Core network dynamic network script checking method and system based on large model
By adopting a core network dynamic network script verification method based on a large model, the system achieves automated and multi-dimensional analysis of core network dynamic network scripts, solving the problem of low efficiency in traditional manual verification, improving verification accuracy and efficiency, and reducing network failure risks.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-12-24
- Publication Date
- 2026-04-07
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
Traditional manual verification of core network scripts is inefficient, error-prone, and unable to meet the growing business demands. Furthermore, network congestion events frequently occur due to configuration errors.
A core network dynamic network script verification method based on a large model is adopted. Through data preprocessing, model training and optimization, knowledge base construction and result feedback iteration, the method realizes automated, multi-dimensional analysis and feedback of script data, thereby improving the verification accuracy.
It significantly improved the accuracy and efficiency of script verification, reduced the probability of network failure, reduced operation and maintenance costs, reduced verification time from 45 minutes to 3 minutes, and increased the error interception rate to over 99%.
Smart Images
Figure CN121814565A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of telecommunications operator operation and maintenance technology, specifically to a core network dynamic script verification method and system based on a large model. Background Technology
[0002] With the continuous expansion of communication network scale and the increasing complexity of services, the accuracy and security of core network and mobile network operations are becoming increasingly critical.
[0003] Traditional manual verification of network configuration scripts is inefficient, error-prone, and unable to meet the ever-increasing business demands. Meanwhile, the widespread adoption of emerging technologies such as 5G and IoT places higher demands on the accuracy and timeliness of network configuration. The core network's operational workload has surged, processing over 2000 configuration change work orders daily, involving parameter adjustments for network elements such as the MME and AMF, and neighbor cell relationship configurations. Traditional manual verification requires three levels of review: initial review by engineers, secondary review by senior experts, and final review by the operations supervisor. Each work order takes an average of 45 minutes, and due to differences in engineers' understanding of cross-vendor equipment configuration specifications, approximately five network congestion events caused by configuration errors occur monthly.
[0004] Improving the efficiency and accuracy of DVR script verification and ensuring stable network operation are technical problems that need to be solved. Summary of the Invention
[0005] The technical objective of this invention is to address the above-mentioned shortcomings by providing a core network dynamic script verification method and system based on a large model, thereby solving the technical problem of how to improve the efficiency and accuracy of dynamic script verification and ensure stable network operation.
[0006] In a first aspect, the present invention provides a core network dynamic script verification method based on a large model, comprising the following steps:
[0007] Data preprocessing: Collect core network script data, preprocess the script data to obtain script data in a predetermined format, and construct a sample set based on the preprocessed script data and the corresponding actual verification results;
[0008] Large model training and optimization: Supervised training of the large model is performed based on the sample set, the model parameters of the large model are adjusted, and a reinforcement learning mechanism is introduced. The large model is optimized according to the actual verification effect to obtain the trained and optimized large model.
[0009] Core network script verification database construction: collect product documents, standardized configuration change scenarios and configuration specifications of core network equipment, and build a configuration rule knowledge base;
[0010] Script verification: Collect the core network footstep data to be verified and preprocess it. Input the preprocessed footstep data into the trained and optimized large model. The trained and optimized large model calls the configuration rule knowledge base to analyze from multiple dimensions including syntax, logic and function, and outputs the predicted verification effect.
[0011] Results generation and feedback: Based on the predicted verification results, the large model generates a detailed verification report after training and optimization, and feeds the verification report back to the user. The user can modify the script data according to the verification report, and the user can interact with the large model. During the interaction, the user can consult the large model about the problems in the verification report, and the large model provides corresponding explanations and modification suggestions.
[0012] Model optimization iteration: Collect user feedback on the verification results and the modified script data and update them to the sample set to continuously train and optimize the large model.
[0013] Preferably, preprocessing the script data includes the following operations:
[0014] Clean the script data to remove redundant information, comments, and invalid characters;
[0015] The script data is processed into words and sentences according to certain rules and then converted into a format suitable for large models.
