Approximating matrix functions by block Krylov methods with randomized vectors

DOI10.48550/ARXIV.2608.13714MaRDI QIDQ7284889arXivFDO

Authors Lucas Onisk, Lothar Reichel, Josh Kane, G. Rodriguez, Giuseppe Rodriguez

Publication date 13 August 2026

Published in arXiv

Tldr This paper explores randomized block Krylov methods to approximate $f(A)\mathbf{b}$ where $A$ is large, by evaluating $f$ at a small matrix. Computed examples suggest that adding $\mathbf{b}$ and random vectors to the initial block vector can reduce computing time and number of steps compared with standard Krylov methods.

Full work available at URL https://doi.org/10.48550/ARXIV.2608.13714

Mathematics Subject Classification ID

Numerical computation of eigenvalues and eigenvectors of matrices (65F15) Ill-posedness and regularization problems in numerical linear algebra (65F22) Numerical computation of matrix exponential and similar matrix functions (65F60)

arXiv classification

math.NA

Cites work

• A Technique for the Numerical Solution of Certain Integral Equations of the First Kind
• Accelerating a restarted Krylov method for matrix functions with randomization
• Analysis of directed networks via the matrix exponential
• Analysis of Some Krylov Subspace Approximations to the Matrix Exponential Operator
• Computational aspects of simultaneous Gaussian quadrature
• Computational methods for large-scale inverse problems: a survey on hybrid projection methods
• Dealing with linear dependence during the iterations of the restarted block Lanczos methods
• Efficient Solution of Parabolic Equations by Krylov Approximation Methods
• Enhancement of Krylov Subspace Spectral Methods by Block Lanczos Iteration
• Error Estimates and Evaluation of Matrix Functions via the Faber Transform
• Finding structure with randomness: probabilistic algorithms for constructing approximate matrix decompositions
• Functions of Matrices
• Global FOM and GMRES algorithms for matrix equations
• Improvement of the resolution of an instrument by numerical solution of an integral equation
• IRBL: An Implicitly Restarted Block-Lanczos Method for Large-Scale Hermitian Eigenproblems
• Iterative exponential filtering for large discrete ill-posed problems
• Iterative Methods for Sparse Linear Systems
• Lanczos-based exponential filtering for discrete ill-posed problems
• Matrix functions in network analysis
• Network properties revealed through matrix functions
• Numerical considerations of block GMRES methods when applied to linear discrete ill-posed problems
• Oblique projection methods for linear systems with multiple right-hand sides
• On computing the zeros of a class of Sobolev orthogonal polynomials
• On Krylov Subspace Approximations to the Matrix Exponential Operator
• On short recurrence Krylov type methods for linear systems with many right-hand sides
• On the global Krylov subspace methods for solving general coupled matrix equations
• Preconditioning without a preconditioner: faster ridge-regression and Gaussian sampling with randomized block Krylov subspace methods
• Randomized algorithms for low-rank matrix approximation: Design, analysis, and applications
• Randomized block Krylov methods for approximating extreme eigenvalues
• Randomized block-Krylov subspace methods for low-rank approximation of matrix functions
• Randomized Sketching for Krylov Approximations of Large-Scale Matrix Functions
• Range restricted iterative methods for linear discrete ill-posed problems
• Rank-Deficient and Discrete Ill-Posed Problems
• Regularization of inverse problems by an approximate matrix-function technique
• Robust port-Hamiltonian representations of passive systems
• Solution methods for linear discrete ill-posed problems for color image restoration
• Solution of large linear discrete ill-posed problems by randomized block Krylov methods
• The Use of Auto-correlation for Pseudo-rank Determination in Noisy III-conditioned Linear Least-squares Problems

This page was built for publication: Approximating matrix functions by block Krylov methods with randomized vectors

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q7284889)