Kandolf, Peter and Relton, Samuel D. (2016) A Block Krylov Method to Compute the Action of the Frechet Derivative of a Matrix Function on a Vector with Applications to Condition Number Estimation. [MIMS Preprint]
PDF
paper.pdf Download (381kB) |
Abstract
We design a block Krylov method to compute the action of the FrÃ©chet derivative of a matrix function on a vector using only matrixâ��vector products. The algorithm we derive is especially effective when the direction matrix in the derivative is of low rank. Our results and experiments are focused mainly on FrÃ©chet derivatives with rank-1 direction matrices. Our analysis applies to all functions with a power series expansion convergent on a subdomain of the complex plane which, in particular, includes the matrix exponential. We perform an a priori error analysis of our algorithm to obtain rigorous stopping criteria. Furthermore, we show how our algorithm can be used to estimate the 2-norm condition number of f(A)b efficiently. Our numerical experiments show that our new algorithm for computing the action of a FrÃ©chet derivative typically requires a small number of iterations to converge and (particularly for single and half precision accuracy) is significantly faster than alternative algorithms. When applied to condition number estimation our experiments show that the resulting algorithm can detect ill-conditioned problems that are undetected by competing algorithms.
Item Type: | MIMS Preprint |
---|---|
Uncontrolled Keywords: | matrix function, matrix exponential, Krylov subspace, block Krylov subspace, Frechet derivative, condition number |
Subjects: | MSC 2010, the AMS's Mathematics Subject Classification > 65 Numerical analysis |
Depositing User: | Dr Samuel Relton |
Date Deposited: | 31 May 2016 |
Last Modified: | 08 Nov 2017 18:18 |
URI: | https://eprints.maths.manchester.ac.uk/id/eprint/2480 |
Available Versions of this Item
- A Block Krylov Method to Compute the Action of the Frechet Derivative of a Matrix Function on a Vector with Applications to Condition Number Estimation. (deposited 31 May 2016) [Currently Displayed]
Actions (login required)
View Item |