发表机构
Beihang University; Xi’an Aeronautics Computing Technique Research Institute（北京航空航天大学; 西安航空计算技术研究所）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
提出FlowGen框架，利用LLM语义处理和关系图生成用例流，在分支点预测和备选流生成上显著优于现有方法。
AI 中文摘要
用例流是用例建模的重要元素，因为它们支持下游软件工程活动，包括需求分析、架构与详细设计以及测试用例生成。然而，手动构建用例流成本高昂且需要专业知识，而现有的自动化方法在保持语义一致性、控制流逻辑、数据流逻辑和预期系统边界方面仍存在困难，尤其是在识别分支点和生成备选流时。为解决这一问题，我们提出了FlowGen用于完整的用例流构建。FlowGen利用基于LLM的语义信息处理（SIP）提取语义元素，构建由增强型R-GAT编码的语义关系图（SRG）以进行基本流生成（BFGen），并通过BPP进一步支持分支点预测，通过AFGen支持分支条件下的备选流生成。在13个公共数据集和7个工业数据集上的评估表明，FlowGen在所有三个核心组件中均持续优于竞争基线。特别是，BFGen在精确率上比最佳基线提高14%，召回率提高7-25%，F1分数提高11-30%，AUC提高10-19%；BPP将精确率提高30-110%，召回率提高33-91%，F1分数提高32-117%；AFGen将精确率提高8-23%，F1分数提高5-18%，AUC提高0.6-2.5%。此外，我们验证了基于LLM的SIP模块和BFGen中注意力保留因子的有效性，分析了需求完整性对BFGen的影响，并考察了不同范围的分支相关上下文如何影响AFGen。
英文摘要
Use case flows are important elements of use case modeling because they support downstream software engineering activities, including requirements analysis, architectural and detailed design, and test case generation. However, constructing them manually is costly and expertise-intensive, while existing automated approaches still struggle to preserve semantic consistency, control-flow logic, data-flow logic, and the intended system boundary, especially when identifying branch points and generating alternative flows. To address this problem, we propose FlowGen for complete use case flow construction. FlowGen uses LLM-based Semantic Information Processing (SIP) to extract semantic elements, constructs a Semantic Relational Graph (SRG) encoded by an enhanced R-GAT for basic flow generation (BFGen), and further supports branch point prediction through BPP and branch-conditioned alternative flow generation through AFGen. Evaluations on 13 public and 7 industrial datasets show that FlowGen consistently outperforms competitive baselines in all three core components. In particular, BFGen improves over the best baseline by 14% in Precision, 7-25% in Recall, 11-30% in F1, and 10-19% in AUC; BPP improves Precision by 30-110%, Recall by 33-91%, and F1 by 32-117%; AFGen improves Precision by 8-23%, F1 by 5-18%, and AUC by 0.6-2.5%. Moreover, we validate the effectiveness of the LLM-based SIP module and the attention preservation factor in BFGen, analyze the impact of requirement completeness on BFGen, and examine how different scopes of branch-related context affect AFGen.
Comments19 pages, 8 figures, 6 tables
