Improved extremal optimization for the Ising spin glass

Abstract

A version of the extremal optimization (EO) algorithm introduced by Boettcher and Percus is tested on 2D and 3D spin glasses with Gaussian disorder. EO preferentially flips spins that are locally ``unfit''; the variant introduced here reduces the probability to flip previously selected spins. Relative to EO, this adaptive algorithm finds exact ground states with a speed-up of order $10^{4}$ ($10^{2}$) for $16^{2}$- ($8^{3}$-) spin samples. This speed-up increases rapidly with system size, making this heuristic a useful tool in the study of materials with quenched disorder.