Industry reuse-oriented intelligent development standard generation system and method
Patent Information
- Application Number
- CN202610896797.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-06-22
- Publication Date
- 2026-08-11
AI Technical Summary
[0002]传统行业开发标准的制定通常依赖专家经验，存在标准制定周期长，难以适应行业快速变化，跨项目复用性差，相似标准需重复编写，标准之间缺乏冲突检测，容易产生逻辑矛盾或时序冲突的问题；
附图说明
Smart Images
Figure CN122548121A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of computer science, and more particularly to an intelligent development standard generation system and method for industry reuse. Background Technology
[0002] Traditional industry development standards typically rely on expert experience, which results in long standard development cycles, difficulty in adapting to rapid industry changes, poor cross-project reusability, the need to repeatedly write similar standards, and a lack of conflict detection between standards, which can easily lead to logical contradictions or timing conflicts.
[0003] Existing technologies, such as rule-based standard management systems, can only perform static retrieval and cannot achieve intelligent generation and adaptive evolution. Therefore, there is an urgent need for a system and method that can automatically learn from industry data, reuse and generate conflict-free development standards. Summary of the Invention
[0004] The purpose of this invention is to address the shortcomings of existing technologies by proposing an intelligent development standard generation system and method for industry reuse.
[0005] In a first aspect, the present invention provides an intelligent development standard generation system and method for industry reuse, including:
[0006] The feature extraction module is used to receive development data sources from the industry and perform semantic annotation on the development data sources according to the industry ontology library pre-built in the industry, so as to extract the implicit association features of the industry. The implicit association features are used to characterize the non-linear dependency relationship between different development activities in the industry.
[0007] The rule mining module, coupled to the feature extraction module, is used to perform cross-project similarity calculations in the historical development standard library based on the implicit association features to generate a reuse rule cluster, which includes multiple standard segments with time decay weights.
[0008] The detection and adjustment module is used to perform static conflict detection and dynamic collaborative evolution analysis on the reuse rule cluster. When a conflict is detected, the reuse rule cluster is structurally adjusted according to the set industry priority arbitration strategy to generate a conflict-free candidate standard skeleton.
[0009] The output module is used to instantiate the candidate standard skeleton with parameters according to the specific project context of this industry, so as to generate an executable industry development standard document.
[0010] Preferably, the feature extraction module further includes:
[0011] A multimodal data unit is used to heterogeneously fuse unstructured text, semi-structured configuration files, and structured code graphs from the development data source to generate a unified industry feature tensor.
[0012] The mining unit performs hierarchical propagation learning on the industry feature tensor through a graph neural network to extract implicit design patterns and anti-patterns across projects. The implicit correlation features include the implicit design patterns and the anti-patterns.
[0013] Preferably, the rule mining module further includes:
[0014] The time decay weight is dynamically calculated based on the actual number of times the standard fragment has been reused in historical applications, the reuse success rate, and the timestamp of the last reuse, according to a nonlinear exponential decay function, so that the standard fragments that are recent and have a high success rate receive higher reuse priority.
[0015] Preferably, the detection and adjustment module further includes:
[0016] The conflict resolution unit is used to detect logical contradictions within the same reuse rule cluster and to detect temporal constraint conflicts between different reuse rule clusters, and to automatically delete or modify the conflicting standard segments according to a pre-set industry priority strategy.
[0017] The verification unit is used to simulate the execution effect of the reuse rule cluster in a future time window based on the development status change diagram of this industry, and to identify and eliminate implicit mismatches caused by changes in the industry environment.
[0018] Preferably, the output module further includes:
[0019] The parameter filling unit is used to map the abstract variables in the candidate standard skeleton to specific parameter values according to the specific technology stack, team size and compliance constraints of the project in this industry.
[0020] The link generation unit is used to generate reverse hyperlinks for the instantiated candidate standard skeleton, pointing to its original reuse source and the implicit association features on which it is based, to support subsequent standard evolution and auditing.
[0021] Preferably, it further includes:
[0022] The learning feedback module, connected to the output module and the rule mining module respectively, is used to collect execution deviation data of the executable industry development standard document in the actual development process, and feed the execution deviation data as new labeled samples back to the rule mining module to update the time decay weight of the reused rule cluster.
