Hierarchical LU Preconditioning for the Time-Harmonic Maxwell Equations

Research output: Chapter in book/report/conference proceedingConference contributionResearchpeer review

View graph of relations

Details

Original languageEnglish
Title of host publicationDomain Decomposition Methods in Science and Engineering XXVII
EditorsZdenek Dostal, Tomas Kozubek, Axel Klawonn, Luca F. Pavarino, Olof B. Widlund, Ulrich Langer, Jakub Sístek
PublisherSpringer Science and Business Media Deutschland GmbH
Pages391-399
Number of pages9
ISBN (print)9783031507687
Publication statusPublished - 2024
Event27th International Conference on Domain Decomposition Methods in Science and Engineering, DD 2022 - Prague, Czech Republic
Duration: 25 Jul 202229 Jul 2022

Publication series

NameLecture Notes in Computational Science and Engineering
Volume149
ISSN (Print)1439-7358
ISSN (electronic)2197-7100

Abstract

The time-harmonic Maxwell equations are used to study the effect of electric and magnetic fields on each other. Although the linear systems resulting from solving this system using FEMs are sparse, direct solvers cannot reach the linear complexity. In fact, due to the indefinite system matrix, iterative solvers suffer from slow convergence. In this work, we study the effect of using the inverse of \(\mathcal{H}\)-matrix approximations of the Galerkin matrices arising from Nédélec's edge FEM discretization to solve the linear system directly. We also investigate the impact of applying an \(\mathcal{H}-LU\) factorization as a preconditioner and we study the number of iterations to solve the linear system using iterative solvers.

Keywords

    math.NA, cs.NA

ASJC Scopus subject areas

Cite this

Hierarchical LU Preconditioning for the Time-Harmonic Maxwell Equations. / Parvizi, Maryam; Khodadadian, Amirreza; Beuchler, Sven et al.
Domain Decomposition Methods in Science and Engineering XXVII. ed. / Zdenek Dostal; Tomas Kozubek; Axel Klawonn; Luca F. Pavarino; Olof B. Widlund; Ulrich Langer; Jakub Sístek. Springer Science and Business Media Deutschland GmbH, 2024. p. 391-399 (Lecture Notes in Computational Science and Engineering; Vol. 149).

Research output: Chapter in book/report/conference proceedingConference contributionResearchpeer review

Parvizi, M, Khodadadian, A, Beuchler, S & Wick, T 2024, Hierarchical LU Preconditioning for the Time-Harmonic Maxwell Equations. in Z Dostal, T Kozubek, A Klawonn, LF Pavarino, OB Widlund, U Langer & J Sístek (eds), Domain Decomposition Methods in Science and Engineering XXVII. Lecture Notes in Computational Science and Engineering, vol. 149, Springer Science and Business Media Deutschland GmbH, pp. 391-399, 27th International Conference on Domain Decomposition Methods in Science and Engineering, DD 2022, Prague, Czech Republic, 25 Jul 2022. https://doi.org/10.48550/arXiv.2211.11303, https://doi.org/10.1007/978-3-031-50769-4_47
Parvizi, M., Khodadadian, A., Beuchler, S., & Wick, T. (2024). Hierarchical LU Preconditioning for the Time-Harmonic Maxwell Equations. In Z. Dostal, T. Kozubek, A. Klawonn, L. F. Pavarino, O. B. Widlund, U. Langer, & J. Sístek (Eds.), Domain Decomposition Methods in Science and Engineering XXVII (pp. 391-399). (Lecture Notes in Computational Science and Engineering; Vol. 149). Springer Science and Business Media Deutschland GmbH. https://doi.org/10.48550/arXiv.2211.11303, https://doi.org/10.1007/978-3-031-50769-4_47
Parvizi M, Khodadadian A, Beuchler S, Wick T. Hierarchical LU Preconditioning for the Time-Harmonic Maxwell Equations. In Dostal Z, Kozubek T, Klawonn A, Pavarino LF, Widlund OB, Langer U, Sístek J, editors, Domain Decomposition Methods in Science and Engineering XXVII. Springer Science and Business Media Deutschland GmbH. 2024. p. 391-399. (Lecture Notes in Computational Science and Engineering). doi: 10.48550/arXiv.2211.11303, 10.1007/978-3-031-50769-4_47
Parvizi, Maryam ; Khodadadian, Amirreza ; Beuchler, Sven et al. / Hierarchical LU Preconditioning for the Time-Harmonic Maxwell Equations. Domain Decomposition Methods in Science and Engineering XXVII. editor / Zdenek Dostal ; Tomas Kozubek ; Axel Klawonn ; Luca F. Pavarino ; Olof B. Widlund ; Ulrich Langer ; Jakub Sístek. Springer Science and Business Media Deutschland GmbH, 2024. pp. 391-399 (Lecture Notes in Computational Science and Engineering).
Download
@inproceedings{70723e9d33144048a6940d7635e49390,
title = "Hierarchical LU Preconditioning for the Time-Harmonic Maxwell Equations",
abstract = "The time-harmonic Maxwell equations are used to study the effect of electric and magnetic fields on each other. Although the linear systems resulting from solving this system using FEMs are sparse, direct solvers cannot reach the linear complexity. In fact, due to the indefinite system matrix, iterative solvers suffer from slow convergence. In this work, we study the effect of using the inverse of \(\mathcal{H}\)-matrix approximations of the Galerkin matrices arising from N{\'e}d{\'e}lec's edge FEM discretization to solve the linear system directly. We also investigate the impact of applying an \(\mathcal{H}-LU\) factorization as a preconditioner and we study the number of iterations to solve the linear system using iterative solvers. ",
keywords = "math.NA, cs.NA",
author = "Maryam Parvizi and Amirreza Khodadadian and Sven Beuchler and Thomas Wick",
note = "Funding Information: Acknowledgements The authors acknowledge the Deutsche Forschungsgemeinschaft (DFG) under Germany Excellence Strategy within the Cluster of Excellence PhoenixD (EXC 2122, Project ID 390833453). Maryam Parvizi is funded by Alexander von Humboldt Foundation project named H-matrix approximability of the inverses for FEM, BEM and FEM-BEM coupling of the electromagnetic problems. Finally, the authors thank Sebastian Kinnewig for fruitful discussions.; 27th International Conference on Domain Decomposition Methods in Science and Engineering, DD 2022 ; Conference date: 25-07-2022 Through 29-07-2022",
year = "2024",
doi = "10.48550/arXiv.2211.11303",
language = "English",
isbn = "9783031507687",
series = "Lecture Notes in Computational Science and Engineering",
publisher = "Springer Science and Business Media Deutschland GmbH",
pages = "391--399",
editor = "Zdenek Dostal and Tomas Kozubek and Axel Klawonn and Pavarino, {Luca F.} and Widlund, {Olof B.} and Ulrich Langer and Jakub S{\'i}stek",
booktitle = "Domain Decomposition Methods in Science and Engineering XXVII",
address = "Germany",

