Mathematical Physics
[Submitted on 28 Feb 2020 (v1), last revised 11 Mar 2020 (this version, v3)]
Title:Dynamical perturbation theory for eigenvalue problems
View PDFAbstract:Many problems in physics, chemistry and other fields are perturbative in nature, i.e. differ only slightly from related problems with known solutions. Prominent among these is the eigenvalue perturbation problem, wherein one seeks the eigenvectors and eigenvalues of a matrix with small off-diagonal elements. Here we introduce a novel iterative algorithm to compute these eigenpairs based on fixed-point iteration for an algebraic equation in complex projective space. We show from explicit and numerical examples that our algorithm outperforms the usual Rayleigh-Schrödinger expansion on three counts. First, since it is not defined as a power series, its domain of convergence is not a priori confined to a disk in the complex plane; we find that it indeed usually extends beyond the standard perturbative radius of convergence. Second, it converges at a faster rate than the Rayleigh-Schrödinger expansion, i.e. fewer iterations are required to reach a given precision. Third, the (time- and space-) algorithmic complexity of each iteration does not increase with the order of the approximation, allowing for higher precision computations. Because this complexity is merely that of matrix multiplication, our dynamical scheme also scales better with the size of the matrix than general-purpose eigenvalue routines such as the shifted QR or divide-and-conquer algorithms. Whether they are dense, sparse, symmetric or unsymmetric, we confirm that dynamical diagonalization quickly outpaces LAPACK drivers as the size of matrices grows; for the computation of just the dominant eigenvector, our method converges order of magnitudes faster than the Arnoldi algorithm implemented in ARPACK.
Submission history
From: Matteo Smerlak [view email][v1] Fri, 28 Feb 2020 17:13:22 UTC (2,134 KB)
[v2] Mon, 2 Mar 2020 08:27:49 UTC (2,134 KB)
[v3] Wed, 11 Mar 2020 15:39:46 UTC (2,134 KB)
Current browse context:
math
References & Citations
Bibliographic and Citation Tools
Bibliographic Explorer (What is the Explorer?)
Connected Papers (What is Connected Papers?)
Litmaps (What is Litmaps?)
scite Smart Citations (What are Smart Citations?)
Code, Data and Media Associated with this Article
alphaXiv (What is alphaXiv?)
CatalyzeX Code Finder for Papers (What is CatalyzeX?)
DagsHub (What is DagsHub?)
Gotit.pub (What is GotitPub?)
Hugging Face (What is Huggingface?)
Papers with Code (What is Papers with Code?)
ScienceCast (What is ScienceCast?)
Demos
Recommenders and Search Tools
Influence Flower (What are Influence Flowers?)
CORE Recommender (What is CORE?)
arXivLabs: experimental projects with community collaborators
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.