[0023] Secondly, it provides a method for generating intelligent development standards that can be reused across industries, including the following steps:
[0024] Step 1: Receive the development data source of this industry, and perform semantic annotation based on the industry ontology library to extract at least one implicit association feature of this industry. The implicit association feature is used to characterize the non-linear dependency relationship between different development activities.
[0025] Step 2: Based on the implicit association features, perform cross-project similarity calculations in the historical development standard library and generate at least one reuse rule cluster with time decay weights;
[0026] Step 3: Perform static conflict detection and dynamic collaborative evolution analysis on the reuse rule cluster, and adjust the structure according to the industry priority arbitration strategy to generate a conflict-free candidate standard skeleton.
[0027] Step 4: Instantiate the candidate standard skeleton with parameters according to the specific project context of this industry, and output an executable industry development standard document.
[0028] Preferably, step one further includes:
[0029] Heterogeneous multimodal fusion is performed on the aforementioned development data sources to generate a unified industry feature tensor;
[0030] By using a graph neural network to propagate cross-project information from the industry feature tensor, implicit cross-project design patterns or anti-patterns are extracted as implicit correlation features.
[0031] Preferably, step three further includes:
[0032] Based on the development status transition diagram of this industry, the execution effect of the reuse rule cluster in the future time window is simulated, and the decision boundary of the industry priority arbitration strategy is adaptively adjusted by the reinforcement learning agent to minimize the expected conflict rate.
[0033] Preferably, step four further includes:
[0034] Step 5: Collect deviation data of the executable industry development standard document in actual execution, use the deviation data as a feedback signal to incrementally update the time decay weight of the reuse rule cluster, and re-execute Step 3, thereby realizing the closed-loop adaptive evolution of the development standard generation system.
[0035] Compared with the prior art, the present invention has the following beneficial effects: Attached Figure Description
[0036] Figure 1 This is a schematic diagram of the method flow of the present invention. Detailed Implementation
[0037] The following description is intended to disclose the invention and enable those skilled in the art to implement it. The preferred embodiments described below are merely examples, and other obvious variations will occur to those skilled in the art.
[0038] An intelligent development standard generation system for industry reuse includes:
[0039] The feature extraction module is used to receive development data sources from the industry and perform semantic annotation on the development data sources according to the industry ontology library pre-built in the industry to extract the implicit association features of the industry. The implicit association features are used to characterize the non-linear dependency relationship between different development activities in the industry.
[0040] The rule mining module, coupled with the feature extraction module, is used to perform cross-project similarity calculations based on implicit association features in a historical development standard library to generate reusable rule clusters, which include multiple standard fragments with time decay weights.
[0041] The detection and adjustment module is used to perform static conflict detection and dynamic collaborative evolution analysis on the reuse rule cluster. When a conflict is detected, the reuse rule cluster is structurally adjusted according to the set industry priority arbitration strategy to generate a conflict-free candidate standard skeleton.
[0042] The output module is used to instantiate the candidate standard skeleton with parameters according to the specific project context of this industry to generate an executable industry development standard document.
[0043] The feature extraction module first receives development data from the target industry and uses a pre-built industry ontology library to semantically annotate the data source, thereby extracting implicit association features. These features differ from explicit dependencies; instead, they represent non-linear dependencies obtained through data mining. The rule mining module, coupled to the feature extraction module, performs cross-project similarity calculations based on implicit association features in the historical development standard library. It aggregates similar standard fragments into reusable rule clusters and assigns a weight that decays over time to each fragment, giving higher priority to fragments that have performed well recently. The detection and adjustment module performs static conflict detection and dynamic co-evolution analysis on the reusable rule clusters, automatically adjusting the rule structure according to a preset industry priority arbitration strategy, and outputting conflict-free candidate standard skeletons. Finally, the output module instantiates the abstract parameters in the skeleton according to the specific context of the current project, generating a directly executable standard document. Through the collaborative work of these four modules, fully automated and intelligent generation of development standards is achieved. The cross-project reusable rule clusters and time-decay weights help avoid repetitive work and prioritize the latest and most effective practices. The conflict detection and dynamic analysis module eliminates logical contradictions and potential future conflicts in advance, which helps reduce rework costs after the standard is implemented.
[0044] As an optional embodiment, the feature extraction module further includes:
[0045] Multimodal data units are used to heterogeneously fuse unstructured text, semi-structured configuration files, and structured code graphs from development data sources to generate a unified industry feature tensor.
