Solution verification method and device based on large model, equipment and medium

Abstract

The invention provides a solution verification method and device based on a large model, equipment and a medium, and relates to the technical field of artificial intelligence. Comprising the steps that scheme attribute information is extracted from a to-be-audited solution of a target problem according to a target format, and the scheme attribute information at least comprises problem description, root cause analysis, triggering conditions and avoidance measures; acquiring a target cloud platform environment matched with the trigger condition, simulating a target problem in the target cloud platform environment according to the problem description, and outputting an evaluation result about whether the avoidance measure can solve the target problem; if yes, aggregation and conversion are carried out on the avoidance measures, execution steps of the solutions are generated, and corresponding execution environments are obtained; the execution step and the execution environment are sent to an execution scheme program, and the execution scheme program is used for implementing the execution step in the execution environment and returning an obtained return value; and judging whether the solution can solve the target problem according to the return value. The verification efficiency of the solution can be improved.

Description

Technical Field

[0001] The present application relates to the field of artificial intelligence technology, and in particular to a method, apparatus, device and medium for verifying a solution based on a large model. Background Art

[0002] Currently, the verification process for cloud platform solutions relies primarily on manual effort. However, this manual approach requires not only specialized knowledge but also a comprehensive understanding of the cloud platform architecture, operating environment, and potential issues within the existing environment.

[0003] Therefore, the existing solution verification process has problems such as limited manual review and evaluation speed, difficulty in screening environmental information, and slow node selection and solution implementation. Summary of the Invention

[0004] The present application provides a large-model-based solution verification method, apparatus, device, and medium to solve the problem of low efficiency in the solution verification process in the prior art.

[0005] In a first aspect, the present application provides a solution verification method based on a large model, comprising:

[0006] Extracting solution attribute information from the pending solution to the target problem according to the target format of the solution, wherein the solution attribute information at least includes a problem description, a root cause analysis, a triggering condition, and a circumvention measure;

[0007] Acquire a target cloud platform environment that matches the trigger condition, simulate the target problem in the target cloud platform environment according to the problem description, and output an evaluation result of whether the avoidance measure can solve the target problem;

[0008] In response to the evaluation result being a passed evaluation, aggregating and transforming the avoidance measures, generating execution steps of the solution, and obtaining an execution environment corresponding to the execution steps;

[0009] Sending the execution steps and the execution environment to a pre-customized execution program, wherein the execution program is used to implement the execution steps in the execution environment and return a return value obtained by implementing the execution steps;

[0010] It is determined whether the solution can solve the target problem according to the return value.

[0011] In a second aspect, the present application provides a solution verification device based on a large model, comprising:

[0012] An information extraction module is configured to extract solution attribute information from a pending solution to a target problem according to a target format of the solution, wherein the solution attribute information includes at least a problem description, a root cause analysis, a trigger condition, and a circumvention measure;

[0013] An evaluation module is configured to obtain a target cloud platform environment that matches the trigger condition, simulate the target problem in the target cloud platform environment according to the problem description, and output an evaluation result of whether the circumvention measure can solve the target problem;

[0014] an execution step generation module for aggregating and transforming the avoidance measures in response to the evaluation result being a passed evaluation, generating execution steps for the solution, and obtaining an execution environment corresponding to the execution steps;

[0015] An execution step sending module, configured to send the execution step and the execution environment to a pre-customized execution solution program, wherein the execution solution program is configured to implement the execution step in the execution environment and return a return value obtained by implementing the execution step;

[0016] A verification module is used to determine whether the solution can solve the target problem based on the return value.

[0017] In a third aspect, the present application also provides an electronic device comprising a memory, a processor, and a computer program stored on the memory and runnable on the processor, wherein when the processor executes the program, it implements a large model-based solution verification method as described in any one of the embodiments of the present application.

[0018] In a fourth aspect, the present application also provides a computer-readable storage medium on which a computer program is stored, which, when executed by a processor, implements a large model-based solution verification method as described in any of the embodiments of the present application.

