Performance Evaluation of State-of-the-Art Linear Iterative Solvers Based on IDR Theorem for Large Scale Electromagnetic Multiple Scattering Simulations

Authors

  • Norimasa Nakashima Faculty of Information Science and Electrical Engineering Kyushu University, Fukuoka, 819-0395, Japan
  • Seiji Fujino Research Institute for Information Technology Kyushu University, Fukuoka, 812-8581, Japan
  • Mitsuo Tateiba Ariake National College of Technology Omuta, 836-8585, Japan

Keywords:

Performance Evaluation of State-of-the-Art Linear Iterative Solvers Based on IDR Theorem for Large Scale Electromagnetic Multiple Scattering Simulations

Abstract

The present paper treats with the latest linear iterative solver IDR(s) method and its variants proposed by P. Sonneveld and M. van Gijzen. We derive preconditioned algorithms of the solvers based on right preconditioning and list them. The solvers are numerically tested in terms of convergence and accuracy for the computation of electromagnetic wave scattering from over 104 dielectric cylinders. Consequently, minimization schemes for residual vectors refine not only convergence but also accuracy for the original IDR(s) method. However, a spurious convergence may be confirmed and its influence is 1 or 2 digit error independently of parameter s.

Downloads

Download data is not yet available.

References

N. Morita, N. Kumagai, and J. R. Mautz,

Integral Equation Methods for Electromagnet-

ics, Artech House, 1990.

N. Nakashima and M. Tateiba, “A Wideband

Fast Multipole Algorithm for Two-

dimensional Volume Integral Equations,” Int.

J. Numer. Meth. Eng., vol. 77, pp. 195-213,

Y. Saad, Iterative Methods for Sparse Linear

Systems, SIAM, 2003.

Y. Saad and M. Schultz, “GMRES: A Genera-

lized Minimal Residual Algorithm for Solving

Nonsymmetric Linear Systems,” SIAM J. Sci.

Stat. Comput., vol. 7, no. 3, pp. 856-869,

N. Nakashima, S. Fujino, and M. Tateiba,

“Fast Computation for Electromagnetic Wave

Scattering Problem with Many Dielectric Cy-

linder (2) --- Two-step Preconditioning Com-

bined with Half Reduction and Block Jacobi,”

Trans. IPSJ, vol. 45, no. SIG1(ACS4), pp. 1-9,

(in Japanese)

P. Sonneveld and M. B. van Gijzen, “IDR(s):

A Family of Simple and Fast Algorithms for

Solving Large Nonsymmetric Systems of Li-

near Equations,” SIAM J. Sci. Comput ., vol.

, pp. 1035-1062, 2008.

N. Nakashima, S. Fujino, M. Tateiba, and Y.

Onoue, “An Application of IDR(s) Method to

Boundary Element Analyses of Two-

dimensional Electromagnetic Multiple Scatter-

ing,” IEICE Trans. Electron., vol. J92-C, no.

, pp. 111-118, 2009. (in Japanese)

G. L. G. Sleijpen and H. A. van der Vorst,

“Maintaining Convergence Properties of

BiCGStab Methods in Finite Precision Arith-

metic,” Numer. Algo., vol. 2, no. 2, pp. 203-

, 1995.

T. Sakurai, K. Naono, M. Egi, M. Igai, H. Ki-

dachi, and M. Shoji, “Proposal and Evaluation

of AC-IDR(s) Method That Combines High-

NAKASHIMA, FUJINO, TATEIBA: EVALUATION OF STATE-OF-THE-ART LINEAR ITERATIVE SOLVERS BASED ON IDR THEOREM

Performance with High Accuracy,” IPSJ Tech.

Rep., HPC, pp. 49-54, 2008. (in Japanese)

M. B. van Gijzen and P. Sonneveld, “An

IDR(s) variant with minimal intermediate re-

sidual norms,” Proc. Int. Kyoto-Forum on

Krylov Subspace Method, pp. 85-92, 2008.

M. B. van Gijzen and P. Sonneveld, “An Ele-

gant IDR(s) Variant that Efficiently Exploits

Bi-orthogonality Properties,” Delft University

of Technology, Reports of the Department of

Applied Mathematical Analysis, Report 08-21,

Downloads

Published

2022-05-02

How to Cite

[1]
N. . Nakashima, S. . Fujino, and M. . Tateiba, “Performance Evaluation of State-of-the-Art Linear Iterative Solvers Based on IDR Theorem for Large Scale Electromagnetic Multiple Scattering Simulations”, ACES Journal, vol. 26, no. 1, pp. 37–44, May 2022.

Issue

Section

General Submission