ES22-Chen
2022 Workshop on Recent Developments in Electronic Structure (ES22) Poster Session
Author: Chen, Siyuan - College of William & Mary
Title: Accurate structural optimizations in solids with auxiliary-field quantum Monte Carlo
Abstract: Structural optimization by accurate, non-perturbative methods has been an outstanding challenge in many-body electronic structure computations. We present the direct computation of forces and stresses in solids by plane-wave auxiliary-field quantum Monte Carlo (AFQMC) [1]. With them, we perform full structural optimizations in several solids [2]. Additionally, we propose a general optimization algorithm, FSSDxSET, for gradients which have intrinsic stochastic noise. This algorithm is found to outperform standard optimization methods and several machine learning algorithms in efficiency and robustness [3].
References:
[1] S. Zhang and H. Krakauer, Phys. Rev. Lett. 90, 136401 (2003); M. Suewattana, W. Purwanto, S. Zhang, H. Krakauer, and E. J. Walter, Phys. Rev. B 75, 245123 (2007)
[2] S. Chen and S. Zhang, in preparation
[3] S. Chen and S. Zhang, arXiv:2204.12074