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present application. These improvements and modifications are also considered to be within the scope of protection of the present application.

Claims

1. An artificial intelligence operation and maintenance method based on a large model, characterized in that: include: Obtain the target platform's operating data over multiple time periods; Based on at least one preset operation and maintenance instruction set and the operation data, determining a first operation and maintenance instruction corresponding to each of the multiple time periods using the operation and maintenance macro model, wherein the at least one preset operation and maintenance instruction set is associated with at least one instruction characteristic, the preset operation and maintenance instruction set includes at least one operation and maintenance instruction, the operation and maintenance instruction being suitable for indicating a platform operation and maintenance policy, and the first operation and maintenance instruction being determined by at least some of the operation and maintenance instructions in the preset operation and maintenance instruction set associated with the instruction characteristic that matches the operation data for the corresponding time period; Based on the first operation and maintenance instruction, obtaining simulated operation and maintenance information for each of the multiple time periods; Reinforcement learning is performed on the large operation and maintenance model based on the simulated operation and maintenance information to obtain systematic deviations corresponding to each of the multiple time periods, wherein the systematic deviation corresponding to each time period represents a systematic error of the target platform with respect to the operation and maintenance evaluation indicator under the action of the first operation and maintenance instruction corresponding to the time period; Based on the systematic deviations and operation data corresponding to each of the multiple time periods, a second operation and maintenance instruction for the target platform is generated using the target large model, and the second operation and maintenance instruction is sent to the target platform.

2. The method according to claim 1, characterized in that The performing reinforcement learning on the large operation and maintenance model based on the simulated operation and maintenance information includes: Selecting at least part of the time periods from the plurality of time periods as enhanced time periods; For each enhancement period, performing enhancement processing on the simulated operation and maintenance information of the enhancement period to obtain enhanced operation and maintenance information of the enhancement period; Based on the enhanced operation and maintenance information and the simulated operation and maintenance information of each of the multiple time periods, reinforcement learning is performed on the operation and maintenance large model.

3. The method according to claim 2, characterized in that The enhancing the simulated operation and maintenance information of the enhanced period to obtain the enhanced operation and maintenance information of the enhanced period includes: Performing disturbance processing on the simulated operation and maintenance information during the enhancement period to obtain disturbed operation and maintenance information; Randomly adding noise to the disturbed operation and maintenance information to obtain noisy operation and maintenance information; Determining a first probability distribution of the noisy operation and maintenance information, and determining a second probability distribution of the simulated operation and maintenance information during the enhancement period; Based on the first probability distribution and the second probability distribution, the noisy operation and maintenance information is converted into enhanced operation and maintenance information for the enhancement period.

4. The method according to claim 3, characterized in that The converting the noisy operation and maintenance information into enhanced operation and maintenance information for the enhancement period based on the first probability distribution and the second probability distribution includes: Determining a target noise addition strategy by analyzing a difference between the first probability distribution and the second probability distribution; Noise processing is performed on the noisy operation and maintenance information according to the target noising strategy to obtain enhanced operation and maintenance information for the enhancement period.

5. The method according to claim 1, wherein The target large model is obtained by fine-tuning the operation and maintenance large model through reinforcement learning.

6. The method according to claim 1, characterized in that The acquiring, based on the first operation and maintenance instruction, simulated operation and maintenance information of each of the multiple time periods includes: For the first operation and maintenance instruction corresponding to each time period, simulation is performed based on the first operation and maintenance instruction to simulate the operation status information of the target platform under the action of the first operation and maintenance instruction, and simulated operation and maintenance information of the time period is generated based on the operation status information.

7. The method according to claim 6, characterized in that In the process of reinforcement learning, the learning weight of the operation and maintenance model for the simulated operation and maintenance information corresponding to each time period is determined by the operating status information corresponding to the time period.

8. The method according to claim 1, characterized in that The determining, based on at least one preset operation and maintenance instruction set and the operation data, of first operation and maintenance instructions corresponding to each of the plurality of time periods using the operation and maintenance macro model includes: Inputting the at least one preset operation and maintenance instruction set and the operation data of the multiple time periods into the operation and maintenance macro model; For each time period, the operation and maintenance big model generates a corresponding first operation and maintenance instruction based on the operation data of the time period, under the prompt of at least part of the operation and maintenance instructions in the preset operation and maintenance instruction set associated with the instruction characteristics matching the operation data of the time period.

9. The method according to claim 1, characterized in that The generating of the second operation and maintenance instruction of the target platform by using the target macro model based on the systematic deviations and operation data corresponding to each of the multiple time periods includes: Inputting the systematic deviations and operating data corresponding to each of the plurality of time periods into the target macro model; The target large model determines the time period weight corresponding to each time period according to the systematic deviation corresponding to each time period; The target large model generates the second operation and maintenance instruction based on the time period weights and operation data corresponding to each of the multiple time periods.

10. An artificial intelligence operation and maintenance system based on a large model, characterized in that: include: The data acquisition module is used to obtain the operating data of the target platform in multiple time periods; a first operation and maintenance instruction module, configured to determine, based on at least one preset operation and maintenance instruction set and the operation data, a first operation and maintenance instruction corresponding to each of the plurality of time periods using the operation and maintenance macro model, wherein the at least one preset operation and maintenance instruction set is associated with at least one instruction characteristic, the preset operation and maintenance instruction set includes at least one operation and maintenance instruction, the operation and maintenance instruction being suitable for indicating a platform operation and maintenance strategy, and the first operation and maintenance instruction being determined by at least some of the operation and maintenance instructions in the preset operation and maintenance instruction set that are associated with the instruction characteristic that matches the operation data for the corresponding time period; a simulation module, configured to obtain simulated operation and maintenance information of each of the plurality of time periods based on the first operation and maintenance instruction; a reinforcement learning module, configured to perform reinforcement learning on the large operation and maintenance model based on the simulated operation and maintenance information to obtain systematic deviations corresponding to each of the multiple time periods, wherein the systematic deviation corresponding to each time period represents a systematic error of the target platform with respect to the operation and maintenance evaluation indicator under the action of the first operation and maintenance instruction corresponding to the time period; The second operation and maintenance instruction module is used to generate a second operation and maintenance instruction for the target platform using the target large model based on the systematic deviations and operation data corresponding to each of the multiple time periods, and send the second operation and maintenance instruction to the target platform.

Images