[0016] As a preferred option, the Dongwang script verification knowledge base includes L1, L2, and L3 layers. The L1 layer integrates a syntax rule library, covering various script syntax paradigms, keyword specifications, and error patterns to achieve accurate matching of syntax errors. The L2 layer builds a standard library, which includes industry security guidelines, coding conventions, and interface call specifications, and identifies compliance issues through rule mapping. The L3 layer creates a semantic logic library, which includes business scenario logic models, common semantic conflict cases, and risk patterns, and identifies deep semantic problems by combining contextual reasoning.
[0017] As a preferred approach, script verification involves checking whether the script conforms to the syntax specifications of the core network and dynamic network scripts, including the correctness of statement structure, variable declaration and usage, and function calls; analyzing whether the script's execution flow is reasonable and whether there are logical contradictions or infinite loops; and determining whether the script can achieve the corresponding core network functions as expected and whether it meets business requirements.
[0018] Ideally, the verification report should include the type of problem, its specific location, its severity, and its potential impact.
[0019] Secondly, the present invention provides a core network dynamic network script verification system based on a large model, comprising a data preprocessing module, a large model training and optimization module, a dynamic network script verification database construction module, a script verification module, a result generation and feedback module, and a model optimization and iteration module;
[0020] The data preprocessing module is used to perform the following: collect core network dynamic network script data, preprocess the script data to obtain script data in a predetermined format, and construct a sample set based on the preprocessed script data and the corresponding actual verification results;
[0021] The large model training and optimization module is used to perform the following: supervised training of the large model based on the sample set, adjust the model parameters of the large model, introduce a reinforcement learning mechanism, optimize the large model according to the actual verification effect, and obtain the trained and optimized large model;
[0022] The Dynamic Network Script Verification Database Construction Module is used to perform the following: collect product documents, standardized configuration change scenarios and configuration specifications of core network equipment, and build a configuration rule knowledge base;
[0023] The script verification module is used to perform the following: collect the core network footstep data to be verified and preprocess it; input the preprocessed footstep data into the trained and optimized large model; the trained and optimized large model calls the configuration rule knowledge base to analyze from multiple dimensions including syntax, logic and function, and outputs the predicted verification results.
[0024] The results generation and feedback module is used to perform the following: Based on the predicted verification effect, a detailed verification report is generated by the large model after training and optimization. The verification report is then fed back to the user, allowing the user to modify the script data according to the verification report. The module also supports the user to interact with the large model. During the interaction, the user can consult the large model about the issues in the verification report, and the large model will provide corresponding explanations and modification suggestions.
[0025] The model optimization iteration module is used to perform the following: collect user feedback on the verification results and the modified script data and update them to the sample set, and continuously train and optimize the large model.
[0026] Preferably, the data preprocessing module is used to perform the following operations to preprocess the script data:
[0027] Clean the script data to remove redundant information, comments, and invalid characters;
[0028] The script data is processed into words and sentences according to certain rules and then converted into a format suitable for large models.
[0029] As a preferred option, the Dongwang script verification knowledge base includes L1, L2, and L3 layers. The L1 layer integrates a syntax rule library, covering various script syntax paradigms, keyword specifications, and error patterns to achieve accurate matching of syntax errors. The L2 layer builds a standard library, which includes industry security guidelines, coding conventions, and interface call specifications, and identifies compliance issues through rule mapping. The L3 layer creates a semantic logic library, which includes business scenario logic models, common semantic conflict cases, and risk patterns, and identifies deep semantic problems by combining contextual reasoning.
[0030] As a preferred approach, in terms of syntax, the script verification module is used to check whether the script conforms to the syntax specifications of the core network script, including whether the statement structure, variable declaration and usage, and function calls are correct; in terms of logic, the script verification module is used to analyze whether the execution flow of the script is reasonable, and whether there are logical contradictions and infinite loops; in terms of functional implementation, the script verification module is used to determine whether the script can achieve the corresponding core network functions as expected and whether it meets business requirements.
[0031] Ideally, the verification report should include the type of problem, its specific location, its severity, and its potential impact.
