标签
本文提出了一种基于自动机的新型方法，用于从信号时序逻辑规范中进行控制综合，利用强化学习，相比现有方法，提高了鲁棒性分数和满足率。
This paper presents Prove-RT, an LLM-assisted framework for generating Prosa/Rocq mechanized theorem prover scripts for schedulability analysis in real-time systems, achieving a 44.7% success rate on a curated evaluation set.