[0019] In a fifth aspect, the present application also provides a computer program product, including a computer program, which, when executed by a processor, implements the large model-based solution verification method as described in any one of the embodiments of the present application.

[0020] The solution verification method, device, equipment and medium based on the big model provided in this application realize the automatic extraction of solution attribute information based on the big model, obtain the matching target cloud platform environment according to the extracted attribute information, simulate the target problem, and then evaluate whether the avoidance measures can solve the target problem. After the evaluation is passed, the execution steps and execution environment can be automatically obtained, and the execution solution program can be called to implement the execution steps in the execution environment, and whether the solution can solve the target problem can be judged based on the return value of the program. Therefore, the solution is reviewed and evaluated based on the big model, and the solution is implemented using a customized execution solution program. This can not only improve the accuracy of the solution content review and the speed of verification, but also reduce the complexity of the solution verification of technical personnel and improve work efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the embodiments. It should be understood that the following drawings only illustrate certain embodiments of the present invention and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other relevant drawings can be obtained based on these drawings without paying any creative work.

[0022] Figure 1 A flowchart of a solution verification method based on a large model provided in an embodiment of the present application;

[0023] Figure 2 A flowchart of another large model-based solution verification method provided in an embodiment of the present application;

[0024] Figure 3 This is a diagram of an implementation architecture of a large-scale model-based solution verification method provided in an embodiment of the present application;

[0025] Figure 4 A schematic diagram of the structure of a solution verification device based on a large model provided in an embodiment of the present application;

[0026] Figure 5 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application.

[0027] The above drawings illustrate specific embodiments of the present application, which will be described in more detail below. These drawings and the textual description are not intended to limit the scope of the present application in any way, but rather to illustrate the concepts of the present application to those skilled in the art by reference to specific embodiments. DETAILED DESCRIPTION

[0028] The present application will be further described in detail below with reference to the accompanying drawings and examples. It should be understood that the specific embodiments described herein are intended only to illustrate the present application and are not intended to limit the present application. It should also be noted that, for ease of description, the accompanying drawings only illustrate portions of the present application, not all of the structures.

[0029] It should be noted that similar numbers and letters represent similar items in the following figures. Therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures. At the same time, in the description of this application, the terms "first", "second", etc. are only used to distinguish the description and cannot be understood as indicating or implying relative importance. The acquisition, storage, use, processing, etc. of data in the technical solution of this application comply with the relevant provisions of national laws and regulations. It should be noted that in the embodiments of this application, certain software, components, models and other existing solutions in the industry may be mentioned, and they should be regarded as exemplary. Their purpose is only to illustrate the feasibility of the implementation of the technical solution of this application, but it does not mean that the applicant has or must use the relevant content of the solution.

[0030] Figure 1 This is a flowchart of a solution verification method based on a large model provided in an embodiment of the present application. This embodiment is applicable to the case of verifying solutions to cloud platform problems and relates to the field of artificial intelligence technology. The method can be executed by a solution verification device based on a large model, which can be implemented in software and / or hardware, and is preferably configured in an electronic device, such as a computer device or server equipped with a large model. Figure 1 As shown, the method specifically includes:

[0031] S101. Extract solution attribute information from a pending solution to a target problem according to a target format of the solution, wherein the solution attribute information includes at least a problem description, a root cause analysis, a triggering condition, and a circumvention measure.

[0032] S102. Obtain a target cloud platform environment that matches the trigger condition, simulate the target problem in the target cloud platform environment according to the problem description, and output an evaluation result on whether the avoidance measures can solve the target problem.

[0033] S103: In response to the evaluation result being a passed evaluation, the avoidance measures are aggregated and transformed, execution steps of the solution are generated, and an execution environment corresponding to the execution steps is obtained.

[0034] S104: Send the execution steps and the execution environment to a pre-customized execution program, wherein the execution program is used to implement the execution steps in the execution environment and return a return value obtained by implementing the execution steps.