[0032] The core network script verification method and system based on a large model of the present invention has the following advantages: Through the rapid processing capabilities of the large model and the learning and training of large amounts of script data, it can accurately identify various complex syntactic, semantic, and specification errors in the scripts, avoiding oversights and misjudgments that are prone to occur during manual verification. This significantly improves the accuracy of script verification and effectively reduces the probability of network failures caused by script errors. It also reduces reliance on extensive manual verification, saving on operation and maintenance costs. Attached Figure Description
[0033] To more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.
[0034] The invention will be further described below with reference to the accompanying drawings.
[0035] Figure 1 Here is a flowchart of a core network dynamic network script verification method based on a large model, as shown in Example 1.
[0036] Figure 2 This is a script flowchart from Example 1, a core network dynamic network script verification method based on a large model. Detailed Implementation
[0037] The present invention will be further described below with reference to the accompanying drawings and specific embodiments, so that those skilled in the art can better understand and implement the present invention. However, the embodiments are not intended to limit the present invention. In the absence of conflict, the embodiments of the present invention and the technical features in the embodiments can be combined with each other.
[0038] This invention provides a core network dynamic network script verification method and system based on a large model, which is used to solve the technical problem of how to improve the efficiency and accuracy of dynamic network script verification and ensure stable network operation.
[0039] Example 1:
[0040] This invention provides a core network dynamic network script verification method based on a large model, comprising six steps: data preprocessing, large model training and optimization, construction of dynamic network script verification database, script verification, result generation and feedback, and model optimization iteration.
[0041] Step S100 Data Preprocessing: Collect core network dynamic network script data, preprocess the script data to obtain script data in a predetermined format, and construct a sample set based on the preprocessed script data and the corresponding actual verification results.
[0042] As a specific implementation of data preprocessing, core network script data is collected, and the script data is cleaned to remove redundant information, comments, and invalid characters. The scripts are then segmented into words and sentences according to certain rules and converted into a format suitable for large model input.
[0043] Step S200 Large Model Training and Optimization: Supervised training of the large model is performed based on the sample set, the model parameters of the large model are adjusted, and a reinforcement learning mechanism is introduced. The large model is optimized according to the actual verification effect to obtain the trained and optimized large model.
[0044] As a specific implementation of large-scale model training and optimization, this step constructs a training dataset containing a large number of core network dynamic network script samples and corresponding verification results. This dataset is used to perform supervised training on the large-scale model, adjusting its parameters to enable it to accurately understand the semantics and logic of the core network dynamic network scripts. Simultaneously, a reinforcement learning mechanism is introduced to optimize the large-scale model based on actual verification results, improving the model's verification accuracy and adaptability.
[0045] Step S300 Dynamic Network Script Verification Database Construction: Collect product documents, standardized configuration change scenarios and configuration specifications of core network equipment, and build a configuration rule knowledge base.
[0046] The Dongwang script verification knowledge base comprises three layers: L1, L2, and L3. Layer L1 integrates a syntax rule library, covering various script syntax paradigms, keyword specifications, and error patterns to achieve accurate syntax error matching. Layer L2 constructs a standard library, incorporating industry security guidelines, coding conventions, and API call specifications, using rule mapping to identify compliance issues. Layer L3 creates a semantic logic library, including business scenario logic models, common semantic conflict cases, and risk patterns, combining contextual reasoning to identify deep semantic problems. The knowledge base is continuously iterated through a dynamic update mechanism, providing precise rules and case support for verification at each level, improving the comprehensiveness and accuracy of script problem identification.
[0047] Step S400 Script Verification: Collect the core network footstep data to be verified and preprocess it. Input the preprocessed footstep data into the trained and optimized large model. The trained and optimized large model calls the configuration rule knowledge base to analyze from multiple dimensions including syntax, logic and function, and outputs the predicted verification results.
[0048] In this embodiment, the preprocessed core network dynamic network script data is input into a trained and optimized large model. The large model analyzes the script from multiple dimensions, including syntax, logic, and functional implementation. In terms of syntax, it checks whether the script conforms to the syntax specifications of core network dynamic network scripts, including whether the statement structure, variable declarations and usage, and function calls are correct. In terms of logic, it analyzes whether the script's execution flow is reasonable and whether there are logical contradictions, infinite loops, or other problems. In terms of functional implementation, it determines whether the script can implement the corresponding core network functions as expected and whether it meets business requirements.