[0046] The mining unit uses a graph neural network to perform hierarchical propagation learning on the industry feature tensor to extract implicit design patterns and anti-patterns across projects. Implicit correlation features include implicit design patterns and anti-patterns.
[0047] The multimodal data unit integrates three types of data from the development data source into a unified industry feature tensor through entity alignment and feature projection techniques. The three types of data are unstructured text (e.g., emails, design documents), semi-structured configuration files (e.g., deployment scripts in YAML and JSON formats), and structured code graphs (e.g., abstract syntax trees, method call graphs). The mining unit uses a graph neural network to perform hierarchical message propagation on the graph formed by the industry feature tensor. Each node represents a development activity or artifact, and edges represent dependencies. After multiple rounds of neighbor aggregation, the GNN can learn high-order patterns across projects and output implicit design patterns and anti-patterns. These patterns together constitute part of the implicit association features. Through multimodal fusion and graph neural networks, the system's ability to understand complex industry data is improved. The introduction of graph neural networks enables the system to capture deep structural patterns across projects, rather than just surface statistical features. The mined design patterns can be directly used as high-quality reusable materials, while anti-patterns can be used as a negative list for automatic avoidance, thereby improving the advancement and reliability of the generated standards.
[0048] As an optional embodiment, the rule mining module further includes:
[0049] The time decay weight is dynamically calculated based on the actual number of times the standard fragment has been reused in historical applications, the reuse success rate, and the timestamp of the last reuse, according to a non-linear exponential decay function, so that the most recent and high-success-rate standard fragments are given higher reuse priority.
[0050] In the rule mining module, the weight of each standard segment is not static, but dynamically calculated based on three historical metrics: actual reuse count, reuse success rate, and the timestamp of the last reuse. The system employs a non-linear exponential decay function, for example, weight = initial weight · e. -λ·∆tThe success rate coefficient, where ∆t is the difference between the current time and the last reuse time, and λ is the decay rate, causes the weight of segments reused for longer periods to decrease exponentially, while recently reused segments maintain a higher weight. At the same time, segments with high success rates will receive additional positive gains. Ultimately, when generating reuse rule clusters, segments with high weights are preferentially selected into the clusters. The system can automatically adapt to the rapid iteration of industry technologies, allowing outdated standards to naturally decline. Combined with the success rate factor, it helps to reduce the risk of pursuing new standards while neglecting stability.
[0051] As an optional embodiment, the detection and adjustment module further includes:
[0052] The conflict resolution unit is used to detect logical contradictions within the same reuse rule cluster and to detect temporal constraint conflicts between different reuse rule clusters. It also automatically deletes or modifies conflicting standard segments according to a pre-set industry priority strategy.
[0053] The verification unit is used to simulate the execution effect of the reuse rule cluster in a future time window based on the development status change diagram of this industry, and to identify and eliminate implicit mismatches caused by changes in the industry environment.
[0054] The conflict resolution unit detects logical contradictions within the same reused rule cluster and temporal constraint conflicts between different rule clusters. Upon detection of a conflict, the system automatically deletes or modifies low-priority conflicting segments according to a pre-set industry priority strategy. The verification unit simulates the execution effect of the reused rule cluster in multiple future time windows based on the industry development state transition graph. The industry development state transition graph is a directed graph where nodes represent development stages and edges represent state transition probabilities. Through simulation, it can identify implicit mismatches that are currently not conflicting but will become incompatible in the future due to changes in the industry environment, and eliminate them in advance. The automatic detection of logical contradictions and temporal conflicts avoids low-level errors that are easily overlooked during manual review. The dynamic collaborative evolution analysis standard is not only suitable for the present but also for upcoming changes in the industry.
[0055] As an optional embodiment, the output module further includes:
[0056] The parameter filling unit is used to map the abstract variables in the candidate standard skeleton to specific parameter values based on the specific technology stack, team size and compliance constraints of the project in this industry.
[0057] The link generation unit is used to generate reverse hyperlinks for the instantiated candidate standard skeleton, pointing to its original reuse source and the implicit related features on which it is based, to support subsequent standard evolution and auditing.