[0035] S105. Determine whether the solution can solve the target problem based on the return value.

[0036] Cloud platforms are virtual platforms based on internet technologies, integrating hardware resources such as computing, storage, and networking with software services to provide users with flexible and scalable IT resources and services. For example, in the fintech sector, cloud platforms can provide financial services. When problems arise on the cloud platform, technical personnel will provide solutions. However, to ensure the effectiveness of the solutions, they must be verified and evaluated before they can be implemented in production.

[0037] The solution is provided to the big model, which extracts solution attribute information from the pending solution for the target problem based on the solution's target format. The target format refers to the solution's set format, which includes a description of the solution's problem, root cause analysis, triggering conditions, and workarounds. The big model can then use this solution attribute information as a basis for verifying and evaluating the solution. The problem description identifies the current problem; the root cause analysis identifies the cause of the identified problem; the triggering conditions identify the circumstances surrounding the problem; and the workarounds identify the currently identified solutions that can resolve the problem.

[0038] After determining the solution attributes, the large model retrieves the target cloud platform environment that matches the triggering conditions. Based on the problem description, it simulates the target problem in the target cloud platform environment and outputs an evaluation result on whether the circumvention measures can solve the target problem. Specifically, different problems and solutions involve different cloud platform environments. Therefore, to accurately verify the solution, it is necessary to simulate the target problem in the matching target cloud platform environment and verify and evaluate the solution.

[0039] In one embodiment, the large model can first obtain different environmental information parameters, and then generate different cloud platform environments based on these environmental information parameters. Before the evaluation, the target cloud platform environment that matches the trigger conditions is obtained from different cloud platform environments. Since different solutions need to be verified in different environments, in order to quickly determine the environment in which the solution is implemented, the environmental information parameters can be sent to the large model in advance, and the large model can be driven to automatically enter the table in the database, thereby avoiding the failure of the large model context association caused by the long solution verification interval. Specifically, for existing data, it only needs to be imported once, and for incremental data, it needs to be imported each time it is added, so that the latest environmental information is always kept in the database.

[0040] During the evaluation process, the target problem can be simulated in the target cloud platform environment based on the problem description and the accuracy of the description can be evaluated. If the target problem description is accurate, the avoidance measures can be further evaluated and the evaluation results are output to determine whether the avoidance measures can solve the target problem.

[0041] If the assessment passes, the avoidance measures are further aggregated and transformed to generate the solution's execution steps and obtain the corresponding execution environment. These steps and the execution environment are then sent to a pre-defined execution program. The execution program executes the steps in the execution environment and returns the return value. The large model receives the return value and uses it to determine whether the solution solves the target problem.

[0042] The avoidance measures may include commands and descriptive language. Aggregating and converting the avoidance measures to generate execution steps for the solution and obtaining the execution environment corresponding to the execution steps may include: converting the descriptive language into actionable steps; identifying the execution order between the commands and the actionable steps, aggregating them based on the execution order to obtain the execution steps; and obtaining the execution environment corresponding to the execution steps based on the environment information parameters corresponding to the execution steps.

[0043] Specifically, the large model uses semantic understanding technology to understand descriptive language and convert it into actionable steps. Avoidance measures can include multiple commands and descriptive language. Therefore, after converting the actionable steps, it is also necessary to identify the execution order between these commands and actionable steps, aggregate them according to the execution order, and obtain the final executable steps. For example, log in to virtual machine B with account A, or switch the current account to account C. Then, the corresponding execution environment is obtained based on the environmental information parameters corresponding to the execution step. In the above example, the current execution environment can be determined based on the environmental information parameters of virtual machine B.