[0049] Step S500 Result Generation and Feedback: Based on the predicted verification effect, a detailed verification report is generated by the large model after training and optimization. The verification report is then fed back to the user, allowing the user to modify the script data according to the verification report. The user can also interact with the large model. During the interaction, the user can consult the large model about the issues in the verification report, and the large model will provide corresponding explanations and modification suggestions.
[0050] In this embodiment, the large model generates a detailed verification report based on the verification analysis results. The report clearly identifies the problems in the script, including the type of problem, its specific location, severity, and potential impact. Simultaneously, the verification results are fed back to the user, who can modify the script based on the report. The system supports interaction between the user and the large model; users can consult the large model about the problems in the report, and the large model provides corresponding explanations and modification suggestions to help users better understand and resolve the issues.
[0051] Step S600 Model Optimization Iteration: Collect user feedback on the verification results and the modified script data and update them to the sample set, and continuously train and optimize the large model.
[0052] In this embodiment, this step collects user feedback on the verification results and the modified script data, adds this data to the training dataset, and continuously trains and optimizes the large model to continuously improve the verification capability and accuracy of the large model.
[0053] The method in this embodiment utilizes the powerful natural language processing and reasoning capabilities of large models to achieve efficient, comprehensive, and accurate verification of the syntax, semantics, and specifications of core network dynamic network scripts. This overcomes the shortcomings of traditional verification methods, improves the quality of core network dynamic network scripts, ensures the stable and reliable operation of the core network, and reduces the risk of network failures and maintenance costs caused by script errors.
[0054] This embodiment achieves fully automated verification by embedding AI agents into the production process, reducing the review time to within 3 minutes and increasing the error interception rate to over 99%. The research on AI agent technology for core network script verification based on a large model aims to utilize advanced artificial intelligence technology to improve the efficiency and accuracy of network script verification and ensure stable network operation.
[0055] Example 2:
[0056] This invention discloses a core network dynamic network script verification system based on a large model, comprising a data preprocessing module, a large model training and optimization module, a dynamic network script verification database construction module, a script verification module, a result generation and feedback module, and a model optimization and iteration module.
[0057] The data preprocessing module is used to perform the following: collect core network script data, preprocess the script data to obtain script data in a predetermined format, and construct a sample set based on the preprocessed script data and the corresponding actual verification results.
[0058] As a specific implementation of the data preprocessing module, this module collects core network script data, cleans the script data to remove redundant information, comments, and invalid characters, and processes the scripts into words and sentences according to certain rules, transforming them into a format suitable for large model input.
[0059] The large model training and optimization module is used to perform the following: supervised training of the large model based on the sample set, adjust the model parameters of the large model, introduce a reinforcement learning mechanism, optimize the large model according to the actual verification effect, and obtain the trained and optimized large model.
[0060] As a specific implementation of the large-scale model training and optimization module, this module performs the following operations: It constructs a training dataset containing a large number of core network and dynamic network script samples and corresponding verification results; uses this dataset to perform supervised training on the large-scale model, adjusting the model parameters to enable it to accurately understand the semantics and logic of the core network and dynamic network scripts; and simultaneously introduces a reinforcement learning mechanism to optimize the large-scale model based on actual verification results, improving the model's verification accuracy and adaptability.
[0061] The Dynamic Network Script Verification Database Building Module is used to perform the following: collect product documents, standardized configuration change scenarios and configuration specifications of core network equipment, and build a configuration rule knowledge base.
[0062] The Dongwang script verification knowledge base comprises three layers: L1, L2, and L3. Layer L1 integrates a syntax rule library, covering various script syntax paradigms, keyword specifications, and error patterns to achieve accurate syntax error matching. Layer L2 constructs a standard library, incorporating industry security guidelines, coding conventions, and API call specifications, using rule mapping to identify compliance issues. Layer L3 creates a semantic logic library, including business scenario logic models, common semantic conflict cases, and risk patterns, combining contextual reasoning to identify deep semantic problems. The knowledge base is continuously iterated through a dynamic update mechanism, providing precise rules and case support for verification at each level, improving the comprehensiveness and accuracy of script problem identification.