}

Download

TY - GEN

T1 - Hierarchical LU Preconditioning for the Time-Harmonic Maxwell Equations

AU - Parvizi, Maryam

AU - Khodadadian, Amirreza

AU - Beuchler, Sven

AU - Wick, Thomas

N1 - Funding Information: Acknowledgements The authors acknowledge the Deutsche Forschungsgemeinschaft (DFG) under Germany Excellence Strategy within the Cluster of Excellence PhoenixD (EXC 2122, Project ID 390833453). Maryam Parvizi is funded by Alexander von Humboldt Foundation project named H-matrix approximability of the inverses for FEM, BEM and FEM-BEM coupling of the electromagnetic problems. Finally, the authors thank Sebastian Kinnewig for fruitful discussions.

PY - 2024

Y1 - 2024

N2 - The time-harmonic Maxwell equations are used to study the effect of electric and magnetic fields on each other. Although the linear systems resulting from solving this system using FEMs are sparse, direct solvers cannot reach the linear complexity. In fact, due to the indefinite system matrix, iterative solvers suffer from slow convergence. In this work, we study the effect of using the inverse of \(\mathcal{H}\)-matrix approximations of the Galerkin matrices arising from Nédélec's edge FEM discretization to solve the linear system directly. We also investigate the impact of applying an \(\mathcal{H}-LU\) factorization as a preconditioner and we study the number of iterations to solve the linear system using iterative solvers.

AB - The time-harmonic Maxwell equations are used to study the effect of electric and magnetic fields on each other. Although the linear systems resulting from solving this system using FEMs are sparse, direct solvers cannot reach the linear complexity. In fact, due to the indefinite system matrix, iterative solvers suffer from slow convergence. In this work, we study the effect of using the inverse of \(\mathcal{H}\)-matrix approximations of the Galerkin matrices arising from Nédélec's edge FEM discretization to solve the linear system directly. We also investigate the impact of applying an \(\mathcal{H}-LU\) factorization as a preconditioner and we study the number of iterations to solve the linear system using iterative solvers.

KW - math.NA

KW - cs.NA

UR - http://www.scopus.com/inward/record.url?scp=85184280379&partnerID=8YFLogxK

U2 - 10.48550/arXiv.2211.11303

DO - 10.48550/arXiv.2211.11303

M3 - Conference contribution

SN - 9783031507687

T3 - Lecture Notes in Computational Science and Engineering

SP - 391

EP - 399

BT - Domain Decomposition Methods in Science and Engineering XXVII

A2 - Dostal, Zdenek

A2 - Kozubek, Tomas

A2 - Klawonn, Axel

A2 - Pavarino, Luca F.

A2 - Widlund, Olof B.

A2 - Langer, Ulrich

A2 - Sístek, Jakub

PB - Springer Science and Business Media Deutschland GmbH

T2 - 27th International Conference on Domain Decomposition Methods in Science and Engineering, DD 2022

Y2 - 25 July 2022 through 29 July 2022

ER -