ChronoLLM: Customizing Language Models for Physics-Based Simulation Code Generation

2025-08-19

AI Summary

This study addresses the challenge of adapting pre-trained large language models (LLMs) into domain-specific virtual assistants for physics-based simulation. Focusing on PyChrono—a multi-body dynamics simulation framework—we propose the first systematic framework unifying adaptation strategies for both open- and closed-weight LLMs, enabling robust generation of simulation scripts across scenarios ranging from simple pendulums to complex vehicle–terrain interactions. Our methodology integrates instruction fine-tuning, domain-specific data augmentation, explicit injection of PyChrono API knowledge, and a quantitative evaluation protocol grounded in executable correctness and functional accuracy. Empirical results demonstrate substantial improvements over baseline LLMs: generated scripts exhibit markedly higher runtime validity and physical fidelity; the assistant reliably answers API queries, recommends appropriate modeling strategies, and delivers production-ready script templates for expert users. Collectively, this work lowers the entry barrier for cross-domain physics simulation tools while advancing the principled integration of LLMs into scientific computing workflows.

Technology Category

• Machine Learning: Large Multimodal Models (LMMs)
• Cognitive Modeling & Cognitive Systems: Simulating Human Behavior
• Search and Optimization: Sampling/Simulation-based Search

Application Category

• Search and Retrieval-Augmented AI: Search Tool Learning with LLM: Teaching LLMs to invoke search and make use of retrieved information
• User Modeling, Personalization and Recommendation: Large Language Models (LLM) for user modeling and recommendation
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions

Abstract

This contribution is concerned with the following issue: can pretrained large language models (LLMs) be refined and customized to the point where they become virtual assistants helping experts with the effective use of a simulation tool? In this case study, the ``simulation tool'' considered is PyChrono, an open source multi-physics dynamics engine for multibody systems. We present a framework for refining and customizing both open- and closed-source LLMs to harness the power of AI in generating scripts that perform PyChrono virtual experiments. We refine and customize several classes of LLMs through a process that leads to a quantifiable improvement in the quality of the generated PyChrono simulation scripts. These scripts can range from simple single-pendulum simulations to complex virtual experiments involving full vehicles on deformable terrain. While the generated scripts are rarely perfect, they often serve as strong starting points for the user to modify and improve on. Additionally, the LLM can answer specific API questions about the simulator, or recommend modeling approaches. The framework discussed is general and can be applied to lower the entry barrier for simulation tools associated with other application domains.

Problem

• Customizing LLMs for physics simulation code generation
• Improving quality of PyChrono simulation scripts via AI
• Lowering entry barrier for complex simulation tools

Innovation

• Refining LLMs for PyChrono simulation script generation
• Customizing open- and closed-source language models
• Generating physics simulation scripts through AI framework