Localized density matrix minimization and linear-scaling algorithms

JOURNAL OF COMPUTATIONAL PHYSICS

卷:315

Article

Lai, Rongjie¹  Lu, Jianfeng^2,3,4 

[1] Rensselaer Polytech Inst, Dept Math, Troy, NY 12181 USA

[2] Duke Univ, Dept Math, Durham, NC 27706 USA

[3] Duke Univ, Dept Phys, Durham, NC 27706 USA

[4] Duke Univ, Dept Chem, Durham, NC 27706 USA

关键词: Localized density matrix;    l(1) norm;    Hamiltonian;    Finite temperature;    Linear-scaling algorithms;   

DOI  :  10.1016/j.jcp.2016.02.076

来源: Elsevier

【 摘 要 】

We propose a convex variational approach to compute localized density matrices for both zero temperature and finite temperature cases, by adding an entry-wise l(1) regularization to the free energy of the quantum system. Based on the fact that the density matrix decays exponentially away from the diagonal for insulating systems or systems at finite temperature, the proposed l(1) regularized variational method provides an effective 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 l(1) 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. (C) 2016 Elsevier Inc. All rights reserved.

【 授权许可】

Free