Accelerating Multicanonical Monte Carlo Simulations with Irreversibility

Monte Carlo simulations are robust methods to study statistical physics. However, the unpredictable convergence time and the ease of being trapped in local minima have plagued the efficiency of both traditional and modern Monte Carlo algorithms. We review and design new strategies of introducing irreversibility to suppress the random walk behavior in Monte Carlo simulations. These strategies violate detailed balance condition, yet they satisfy the global balance condition that ensures correct statistics. When applied to multicanonical sampling, our new strategies showed significant speedup compared to the original algorithm. We will demonstrate the advantage of our new algorithms with Monte Carlo simulations on spin systems and alloys.