Tool-Augmented Agent for Closed-loop Optimization, Simulation, and Modeling Orchestration

用于闭环优化、仿真与建模编排的工具增强型智能体

Abstract: Iterative industrial design-simulation optimization is bottlenecked by the CAD-CAE semantic gap: translating simulation feedback into valid geometric edits under diverse, coupled constraints. To fill this gap, we propose COSMO-Agent (Closed-loop Optimization, Simulation, and Modeling Orchestration), a tool-augmented reinforcement learning (RL) framework that teaches LLMs to complete the closed-loop CAD-CAE process.

摘要： 迭代式工业设计-仿真优化面临着 CAD-CAE 语义鸿沟的瓶颈：即在多种耦合约束下，将仿真反馈转化为有效的几何编辑。为了填补这一空白，我们提出了 COSMO-Agent（闭环优化、仿真与建模编排），这是一个工具增强型强化学习（RL）框架，旨在教会大语言模型（LLM）完成闭环的 CAD-CAE 流程。

Specifically, we cast CAD generation, CAE solving, result parsing, and geometry revision as an interactive RL environment, where an LLM learns to orchestrate external tools and revise parametric geometries until constraints are satisfied.

具体而言，我们将 CAD 生成、CAE 求解、结果解析和几何修正建模为一个交互式强化学习环境，其中 LLM 学习如何编排外部工具并修改参数化几何结构，直到满足约束条件为止。

To make this learning stable and industrially usable, we design a multi-constraint reward that jointly encourages feasibility, toolchain robustness, and structured output validity. In addition, we contribute an industry-aligned dataset that covers 25 component categories with executable CAD-CAE tasks to support realistic training and evaluation.

为了使这种学习过程稳定且具备工业可用性，我们设计了一种多约束奖励机制，共同促进了方案的可行性、工具链的鲁棒性以及结构化输出的有效性。此外，我们还贡献了一个符合工业标准的数据集，涵盖了 25 个组件类别及可执行的 CAD-CAE 任务，以支持真实场景下的训练与评估。

Experiments show that COSMO-Agent training substantially improves small open-source LLMs for constraint-driven design, exceeding large open-source and strong closed-source models in feasibility, efficiency, and stability.

实验表明，COSMO-Agent 训练显著提升了小型开源 LLM 在约束驱动设计任务中的表现，在可行性、效率和稳定性方面均超过了大型开源模型及强大的闭源模型。