Symmetry projection schemes for Gaussian Monte Carlo methods

TL;DR

This paper introduces symmetry projection schemes to improve the accuracy of Gaussian Monte Carlo methods for the Hubbard model by ensuring the density matrix respects the Hamiltonian's symmetries.

Contribution

The paper proposes symmetry projection techniques to enhance Gaussian Monte Carlo methods, addressing previous inaccuracies in ground state property estimations.

Findings

Symmetry projection schemes improve ground state correlation function accuracy.

The method ensures the density matrix maintains Hamiltonian symmetries.

Enhanced reliability of Monte Carlo estimates for the Hubbard model.

Abstract

A novel sign-free Monte Carlo method for the Hubbard model has recently been proposed by Corney and Drummond. High precision measurements on small clusters show that ground state correlation functions are not correctly reproduced. We argue that the origin of this mismatch lies in the fact that the low temperature density matrix does not have the symmetries of the Hamiltonian. Here we show that supplementing the algorithm with symmetry projection schemes provides reliable and accurate estimates of ground state properties.