Efficient n-body simulations using physics informed graph neural networks

• Víctor Ramos Osuna ^[1] ; Alberto Díaz Álvarez ; Raúl Lara Cabrera
  1. [1] Universidad Politécnica de Madrid

     Universidad Politécnica de Madrid

     Madrid, España

     
• Localización: Actas del XVI Congreso Español de Metaheurísticas, Algoritmos Evolutivos y Bioinspirados: (MAEB 2025) 28-30 de mayo, Donostia/San Sebastián / coord. por Leticia Hernando Rodríguez, Josu Ceberio Uribe, Jon Vadillo Jueguen, 2025, ISBN 978-84-1319-656-5, págs. 71-80
• Idioma: inglés
• Enlaces
  • Texto Completo Libro
• Resumen

  • This paper presents a novel approach for accelerating N-bodies simulations by integrating a physics-informed graph neural networks (GNNs) with traditional numerical methods. Our method implements a leapfrog-based simulation engine to generate datasets from diverse astrophysical scenarios which are then transformed into graph representations. A custom-designed GNN is trained to predict particle accelerations with high precision. Experiments, conducted on 60 training and 6 testing simulations spanning from 3 to 500 bodies over 1000 time steps, demonstrate that the proposed model achieves extremely low prediction errors—loss values while maintaining robust long-term stability, with accumulated errors in position, velocity, and acceleration remaining insignificant. Furthermore, our method yields a modest speedup of approximately 17% over conventional simulation techniques.

    These results indicate that the integration of deep learning with traditional physical simulation methods offers a promising pathway to significantly enhance computational efficiency without compromising accuracy