Login / Signup

Iterative stochastic subspace self-consistent field method.

Pierre-François LoosJean-Louis RivailXavier Assfeld
Published in: Journal of molecular modeling (2017)
We propose a new self-consistent field (SCF) algorithm based on an iterative, partially stochastic "Divide & Conquer"-type approach. This new SCF algorithm is a simple variant of the usual SCF procedure and can be easily implemented in parallel. A detailed description of the algorithm is reported. We illustrate this new method on one-dimensional hydrogen chains and three-dimensional hydrogen clusters. Graphical Abstract Stochastic partition of the molecular orbitals.
Keyphrases
  • machine learning
  • deep learning
  • image quality
  • neural network
  • minimally invasive
  • density functional theory
  • computed tomography
  • magnetic resonance imaging
  • single molecule
  • molecular dynamics