Monte Carlo Quasi-Heatbath by approximate inversion

Abstract

When sampling the distribution P(phi) ~ exp(-|A phi|^2), a global heatbath normally proceeds by solving the linear system A phi = eta, where eta is a normal Gaussian vector, exactly. This paper shows how to preserve the distribution P(phi) while solving the linear system with arbitrarily low accuracy. Generalizations are presented.

Comments: 10 pages, 1 figure; typos corrected, reference added; version to appear in Phys. Rev. E