We investigate yet another direct link between linear algebra and electronic structure theory by exploring the usefulness of the Gershgorin Theorem in the estimation of the correlation energies of many-particle systems. For a case study, we focus on a particular class of systems, two-dimensional quantum dots, and perform extensive numerical calculations by means of a 2p2h Configuration Interaction scheme. We find that the so-called Gershgorin radii constitute natural bounds for the correlation energy though –unfortunately– very loose bounds. However, we show that both quantities can be related through a scaling relation. This result opens the path to further investigations and has perspectives in the design of, e.g., highly parallelized algorithms. Furthermore, this conclusion applies only to two-dimensional (parabolic) quantum dots, and the situation might differ in different classes of systems. This possibility is left as an open question. \(^1\) (The initial exploration of Gershgorin bounds for electronic correlation was conducted by one of us (A.O.) during doctoral research. While many procedural aspects in the present study closely follow -for obvious reasons- the framework outlined in Sect. 3.1.1 of Ref. [1], the original analysis faced limitations due to inconsistencies in numerical convergence. As a result, a comprehensive review incorporating fully converged ground-state energies remained an open issue and was never formally published in paper format.)

错误:搜索内容不能为空,请输入英文关键词
错误:关键词超出字数限制,请精简
高级检索

Bounding the Electronic Correlation with Basic Linear Algebra

  • A. Odriazola,
  • N. C. Hernández

摘要

We investigate yet another direct link between linear algebra and electronic structure theory by exploring the usefulness of the Gershgorin Theorem in the estimation of the correlation energies of many-particle systems. For a case study, we focus on a particular class of systems, two-dimensional quantum dots, and perform extensive numerical calculations by means of a 2p2h Configuration Interaction scheme. We find that the so-called Gershgorin radii constitute natural bounds for the correlation energy though –unfortunately– very loose bounds. However, we show that both quantities can be related through a scaling relation. This result opens the path to further investigations and has perspectives in the design of, e.g., highly parallelized algorithms. Furthermore, this conclusion applies only to two-dimensional (parabolic) quantum dots, and the situation might differ in different classes of systems. This possibility is left as an open question. \(^1\) (The initial exploration of Gershgorin bounds for electronic correlation was conducted by one of us (A.O.) during doctoral research. While many procedural aspects in the present study closely follow -for obvious reasons- the framework outlined in Sect. 3.1.1 of Ref. [1], the original analysis faced limitations due to inconsistencies in numerical convergence. As a result, a comprehensive review incorporating fully converged ground-state energies remained an open issue and was never formally published in paper format.)