Analysis and optimization of certain parallel Monte Carlo methods in the low temperature limit

TL;DR

This paper investigates the optimal temperature selection in parallel Monte Carlo methods to reduce metastability effects, demonstrating that geometric temperature ratios are nearly optimal in multi-well models.

Contribution

It introduces an explicit optimization framework for temperature selection in infinite swapping schemes, revealing near-optimal geometric sequences in complex energy landscapes.

Findings

Optimal temperature ratios form a geometric sequence in double well models.

Two sources of variance reduction influence the choice of highest temperature.

Geometric temperature ratios are nearly optimal in multi-well models, with performance gaps decreasing geometrically.

Abstract

Metastability is a formidable challenge to Markov chain Monte Carlo methods. In this paper we present methods for algorithm design to meet this challenge. The design problem we consider is temperature selection for the infinite swapping scheme, which is the limit of the widely used parallel tempering scheme obtained when the swap rate tends to infinity. We use a recently developed tool for the analysis of the empirical measure of a small noise diffusion to transform the variance reduction problem into an explicit optimization problem. Our first analysis of the optimization problem is in the setting of a double well model, and it shows that the optimal selection of temperature ratios is a geometric sequence except possibly the highest temperature. In the same setting we identify two different sources of variance reduction, and show how their competition determines the optimal highest temperature. In the general multi-well setting we prove that a pure geometric sequence of temperature ratios is always nearly optimal, with a performance gap that decays geometrically in the number of temperatures.