[0044] For example, a current cloud platform problem occurs when a cluster adds a pure VBS node (the cloud platform's virtual block storage node) for a certain patch version. The database records that more components are installed on the pure VBS node than are actually installed. As a result, upgrade tasks are issued for these extra components. However, since they are not actually deployed on the node, the upgrade of these components cannot be completed, and the interface reports that the upgrade is paused. After obtaining the solution, the large model first performs semantic recognition and understanding. Then, based on trigger conditions, it simulates the problem in a simulated cloud platform environment (i.e., it finds the patch version, adds the pure VBS node, and simulates its version upgrade). The large model then outputs a description of the problem. If the generated problem description matches the problem description in the obtained solution, it further outputs the root cause and solution. If the output is consistent with the solution, that is, manually deleting the remaining component data to avoid pausing the upgrade process, it then generates an aggregation program based on the steps of the solution. In this example, it ultimately generates a corresponding script program, which, by executing it, deletes the remaining component data. It should be noted that during this process, for necessary operations such as logging into the storage master management node and switching to the root user, the large model can search and complete the login through environmental information.

[0045] The method of the embodiment of the present application is implemented based on a big model. The big model is determined after learning the historical data of the problem and its solution. Utilize the learning ability of the big model to learn each group of solved problems and their solutions. Specifically, for any group of solved problems and their solutions, the solution attribute information of the solution is first input into the big model. The big model determines the cloud platform environment based on the trigger conditions in the solution, and theoretically simulates the problem in the cloud platform environment. Then, the big model evaluates whether the problem description is accurate and outputs the evaluation results. The technician can choose whether the big model needs to be re-learned based on the evaluation results of the big model. If the big model does not understand the problem description in place, the corrected evaluation information can be fed back to the big model for re-learning. If the evaluation results fed back by the big model meet expectations, the next step of learning is entered.

[0046] Next, the large model continues to analyze whether the workaround measures in the solution can solve the problem. First, the large model obtains a workaround measure based on the simulation of the problem, and compares the workaround measure with the workaround measure in the solution. Based on the comparison results, it gives an evaluation result on whether the workaround measure in the solution can solve the problem. If the large model feedback cannot solve the problem, the technician can correct the result and feed it back to the large model for re-learning. After such repeated learning, the large model has the ability to simulate problems and obtain workaround measures, as well as the ability to evaluate the workaround measures in the solution. When the result of whether the workaround measures fed back by the large model can solve the problem meets expectations, it proceeds to the next step of learning.

[0047] Next, the large model aggregates and transforms the avoidance measures, generating the solution's execution steps and obtaining the corresponding execution environment. Similarly, if the execution steps and execution environment generated by the large model don't meet expectations, technicians will revise the results and feed them back to the large model for re-learning until they meet expectations.

[0048] Through this learning process, the large model acquires the ability to understand problems and solutions. It not only acquires the corresponding cloud platform environment and simulates the problem within it, but also accurately assesses the accuracy of the problem description and the effectiveness of workarounds. It also aggregates and transforms workarounds to generate specific, executable steps. The large model then interacts with a pre-defined execution solution program, sending the execution steps and environment to it. When the execution solution program returns a return value, it can use this return value to determine whether the solution solves the target problem. Furthermore, the large model can assess the subsequent impact of the solution's implementation on the cloud platform.

[0049] The technical solution of the embodiment of the present application realizes the automatic extraction of solution attribute information based on the big model, obtains the matching target cloud platform environment according to the extracted attribute information, simulates the target problem, and then evaluates whether the avoidance measures can solve the target problem. After the evaluation is passed, the execution steps and execution environment can be automatically obtained, and the execution solution program can be called to implement the execution steps in the execution environment, and whether the solution can solve the target problem can be judged based on the return value of the program. Therefore, the solution is reviewed and evaluated based on the big model, and the solution is implemented using a customized execution solution program. This can not only improve the accuracy of the solution content review and the speed of verification, but also reduce the complexity of the solution verification of technical personnel and improve work efficiency.

[0050] Figure 2 This is a flow chart of another solution verification method based on a large model provided in an embodiment of the present application. Figure 2 As shown, the method includes the following steps:

[0051] S201. Extract solution attribute information from a pending solution to a target problem according to a target format of the solution, wherein the solution attribute information at least includes a problem description, a root cause analysis, a triggering condition, and a circumvention measure.