[0058] Based on the specific technology stack, team size, and compliance constraints of the current industry project, each variable is mapped to a specific parameter value through a rule engine or lightweight machine learning model. The link generation unit generates a reverse hyperlink for each instantiated standard entry. These hyperlinks are embedded in the output document and can be automatically parsed by subsequent evolution modules or auditing tools. This allows the same standard skeleton to adapt to the diverse actual situations of projects. The generation of reverse hyperlinks provides complete traceability for the standard. When the standard causes problems, the source can be quickly traced back to the original basis for easy correction.
[0059] As an optional embodiment, it also includes:
[0060] The learning feedback module, connected to the output module and the rule mining module respectively, is used to collect execution deviation data of executable industry development standard documents in the actual development process. The execution deviation data is fed back to the rule mining module as new labeled samples to update the time decay weight of the reuse rule cluster.
[0061] During actual development, the system collects execution deviation data, which includes deviations made by the team when following the generated standards and the project results after the deviations. This deviation data is used as new labeled samples and fed back to the rule mining module. The rule mining module updates the time decay weight of each standard segment in the reuse rule cluster accordingly. If a segment is frequently deliberately deviated from by the team and the result is better, the system automatically reduces its weight. If a segment is strictly followed and the result is good, its weight is increased. This enables the system to learn from practice and form a positive cycle of self-evolution.
[0062] like Figure 1 The method for generating intelligent development standards for industry reuse, as shown, includes the following steps:
[0063] Step 1: Receive the development data source of this industry, and perform semantic annotation based on the industry ontology library to extract at least one implicit association feature of this industry. The implicit association feature is used to characterize the non-linear dependency relationship between different development activities.
[0064] Step 2: Based on the implicit association features, perform cross-project similarity calculations in the historical development standard library and generate at least one reuse rule cluster with time decay weights;
[0065] Step 3: Perform static conflict detection and dynamic collaborative evolution analysis on the reuse rule cluster, and adjust the structure according to the industry priority arbitration strategy to generate a conflict-free candidate standard skeleton.
[0066] Step 4: Instantiate the candidate standard skeleton with parameters according to the specific project context of this industry, and output an executable industry development standard document.
[0067] Through a four-step process, the method can be quickly deployed and used, gaining the ability to automatically generate development standards, thereby improving the efficiency of standard specification.
[0068] As an optional embodiment, step one further includes:
[0069] Heterogeneous multimodal fusion of development data sources generates a unified industry feature tensor;
[0070] By using graph neural networks to propagate industry feature tensors across projects, implicit cross-project design patterns or anti-patterns are extracted as implicit correlation features.
[0071] As an optional embodiment, step three further includes:
[0072] Based on the development status transition diagram of this industry, the execution effect of the reuse rule cluster in the future time window is simulated, and the decision boundary of the industry priority arbitration strategy is adaptively adjusted by the reinforcement learning agent to minimize the expected conflict rate.
[0073] As an optional embodiment, step four further includes:
[0074] Step 5: Collect deviation data of executable industry development standard documents in actual implementation, use the deviation data as feedback signal to incrementally update the time decay weight of reuse rule clusters, and re-execute Step 3 to achieve closed-loop adaptive evolution of the development standard generation system.
[0075] Through iterative improvements, the solution gains the ability to continuously learn. By introducing incremental updates, it becomes suitable for dynamic and large-scale industry environments. Re-executing step three ensures that the evolution process always meets conflict constraints and will not generate new contradictions due to changes in weights. This closed-loop design allows the standard to evolve in sync with the growth of the industry and the organization, which helps to ensure that the standard is always the latest and most suitable.
[0076] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely principles of the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the claimed invention.
