mVMC - Open-source software for many-variable variational Monte Carlo method

TL;DR

mVMC is an open-source software that enables efficient variational Monte Carlo calculations for complex interacting fermion systems, supporting a wide range of Hamiltonians with high accuracy.

Contribution

It introduces a user-friendly, flexible implementation of the variational Monte Carlo method with extensive variational parameters and demonstrates its effectiveness on various quantum models.

Findings

High-accuracy ground-state wave functions for fermion systems

Effective optimization with stochastic reconfiguration method

Versatile application to multiple quantum lattice models

Abstract

mVMC (many-variable Variational Monte Carlo) is an open-source software based on the variational Monte Carlo method applicable for a wide range of Hamiltonians for interacting fermion systems. In mVMC, we introduce more than ten thousands variational parameters and simultaneously optimize them by using the stochastic reconfiguration (SR) method. In this paper, we explain basics and user interfaces of mVMC. By using mVMC, users can perform the calculation by preparing only one input file of about ten lines for widely studied quantum lattice models, and can also perform it for general Hamiltonians by preparing several additional input files. We show the benchmark results of mVMC for the Hubbard model, the Heisenberg model, and the Kondo-lattice model. These benchmark results demonstrate that mVMC provides ground-state and low-energy-excited-state wave functions for interacting fermion systems with high accuracy.