[0052] S202. Obtain a target cloud platform environment that matches the trigger condition, simulate the target problem in the target cloud platform environment according to the problem description, and output an evaluation result on whether the avoidance measures can solve the target problem.

[0053] S203: In response to the evaluation result being a passed evaluation, the avoidance measures are aggregated and transformed, execution steps of the solution are generated, and an execution environment corresponding to the execution steps is obtained.

[0054] S204: Send the execution steps and the execution environment to a pre-customized execution solution program, wherein the execution solution program is used to implement the execution steps in the execution environment and return a return value obtained by implementing the execution steps.

[0055] S205: Determine whether the solution is successfully implemented based on the return value. If so, execute S206.

[0056] S206: Generate verification steps according to the solution attribute information, and obtain the verification environment corresponding to the verification steps.

[0057] S207: Send the verification steps and the verification environment to a pre-customized verification solution program, wherein the verification solution program is used to execute the verification steps in the verification environment and return the verification results.

[0058] S208. Determine whether the solution can solve the target problem based on the verification result.

[0059] Specifically, the return value can be used to determine whether the solution was successfully implemented. Any exceptions or other special circumstances encountered during the execution of the solution program within the execution environment can be confirmed using the return value. If the solution was not successfully implemented, technical personnel can address it. If the solution was successful, further verification of the solution's resolution is necessary to ensure that it correctly addresses the cloud platform's current target problem.

[0060] First, the large model generates verification steps based on the solution attribute information and obtains the verification environment corresponding to the verification steps. Then, the verification steps and the verification environment are sent to the pre-customized verification solution program, wherein the verification solution program is used to execute the verification steps in the verification environment and return the verification results. Finally, the large model determines whether the solution can solve the target problem based on the verification results. Among them, the verification step is used to verify whether the previous problem still exists after the solution is implemented. For example, in the example above, the measure in the solution is to delete the residual component data, then the verification step can be to add a VBS node for the patch version after deleting the residual component data, and simulate its upgrade process. If the upgrade can be completed smoothly, it means that the verification is successful. Otherwise, the verification failure can be fed back and the technical staff can further revise the solution.

[0061] The large model can be trained using historical data on solved problems and their solutions. Specifically, the verification steps after implementing the solution are obtained from historical data. Simultaneously, the large model generates a set of verification steps based on the attributes of the current problem and solution, drawing on its own understanding. The technicians compare the two. If the verification steps generated by the large model are incorrect, the corrected steps are re-entered into the large model for further learning. This continues until the large model can output accurate verification steps. In this way, the large model possesses the ability to generate verification steps and, based on the verification results, determine whether the solution solves the target problem.

[0062] The technical solution of the embodiment of the present application realizes the automatic extraction of solution attribute information based on the large model, obtains the matching target cloud platform environment according to the extracted attribute information, simulates the target problem, and then evaluates whether the avoidance measures can solve the target problem. After the evaluation is passed, the execution steps and execution environment can be automatically obtained, and the execution solution program can be called to implement the execution steps in the execution environment. After the implementation is successful, the verification steps and verification environment are further generated and sent to the pre-customized verification solution program for implementation and verification. Finally, the verification result returned by the verification solution program is used to determine whether the solution can solve the target problem. Thus, the solution is reviewed and evaluated based on the large model, and the solution is implemented using a customized execution solution program, and verified using a customized verification solution program, forming a complete solution verification process. It can not only improve the accuracy of solution content review and the speed of verification, but also reduce the complexity of solution verification for technical personnel and improve work efficiency.