Claims
1. An intelligent development standard generation system for industry reuse, characterized in that, include: The feature extraction module is used to receive development data sources from the industry and perform semantic annotation on the development data sources according to the industry ontology library pre-built in the industry, so as to extract the implicit association features of the industry. The implicit association features are used to characterize the non-linear dependency relationship between different development activities in the industry. The rule mining module, coupled to the feature extraction module, is used to perform cross-project similarity calculations in the historical development standard library based on the implicit association features to generate a reuse rule cluster, which includes multiple standard segments with time decay weights. The detection and adjustment module is used to perform static conflict detection and dynamic collaborative evolution analysis on the reuse rule cluster. When a conflict is detected, the reuse rule cluster is structurally adjusted according to the set industry priority arbitration strategy to generate a conflict-free candidate standard skeleton. The output module is used to instantiate the candidate standard skeleton with parameters according to the specific project context of this industry, so as to generate an executable industry development standard document. 2.The industry-multipurpose-oriented intelligent development standard generation system according to claim 1, wherein, The feature extraction module further includes: A multimodal data unit is used to heterogeneously fuse unstructured text, semi-structured configuration files, and structured code graphs from the development data source to generate a unified industry feature tensor. The mining unit performs hierarchical propagation learning on the industry feature tensor through a graph neural network to extract implicit design patterns and anti-patterns across projects. The implicit correlation features include the implicit design patterns and the anti-patterns. 3.The industry-multipurpose-oriented intelligent development standard generation system according to claim 2, characterized in that, The rule mining module also includes: The time decay weight is dynamically calculated based on the actual number of times the standard fragment has been reused in historical applications, the reuse success rate, and the timestamp of the last reuse, according to a nonlinear exponential decay function, so that the standard fragments that are recent and have a high success rate receive higher reuse priority.
4. The intelligent development standard generation system for industry reuse according to claim 1, characterized in that, The detection and adjustment module also includes: The conflict resolution unit is used to detect logical contradictions within the same reuse rule cluster and to detect temporal constraint conflicts between different reuse rule clusters, and to automatically delete or modify the conflicting standard segments according to a pre-set industry priority strategy. The verification unit is used to simulate the execution effect of the reuse rule cluster in a future time window based on the development status change diagram of this industry, and to identify and eliminate implicit mismatches caused by changes in the industry environment.
5. The intelligent development standard generation system for industry reuse according to claim 1, characterized in that, The output module also includes: The parameter filling unit is used to map the abstract variables in the candidate standard skeleton to specific parameter values according to the specific technology stack, team size and compliance constraints of the project in this industry. The link generation unit is used to generate reverse hyperlinks for the instantiated candidate standard skeleton, pointing to its original reuse source and the implicit association features on which it is based, to support subsequent standard evolution and auditing.
6. The intelligent development standard generation system for industry reuse according to claim 1, characterized in that, Also includes: The learning feedback module, connected to the output module and the rule mining module respectively, is used to collect execution deviation data of the executable industry development standard document in the actual development process, and feed the execution deviation data as new labeled samples back to the rule mining module to update the time decay weight of the reused rule cluster.
7. A method for generating intelligent development standards for industry reuse, applicable to the intelligent development standard generation system for industry reuse described in any one of claims 1-6, characterized in that, Includes the following steps: Step 1: Receive the development data source of this industry, and perform semantic annotation based on the industry ontology library to extract at least one implicit association feature of this industry. The implicit association feature is used to characterize the non-linear dependency relationship between different development activities. Step 2: Based on the implicit association features, perform cross-project similarity calculations in the historical development standard library and generate at least one reuse rule cluster with time decay weights; Step 3: Perform static conflict detection and dynamic collaborative evolution analysis on the reuse rule cluster, and adjust the structure according to the industry priority arbitration strategy to generate a conflict-free candidate standard skeleton. Step 4: Instantiate the candidate standard skeleton with parameters according to the specific project context of this industry, and output an executable industry development standard document.
8. The method for generating intelligent development standards for industry reuse according to claim 7, characterized in that, Step one also includes: Heterogeneous multimodal fusion is performed on the aforementioned development data sources to generate a unified industry feature tensor; By using a graph neural network to propagate cross-project information from the industry feature tensor, implicit cross-project design patterns or anti-patterns are extracted as implicit correlation features.
9. The method for generating intelligent development standards for industry reuse according to claim 7, characterized in that, Step three also includes: Based on the development status transition diagram of this industry, the execution effect of the reuse rule cluster in the future time window is simulated, and the decision boundary of the industry priority arbitration strategy is adaptively adjusted by the reinforcement learning agent to minimize the expected conflict rate.
10. The method for generating intelligent development standards for industry reuse according to claim 7, characterized in that, Step four is followed by: Step 5: Collect deviation data of the executable industry development standard document in actual execution, use the deviation data as a feedback signal to incrementally update the time decay weight of the reuse rule cluster, and re-execute Step 3, thereby realizing the closed-loop adaptive evolution of the development standard generation system.