A method for accelerating the rate of convergence of the long-time (or small- temperature) limit of quantum Monte Carlo approaches is presented. To do that, a variation of the Lanczos algorithm suitable for QMC data is introduced. This algorithm allows one to extract more information from correlation functions at small times, thus avoiding large statistical fluctuations associated with large times. It is first applied to an exactly soluble system and then to the LiH molecule. Calculations using both the fixed-node and nodal-release approaches are discussed.
ASJC Scopus subject areas
- Physics and Astronomy(all)