[0063] Figure 3This is a diagram illustrating the implementation architecture of a large-scale model-based solution verification method provided in an embodiment of the present application. As shown, existing environmental information is first input into the large-scale model. When updates are required, incremental environmental information is then input into the large-scale model. This environmental information is stored in a database. The large-scale model can generate different cloud platform environments based on the environmental information. Subsequently, a matching target cloud platform environment can be selected based on the solution's attribute information. Solutions 1, 2, ..., and n are the solutions to be verified. These solutions are input into the large-scale model, which uses its learning and understanding capabilities to evaluate and verify them. The model outputs solution review information and evaluation results, along with solution implementation steps and implementation environment information. The solution review information includes at least whether the solution content is incorrect and whether the problem description is accurate. The solution evaluation results include at least whether the solution can solve the problem, the impact of the problem, and whether there are any potential problems after the problem is solved. The large-scale model also outputs verification steps and a verification environment. These implementation steps and environment information are sent to a customized program to execute the solution. After successful implementation, the verification steps and environment are sent to the customized program for verification. Finally, the large-scale model determines whether the solution solves the target problem based on the verification results returned by the verification program. This completes the verification process for a set of solutions. This addresses existing issues such as limited manual review and evaluation speed, difficulty screening environmental information, slow node selection and solution implementation, and inefficient result verification and feedback, improving the efficiency and accuracy of solution verification.

[0064] Figure 4 This is a schematic diagram of the structure of the solution verification device based on the large model provided in the embodiment of the present application. Figure 4 As shown, the device includes:

[0065] An information extraction module 410 is configured to extract solution attribute information from a pending solution to a target problem according to a target format of the solution, wherein the solution attribute information includes at least a problem description, a root cause analysis, a trigger condition, and a workaround;

[0066] Evaluation module 420 is configured to obtain a target cloud platform environment that matches the trigger condition, simulate the target problem in the target cloud platform environment according to the problem description, and output an evaluation result of whether the circumvention measure can solve the target problem;

[0067] an execution step generation module 430 for aggregating and transforming the avoidance measures in response to the evaluation result being a passed evaluation, generating execution steps for the solution, and obtaining an execution environment corresponding to the execution steps;

[0068] An execution step sending module 440 is used to send the execution step and the execution environment to a pre-customized execution solution program, wherein the execution solution program is used to implement the execution step in the execution environment and return a return value obtained by implementing the execution step;

[0069] The verification module 450 is configured to determine, based on the return value, whether the solution can solve the target problem.

[0070] In some embodiments, the apparatus further comprises:

[0071] An environment information acquisition module is used to obtain environment information parameters and generate different cloud platform environments according to the environment information parameters;

[0072] Accordingly, the evaluation module 420 is specifically configured to:

[0073] A target cloud platform environment matching the trigger condition is obtained from the different cloud platform environments.

[0074] In some embodiments, the evaluation module 420 includes:

[0075] A problem simulation unit, configured to simulate the target problem in the target cloud platform environment according to the problem description, and evaluate the accuracy of the description of the target problem;

[0076] An evaluation result output unit is used to output an evaluation result of whether the avoidance measure can solve the target problem if the target problem description is accurate.

[0077] In some embodiments, the circumvention measures include commands and descriptive language;

[0078] Accordingly, the execution step generation module 430 includes:

[0079] a conversion unit, configured to convert the descriptive language into operable steps;

[0080] an aggregation unit, configured to identify an execution order between the command and the operable steps, and perform aggregation according to the execution order to obtain the execution steps;

[0081] The acquiring unit is used to acquire the execution environment corresponding to the execution step according to the environment information parameters corresponding to the execution step.

[0082] In some embodiments, the verification module 450 includes:

[0083] A first judging unit, configured to judge whether the solution is successfully implemented according to the return value;

[0084] a verification environment acquisition unit, configured to generate a verification step according to the solution attribute information and acquire a verification environment corresponding to the verification step if the judgment unit determines that the solution is successfully implemented;

[0085] a sending unit, configured to send the verification steps and the verification environment to a pre-customized verification program, wherein the verification program is configured to execute the verification steps in the verification environment and return a verification result;

[0086] The second judgment unit is used to judge whether the solution can solve the target problem according to the verification result.