[0063] The script verification module is used to perform the following: collect the core network footstep data to be verified and preprocess it; input the preprocessed footstep data into the trained and optimized large model; the trained and optimized large model calls the configuration rule knowledge base to analyze from multiple dimensions including syntax, logic and function, and outputs the predicted verification results.
[0064] In this embodiment, the module inputs the preprocessed core network dynamic network script data into a trained and optimized large model. The large model analyzes the script from multiple dimensions, including syntax, logic, and functional implementation. In terms of syntax, it checks whether the script conforms to the syntax specifications of core network dynamic network scripts, including whether the statement structure, variable declarations and usage, and function calls are correct. In terms of logic, it analyzes whether the script's execution flow is reasonable and whether there are logical contradictions, infinite loops, or other problems. In terms of functional implementation, it determines whether the script can implement the corresponding core network functions as expected and whether it meets business requirements.
[0065] The results generation and feedback module performs the following functions: based on the predicted verification results and the large model after training and optimization, it generates a detailed verification report, which is then fed back to the user. The user can modify the script data based on the verification report, and the user can interact with the large model. During the interaction, the user can consult the large model about the issues in the verification report, and the large model will provide corresponding explanations and modification suggestions.
[0066] In this embodiment, the large model generates a detailed verification report based on the verification analysis results. The report clearly identifies the problems in the script, including the type of problem, its specific location, severity, and potential impact. Simultaneously, the verification results are fed back to the user, who can modify the script based on the report. The system supports interaction between the user and the large model; users can consult the large model about the problems in the report, and the large model provides corresponding explanations and modification suggestions to help users better understand and resolve the issues.
[0067] The model optimization iteration module is used to perform the following: collect user feedback on the verification results and the modified script data and update them to the sample set, and continuously train and optimize the large model.
[0068] In this embodiment, this module is used to collect user feedback on the verification results and the modified script data. This data is then added to the training dataset to continuously train and optimize the large model, thereby continuously improving the verification capabilities and accuracy of the large model.
[0069] The core network dynamic network script verification method and system based on a large model provided by the present invention have been described in detail above. Specific examples have been used to illustrate the principles and implementation methods of the present invention. The description of the above embodiments is only for the purpose of helping to understand the method and core ideas of the present invention. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of the present invention. Therefore, the content of this specification should not be construed as a limitation of the present invention.
Claims
1. A core network dynamic network script verification method based on a large model, characterized in that, Includes the following steps: Data preprocessing: Collect core network script data, preprocess the script data to obtain script data in a predetermined format, and construct a sample set based on the preprocessed script data and the corresponding actual verification results; Large model training and optimization: Supervised training of the large model is performed based on the sample set, the model parameters of the large model are adjusted, and a reinforcement learning mechanism is introduced. The large model is optimized according to the actual verification effect to obtain the trained and optimized large model. Core network script verification database construction: collect product documents, standardized configuration change scenarios and configuration specifications of core network equipment, and build a configuration rule knowledge base; Script verification: Collect the core network footstep data to be verified and preprocess it. Input the preprocessed footstep data into the trained and optimized large model. The trained and optimized large model calls the configuration rule knowledge base to analyze from multiple dimensions including syntax, logic and function, and outputs the predicted verification effect. Results generation and feedback: Based on the predicted verification results, the large model generates a detailed verification report after training and optimization, and feeds the verification report back to the user. The user can modify the script data according to the verification report, and the user can interact with the large model. During the interaction, the user can consult the large model about the problems in the verification report, and the large model provides corresponding explanations and modification suggestions. Model optimization iteration: Collect user feedback on the verification results and the modified script data and update them to the sample set to continuously train and optimize the large model.
2. The core network dynamic network script verification method based on a large model according to claim 1, characterized in that, Preprocessing script data includes the following operations: Clean the script data to remove redundant information, comments, and invalid characters; The script data is processed into words and sentences according to certain rules and then converted into a format suitable for large models.
