arXiv · 1905.00995
Beyond Walkers in Stochastic Quantum Chemistry: Reducing Error using Fast Randomized Iteration
Abstract
We introduce a family of methods for the full configuration interaction problem in quantum chemistry, based on the fast randomized iteration (FRI) framework [L.-H. Lim and J. Weare, SIAM Rev. 59, 547 (2017)]. These methods, which we term "FCI-FRI," stochastically impose sparsity during iterations of the power method and can be viewed as a generalization of full configuration interaction quantum Monte Carlo (FCIQMC) without walkers. In addition to the multinomial scheme commonly used to sample excitations in FCIQMC, we present a systematic scheme where excitations are not sampled independently. Performing ground-state calculations on five small molecules at fixed cost, we find that the systematic FCI-FRI scheme is 11 to 45 times more statistically efficient than the multinomial FCI-FRI scheme, which is in turn 1.4 to 178 times more statistically efficient than the original FCIQMC algorithm.
Explore related subjects
Keep this discovery
Explore connections, maps & timelines
Samuel M. Greene, Robert J. Webber, Jonathan Weare, Timothy C. Berkelbach. 2019-07-09. Beyond Walkers in Stochastic Quantum Chemistry: Reducing Error using Fast Randomized Iteration. https://doi.org/10.1021/acs.jctc.9b00422
Cite the original work for its findings. Save a collection to share your selection of sources.