[0087] In some embodiments, the method is implemented based on a large model, wherein the large model is determined by learning historical data of problems and their solutions.

[0088] The large-model-based solution verification device provided in the embodiment of the present application can be used to execute the technical solution of the large-model-based solution verification method in the above-mentioned embodiment. Its implementation principle and technical effects are similar and will not be repeated here.

[0089] It should be noted that the division of the modules of the above device is merely a division of logical functions. In actual implementation, they can be fully or partially integrated into a single physical entity or physically separated. Furthermore, these modules can be implemented entirely in the form of software called by a processing element; or entirely in the form of hardware; or some modules can be implemented in the form of software called by a processing element, and some modules can be implemented in the form of hardware. For example, evaluation module 420 can be a separate processing element, or it can be integrated into a chip of the above device. Furthermore, it can be stored in the form of program code in the memory of the above device, and called by a processing element of the above device to perform the functions of the above evaluation module 420. The implementation of other modules is similar. Furthermore, these modules can be fully or partially integrated together, or implemented independently. The processing element here can be an integrated circuit with signal processing capabilities. During implementation, each step of the above method or each of the above modules can be completed by hardware integrated logic circuits in the processor element or by software instructions.

[0090] Figure 5 This is a schematic diagram of the structure of the electronic device provided in the embodiment of the present application. Figure 5 As shown, the electronic device may include: a transceiver 121 , a processor 122 , and a memory 123 .

[0091] The processor 122 executes the computer-executable instructions stored in the memory, so that the processor 122 implements the solutions in the above embodiments. The processor 122 can be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it can also be a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components.

[0092] The memory 123 is connected to the processor 122 via a system bus and communicates with the processor 122. The memory 123 is used to store computer program instructions.

[0093] The transceiver 121 may be used to obtain tasks to be executed and configuration information of the tasks to be executed.

[0094] The system bus can be a Peripheral Component Interconnect (PCI) bus or an Extended Industry Standard Architecture (EISA) bus, among others. The system bus can be divided into an address bus, a data bus, a control bus, and so on. For ease of illustration, the figure shows only one thick line, but this does not imply that there is only one bus or only one type of bus. Transceivers are used to enable communication between the database access device and other computers (e.g., clients, read-write libraries, and read-only libraries). Memory may include random access memory (RAM) and non-volatile memory.

[0095] The electronic device provided in the embodiment of the present application may be the terminal device of the above embodiment.

[0096] An embodiment of the present application also provides a computer-readable storage medium, which stores computer instructions. When the computer instructions are executed on a computer, the computer executes the technical solution of the solution verification method based on the large model in the above embodiment.

[0097] An embodiment of the present application also provides a computer program product, which includes a computer program stored in a computer-readable storage medium. At least one processor can read the computer program from the computer-readable storage medium. When the at least one processor executes the computer program, it can implement the technical solution of the large model-based solution verification method in the above embodiment.

[0098] The computer program product, during implementation, may be written in one or more programming languages or a combination thereof, for performing the operations of the present application, including object-oriented programming languages such as Java, Smalltalk, C++, and conventional procedural programming languages such as "C" or similar programming languages. The program code may be executed entirely on the user's computer, partially on the user's computer, as a separate software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving a remote computer, the remote computer may be connected to the user's computer via any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0099] Note that the above are only preferred embodiments of the present application and the technical principles employed. Those skilled in the art will understand that the present application is not limited to the specific embodiments herein, and that various obvious changes, readjustments, and substitutions can be made by those skilled in the art without departing from the scope of protection of the present application. Therefore, although the present application has been described in more detail through the above embodiments, the present application is not limited to the above embodiments and may include many other equivalent embodiments without departing from the scope of the present application. The scope of the present application is determined by the scope of the appended claims.

