Machine Learning Assisted Monte Carlo Simulation: Efficient Overlap Determination for Nonspherical Hard Bodies

Standard molecular dynamics (MD) and Monte Carlo (MC) simulations deal with spherical particles. Extending the standard simulation methodologies to the nonspherical objects is non‐trivial. To circumvent this problem, nonspherical bodies are often treated as a collection of constituent spherical objects. As the number of these constituent objects becomes large, the computational burden to simulate the system also increases. Here, an alternative way is proposed to simulate nonspherical rigid bodies having pairwise repulsive interactions. This approach is based on a machine learning (ML)‐based model, which predicts the overlap between two nonspherical bodies. The ML model is easy to train and the computation cost of its implementation remains independent of the number of constituent spheres used to represent a nonspherical rigid body. When used in MC simulation, this method is faster than the standard implementation, where overlap determination is based on calculating the distance between constituent spheres. This proposed ML‐based MC method produces similar structural features (in comparison to the standard implementation) in both two and three dimensions, and can qualitatively capture the isotropic to nematic transition of rigid rods in three dimensions. It is believed that this work is a step toward a time‐efficient simulation of non‐spherical rigid bodies.