PolyJarvis: LLM Agent for Autonomous Polymer MD Simulations

arXiv cs.CL / 4/6/2026

📰 NewsSignals & Early TrendsIdeas & Deep AnalysisModels & Research

共有:

Key Points

PolyJarvis is an LLM-driven agent that performs end-to-end all-atom polymer molecular dynamics workflows from natural-language inputs (polymer name or SMILES).
It integrates an LLM with the RadonPy simulation platform via Model Context Protocol (MCP) servers to autonomously handle tasks including monomer construction, charge assignment, polymerization, force-field parameterization, GPU-accelerated equilibration, and property extraction.
Validation on polyethylene, atactic polystyrene, PMMA, and PEG shows density predictions within 0.1–4.8% and bulk moduli within 17–24% of reference values.
The agent’s PMMA glass transition temperature (395 K) matches experiment within +10–18 K, while other polymers show Tg overestimation (+38 to +47 K), largely explained by MD cooling-rate bias rather than agent mistakes.
Across 8 polymer–property pairs with comparable experimental references, 5 satisfy strict acceptance criteria, supporting the claim that LLM agents can reliably reproduce expert-like polymer MD results.

Abstract

All-atom molecular dynamics (MD) simulations can predict polymer properties from molecular structure, yet their execution requires specialized expertise in force field selection, system construction, equilibration, and property extraction. We present PolyJarvis, an agent that couples a large language model (LLM) with the RadonPy simulation platform through Model Context Protocol (MCP) servers, enabling end-to-end polymer property prediction from natural language input. Given a polymer name or SMILES string, PolyJarvis autonomously executes monomer construction, charge assignment, polymerization, force field parameterization, GPU-accelerated equilibration, and property calculation. Validation is conducted on polyethylene (PE), atactic polystyrene (aPS), poly(methyl methacrylate) (PMMA), and poly(ethylene glycol) (PEG). Results show density predictions within 0.1--4.8% and bulk moduli within 17--24% of reference values for aPS and PMMA. PMMA glass transition temperature (Tg) (395~K) matches experiment within +10--18~K, while the remaining three polymers overestimate Tg by +38 to +47K (vs upper experimental bounds). Of the 8 property--polymer combinations with directly comparable experimental references, 5 meet strict acceptance criteria. For cases lacking suitable amorphous-phase experimental, agreement with prior MD literature is reported separately. The remaining Tg failures are attributable primarily to the intrinsic MD cooling-rate bias rather than agent error. This work demonstrates that LLM-driven agents can autonomously execute polymer MD workflows producing results consistent with expert-run simulations.