Claims

1. A solution verification method based on a large model, characterized in that: include: Extracting solution attribute information from the pending solution to the target problem according to the target format of the solution, wherein the solution attribute information at least includes a problem description, a root cause analysis, a triggering condition, and a circumvention measure; Acquire a target cloud platform environment that matches the trigger condition, simulate the target problem in the target cloud platform environment according to the problem description, and output an evaluation result of whether the avoidance measure can solve the target problem; In response to the evaluation result being a passed evaluation, aggregating and transforming the avoidance measures, generating execution steps of the solution, and obtaining an execution environment corresponding to the execution steps; Sending the execution steps and the execution environment to a pre-customized execution program, wherein the execution program is used to implement the execution steps in the execution environment and return a return value obtained by implementing the execution steps; It is determined whether the solution can solve the target problem according to the return value.

2. The method according to claim 1, characterized in that Also includes: Obtaining environmental information parameters, and generating different cloud platform environments according to the environmental information parameters; Accordingly, obtaining a target cloud platform environment that matches the trigger condition includes: A target cloud platform environment matching the trigger condition is obtained from the different cloud platform environments.

3. The method according to claim 1, characterized in that The step of simulating the target problem in the target cloud platform environment according to the problem description and outputting an evaluation result of whether the circumvention measure can solve the target problem includes: According to the problem description, simulate the target problem in the target cloud platform environment and evaluate the accuracy of the description of the target problem; If the target problem is accurately described, an evaluation result of whether the avoidance measure can solve the target problem is output.

4. The method according to claim 1, wherein Said circumvention measures include imperative and descriptive language; Accordingly, aggregating and converting the avoidance measures to generate execution steps of the solution and obtaining the execution environment corresponding to the execution steps include: Convert the descriptive language into actionable steps; Identify the execution order between the command and the operable steps, and aggregate according to the execution order to obtain the execution steps; Obtain the execution environment corresponding to the execution step according to the environment information parameters corresponding to the execution step.

5. The method according to claim 1, wherein The determining, based on the return value, whether the solution can solve the target problem includes: Determine whether the solution is successfully implemented according to the return value; If so, generating a verification step according to the scheme attribute information and obtaining a verification environment corresponding to the verification step; Sending the verification steps and the verification environment to a pre-customized verification program, wherein the verification program is used to execute the verification steps in the verification environment and return a verification result; Determine whether the solution can solve the target problem based on the verification result.

6. The method according to claim 1, characterized in that The method is implemented based on a large model, wherein the large model is determined after learning historical data of problems and their solutions.

7. A solution verification device based on a large model, characterized in that: include: An information extraction module is configured to extract solution attribute information from a pending solution to a target problem according to a target format of the solution, wherein the solution attribute information includes at least a problem description, a root cause analysis, a trigger condition, and a circumvention measure; An evaluation module is configured to obtain a target cloud platform environment that matches the trigger condition, simulate the target problem in the target cloud platform environment according to the problem description, and output an evaluation result of whether the circumvention measure can solve the target problem; an execution step generation module for aggregating and transforming the avoidance measures in response to the evaluation result being a passed evaluation, generating execution steps for the solution, and obtaining an execution environment corresponding to the execution steps; An execution step sending module, configured to send the execution step and the execution environment to a pre-customized execution solution program, wherein the execution solution program is configured to implement the execution step in the execution environment and return a return value obtained by implementing the execution step; A verification module is used to determine whether the solution can solve the target problem based on the return value.

8. An electronic device, characterized in that: include: a processor, and a memory communicatively connected to the processor; The memory stores computer-executable instructions; The processor executes the computer-executable instructions stored in the memory to implement the large model-based solution verification method according to any one of claims 1 to 6.

9. A computer-readable storage medium, characterized in that The computer-readable storage medium stores computer-executable instructions, which, when executed by a processor, are used to implement the large-model-based solution verification method according to any one of claims 1 to 6.

10. A computer program product, characterized in that The invention comprises a computer program, which, when executed by a processor, implements the large model-based solution verification method according to any one of claims 1 to 6.

Images