Population Annealing and Its Application to a Spin Glass

A way to modify simulated annealing to a Monte Carlo algorithm for calculating canonical averages is presented. The proposed algorithm is based on the idea of population based Monte Carlo method where multiple replicas of the original system are used to represent a target distribution. Inspired by non‐equilibrium work relation of Jarzynski, an appropriate weight of the replicas is introduced, which enable correct computation of the canonical averages in the limit of infinite number of replicas. The method is applied to a spin glass model and its efficiency is discussed.