Quantum Monte Carlo Method for Fermions, Free of Discretization Errors

In this Letter we present a novel quantum Monte Carlo method for fermions, based on an exact decomposition of the Boltzmann operator $\exp (-\beta \hat{H})$. It can be seen as a synthesis of several related methods. It has the advantage that it is free of discretization errors, and applicable to general interactions, both for ground-state and finite-temperature calculations. The decomposition is based on low-rank matrices, which allows faster calculations. As an illustration, the method is applied to an analytically solvable model (pairing in a degenerate shell) and to the Hubbard model.