Decorrelation of samples in Quantum Monte Carlo calculations and scaling of autocorrelation time in Li and H$_{2}$O clusters

TL;DR

This paper explores how the autocorrelation time in Quantum Monte Carlo calculations scales with system size, enabling accurate energy estimations for large clusters without the need for sample decorrelation.

Contribution

It introduces a method to perform accurate QMC energy calculations by leveraging the scaling behavior of autocorrelation time, reducing the need for data binning in large systems.

Findings

Autocorrelation time scales with system size in large clusters.

Sample decorrelation becomes impractical for large systems.

Accurate energies achieved without decorrelation by exploiting autocorrelation scaling.

Abstract

We have investigated decorrelation of samples in Quantum Monte Carlo (QMC) ground-state energy calculations for large Li and H$_{2}$O nanoclusters. Binning data as a way of eliminating statistical correlations, as is the common practice, is found to become increasingly impractical as the system size grows. We demonstrate nevertheless that it is possible to perform accurate energy calculations - without decorrelating samples - by exploiting the scaling of the integrated autocorrelation time $\tau$ as a function of the number of electrons in the system.