Irreversible simulated tempering

Abstract

An extended ensemble Monte Carlo algorithm is proposed by introducing a violation of the detailed balance condition to the update scheme of the inverse temperature in simulated tempering. Our method, irreversible simulated tempering, is constructed based on the framework of the skew detailed balance condition. By applying this method to the ferromagnetic Ising model in two dimensions on a square lattice as a benchmark, the dynamical behavior of the inverse temperature and an autocorrelation function of the magnetization are studied numerically. It is found that the relaxation dynamics of the inverse temperature changes qualitatively from diffusive to ballistic by violating the detailed balance condition. Consequently, the autocorrelation time of magnetization is several times smaller than that for the conventional algorithm satisfying the detailed balance condition.