Approximating the phi-function action on linear combinations of Krylov bases

DOI10.1080/03081087.2025.2567423MaRDI QIDQ6898820zbMATH OpenFDO

Authors F. S. Naranjo-Noda, J. C. Jimenez

Publication date 17 November 2025

Published in Linear and Multilinear Algebra (Search for Journal in Brave)

Full work available at URL https://doi.org/10.1080/03081087.2025.2567423

Mathematics Subject Classification ID

Matrix exponential and similar functions of matrices (15A16) Basic linear algebra (15A99) Numerical computation of matrix exponential and similar matrix functions (65F60) Numerical linear algebra (65F99)

Review text

This interesting paper studies approximating the phi-function action on linear combinations of Krylov bases. To set the scene, let \(\phi(z):=\frac{\exp(z)-1}{z}.\) This is the well known \(\phi\) function which appears in many places for example in exponential-type integrators for differential equations, in the representation of multiple integrals involving matrix exponentials and in many other places. In high dimensional integrals or in large systems of equations, the action of \(\phi\) is typically computed through Krylov-subspace methods and in many cases, the same operator acts on different vectors typically implying the use of Arnoldi algorithms and in this case, natural questions come to mind as in, is it possible to reduce the number of calls to the Arnoldi algorithm and under what conditions. The authors of the paper show that the computation of the \(\phi\) action on different vectors can be carried out with a single Arnoldi algorithm in the case that these vectors are linear combinations of Krylov bases. Errors and approximations are studied on specific examples.\N\NThe paper is well written with a good set of references.

Cites work

• Algorithm 919: A Krylov subspace algorithm for evaluating the -functions appearing in exponential integrators
• Analysis of Some Krylov Subspace Approximations to the Matrix Exponential Operator
• Computing high dimensional multiple integrals involving matrix exponentials
• Efficient computation of phi-functions in exponential integrators
• Expokit
• Local Linearization-Runge Kutta (LLRK) Methods for Solving Ordinary Differential Equations
• Local linearization-Runge-Kutta methods: a class of A-stable explicit integrators for dynamical systems
• Locally linearized Runge Kutta method of Dormand and Prince
• Locally linearized Runge-Kutta method of Dormand and Prince for large systems of initial value problems
• Solving Ordinary Differential Equations I

This page was built for publication: Approximating the phi-function action on linear combinations of Krylov bases

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q6898820)