Sign problem in the numerical simulation of many-electron systems

TLDR: A recent proposal for dealing with the sign problem due to Sorella leads to an uncontrolled approximation for the ground-state energy, and a method for calculating the correction needed to make it exact is presented.

Abstract: We discuss the problems that arise in the numerical simulation of many-electron systems when the measure of the functional integrals is not positive definite. We present theoretical arguments and numerical data which indicate that the expectation value of the sign of the measure decreases exponentially as the inverse temperature \ensuremath{\beta} increases, unless the measure is forced to be positive by an explicit symmetry. We therefore conclude that a recent proposal for dealing with the sign problem due to Sorella et al. Leads to an uncontrolled approximation. In the cases we have studied it is a good approximation for the ground-state energy, and we present a method for calculating the correction needed to make it exact. However, for some physical quantities, such as the d-wave pair field susceptibility, the neglect of signs can yield misleading results.