Localized density matrix minimization and linear scaling algorithms
arXiv:1506.01610 · doi:10.1016/j.jcp.2016.02.076
Abstract
We propose a convex variational approach to compute localized density matrices for both zero temperature and finite temperature cases, by adding an entry-wise regularization to the free energy of the quantum system. Based on the fact that the density matrix decays exponential away from the diagonal for insulating system or system at finite temperature, the proposed regularized variational method provides a nice way to approximate the original quantum system. We provide theoretical analysis of the approximation behavior and also design convergence guaranteed numerical algorithms based on Bregman iteration. More importantly, the regularized system naturally leads to localized density matrices with banded structure, which enables us to develop approximating algorithms to find the localized density matrices with computation cost linearly dependent on the problem size.
arXiv comment22 pages, 7 figures
References in corpus (7)
- The SIESTA method for ab initio order-N materials simulation
- Efficient index handling of multidimensional periodic boundary conditions
- O(N) methods in electronic structure calculations
- Compressed Modes for Variational Problems in Mathematics and Physics
- Adaptive local basis set for Kohn-Sham density functional theory in a discontinuous Galerkin framework I: Total energy calculation
- Expansion algorithm for the density matrix
- A Nonlinear Model for Relativistic Electrons at Positive Temperature