Non-Reversible Monte Carlo Simulations of Long-Range Interacting Molecular Systems

We present current progress of developing non-reversible Markov-chain Monte Carlo (MCMC) algorithms for efficient simulations of atom-based models of molecules that include long-ranged interactions. The event-chain Monte Carlo (ECMC) algorithm samples the Boltzmann distribution exactly without computing energy changes, which removes the computational bottleneck of traditional reversible MCMC simulations. Also, in contrast to molecular dynamics, the mixing and autocorrelation times of MCMC are not locked to the physical dynamics.

We introduce our open-source JeLLyFysh (JF) application that implements ECMC in a general way [1] by demonstrating number of worked out molecular-simulation examples that include, e.g., liquid water. We then highlight recent improvements of the application and ECMC itself. This includes, in particular, the concept of fast sequential Markov chains where ECMC's direction of motion is sequentially chosen from a set [2]. Choosing a large direction set leads to much shorter mixing times of the rotational degree of freedom, and may thus greatly accelerate ECMC simulations of molecular systems.

[1] P. Höllmer, L. Qin, M. F. Faulkner, A. C. Maggs, and W. Krauth, Comput. Phys. Commun. 253, 107168 (2020)
[2] L. Qin, P. Höllmer, and W. Krauth, arXiv: 2007.15615 (2020)