TRACE: A spatiotemporal contact memory graph network simulator for granular dynamics
Abstract
Learned graph simulators provide an efficient alternative to high-fidelity solvers for granular dynamics. However, granular motion depends strongly on inter-granular contact history, which is difficult to preserve when particle contacts form, break, and rearrange. Existing simulators mainly store temporal information in node features or node-level memory. Here we introduce TRACE, a graph-network simulator that stores interaction history directly on contact edges. Each edge maintains a persistent memory updated by attention-based message passing and a gated recurrent unit, while an edge-identity dictionary preserves this memory as the contact graph changes. A physics-structured decoder predicts inter-granular normal and tangential contact forces, enforces the Coulomb friction limit, and applies equal-and-opposite internal forces. The model is trained with single-step pretraining followed by autoregressive rollout fine-tuning. We evaluate TRACE on 2D and 3D granular column-collapse benchmarks. In both cases, TRACE produces stable, physically consistent long-horizon rollouts, closely reproducing the final deposit geometry and the kinetic energy released during collapse. Compared with graph network simulator (GNS) and node-memory graph neural simulator (NMGNS), TRACE reduces long-rollout position error by 31-62% and final-deposit error by 58-89% across the two benchmarks, while using fewer parameters and maintaining near-zero particle interpenetration. TRACE also achieves 12.2× and 8.9× speedups over the material point method (MPM) reference solver in 2D and 3D, respectively. Our code is available at https://github.com/Data-Driven-Computational-Geotechnics/TRACE.
Related
- Spatiotemporal Dynamics of Conway's Game of Life in the Entropy--Complexity Plane
- Dynamics of Secondary Infectious Diseases with Spatiotemporal Heterogeneity
- Toward Human-Perceptive 6G: Privacy-Preserving Network Intelligence from Human Dynamics
- Learning Memory-Dependent Neural Network Correlation Potentials for Accurate Electron Dynamics
- Stochastic dynamics generation of directed percolation using a two-channel convolutional neural network
BibTeX
@misc{zhou2026trace,
title = {{TRACE: A spatiotemporal contact memory graph network simulator for granular dynamics}},
author = {Changjian Zhou and Negin Yousefpour and Jie Qi and Junfeng Fang and Guillermo A. Narsilio and Hans Petter Jostad},
howpublished = {arXiv},
year = {2026},
month = sep,
doi = {10.48550/arxiv.2609.02991},
eprint = {2609.02991},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2609.02991},
}
RIS
TY - UNPB
TI - TRACE: A spatiotemporal contact memory graph network simulator for granular dynamics
AU - Changjian Zhou
AU - Negin Yousefpour
AU - Jie Qi
AU - Junfeng Fang
AU - Guillermo A. Narsilio
AU - Hans Petter Jostad
PY - 2026
DA - 2026/09/26
PB - arXiv
DO - 10.48550/arxiv.2609.02991
UR - https://arxiv.org/abs/2609.02991
AB - Learned graph simulators provide an efficient alternative to high-fidelity solvers for granular dynamics. However, granular motion depends strongly on inter-granular contact history, which is difficult to preserve when particle contacts form, break, and rearrange. Existing simulators mainly store temporal information in node features or node-level memory. Here we introduce TRACE, a graph-network simulator that stores interaction history directly on contact edges. Each edge maintains a persistent memory updated by attention-based message passing and a gated recurrent unit, while an edge-identity dictionary preserves this memory as the contact graph changes. A physics-structured decoder predicts inter-granular normal and tangential contact forces, enforces the Coulomb friction limit, and applies equal-and-opposite internal forces. The model is trained with single-step pretraining followed by autoregressive rollout fine-tuning. We evaluate TRACE on 2D and 3D granular column-collapse benchmarks. In both cases, TRACE produces stable, physically consistent long-horizon rollouts, closely reproducing the final deposit geometry and the kinetic energy released during collapse. Compared with graph network simulator (GNS) and node-memory graph neural simulator (NMGNS), TRACE reduces long-rollout position error by 31-62% and final-deposit error by 58-89% across the two benchmarks, while using fewer parameters and maintaining near-zero particle interpenetration. TRACE also achieves 12.2× and 8.9× speedups over the material point method (MPM) reference solver in 2D and 3D, respectively. Our code is available at https://github.com/Data-Driven-Computational-Geotechnics/TRACE.
ER -