Monte Carlo implementation of density-functional theory

Abstract

We propose a conceptually easy and relatively straigthforward numerical method for calculating the ground-state properties of many-particle systems based on the Hohenberg-Kohn theorems. In this "density-functional Monte Carlo" method a direct numerical minimization of the energy functional is performed by a Monte Carlo algorithm in which the density is simulated by a distribution of Bernoulli walkers. The total number of particles is conserved by construction, unlike for other implementations of density-functional theory. The feasibility of the method is illustrated by applying it to a nanoshell. © 2012 American Physical Society.