3. The core network dynamic network script verification method based on a large model according to claim 1, characterized in that, The Dongwang script verification knowledge base consists of three layers: L1, L2, and L3. The L1 layer integrates a syntax rule library, covering various script syntax paradigms, keyword specifications, and error patterns to achieve accurate matching of syntax errors. The L2 layer builds a standard library, which includes industry security guidelines, coding conventions, and interface call specifications, and uses rule mapping to identify compliance issues. The L3 layer creates a semantic logic library, which includes business scenario logic models, common semantic conflict cases, and risk patterns, and uses contextual reasoning to identify deep semantic problems.
4. The core network dynamic network script verification method based on a large model according to claim 1, characterized in that, During script verification, in terms of syntax, the script is checked to see if it conforms to the syntax specifications of the core network script, including whether the statement structure, variable declaration and usage, and function calls are correct; in terms of logic, the execution flow of the script is analyzed to see if it is reasonable and whether there are any logical contradictions or infinite loops; in terms of functional implementation, it is determined whether the script can achieve the corresponding core network functions as expected and whether it meets business requirements.
5. The core network dynamic network script verification method based on a large model according to claim 1, characterized in that, The verification report includes the type of problem, its specific location, its severity, and its potential impact.
6. A core network dynamic network script verification system based on a large model, characterized in that, It includes a data preprocessing module, a large model training and optimization module, a dynamic network script verification database construction module, a script verification module, a result generation and feedback module, and a model optimization and iteration module; The data preprocessing module is used to perform the following: collect core network dynamic network script data, preprocess the script data to obtain script data in a predetermined format, and construct a sample set based on the preprocessed script data and the corresponding actual verification results; The large model training and optimization module is used to perform the following: supervised training of the large model based on the sample set, adjust the model parameters of the large model, introduce a reinforcement learning mechanism, optimize the large model according to the actual verification effect, and obtain the trained and optimized large model; The Dynamic Network Script Verification Database Construction Module is used to perform the following: collect product documents, standardized configuration change scenarios and configuration specifications of core network devices, and build a configuration rule knowledge base; The script verification module is used to perform the following: collect the core network footstep data to be verified and preprocess it; input the preprocessed footstep data into the trained and optimized large model; the trained and optimized large model calls the configuration rule knowledge base to analyze the data from multiple dimensions, including syntax, logic, and function; and outputs the predicted verification results. The results generation and feedback module is used to perform the following: Based on the predicted verification effect, a detailed verification report is generated by the large model after training and optimization. The verification report is then fed back to the user, allowing the user to modify the script data according to the verification report. The module also supports the user to interact with the large model. During the interaction, the user can consult the large model about the issues in the verification report, and the large model will provide corresponding explanations and modification suggestions. The model optimization iteration module is used to perform the following: collect user feedback on the verification results and the modified script data and update them to the sample set, and continuously train and optimize the large model.
7. The core network dynamic network script verification system based on a large model according to claim 6, characterized in that, The data preprocessing module performs the following operations to preprocess the script data: Clean the script data to remove redundant information, comments, and invalid characters; The script data is processed into words and sentences according to certain rules and then converted into a format suitable for large models.
8. The core network dynamic network script verification system based on a large model according to claim 6, characterized in that, The Dongwang script verification knowledge base consists of three layers: L1, L2, and L3. The L1 layer integrates a syntax rule library, covering various script syntax paradigms, keyword specifications, and error patterns to achieve accurate matching of syntax errors. The L2 layer builds a standard library, which includes industry security guidelines, coding conventions, and interface call specifications, and uses rule mapping to identify compliance issues. The L3 layer creates a semantic logic library, which includes business scenario logic models, common semantic conflict cases, and risk patterns, and uses contextual reasoning to identify deep semantic problems.
9. The core network dynamic network script verification system based on a large model according to claim 6, characterized in that, In terms of syntax, the script verification module checks whether the script conforms to the syntax specifications of the core network script, including whether the statement structure, variable declaration and usage, and function calls are correct. In terms of logic, the script verification module analyzes whether the execution flow of the script is reasonable and whether there are logical contradictions or infinite loops. In terms of functional implementation, the script verification module determines whether the script can achieve the corresponding core network functions as expected and whether it meets business requirements.
10. The core network dynamic network script verification system based on a large model according to claim 6, characterized in that, The verification report includes the type of problem, its specific location, its severity, and its potential impact.