Nested Reduction Algorithm for Generating \mathcal^-Matrix Representation of Electrically Large Surface Integral Operators from FMM

In this paper, we develop a fast algorithm to generate a rank minimized \mathcal{H}^{2} -matrix from multi-level FMM to represent electrically large surface integral equations (SIEs). The rank of the new \mathcal{H}^{2} -matrix is determined based on prescribed accuracy and is found to be much small...

Full description

Saved in:
Bibliographic Details
Published inDigest - IEEE Antennas and Propagation Society. International Symposium (1995) pp. 1061 - 1062
Main Authors Yang, Chang, Jiao, Dan
Format Conference Proceeding
LanguageEnglish
Published IEEE 05.07.2020
Subjects
Acceleration
Antennas
Integral equations
Meetings
Scattering
Online AccessGet full text
ISSN1947-1491
DOI10.1109/IEEECONF35879.2020.9329651

Cover

More Information
Summary:In this paper, we develop a fast algorithm to generate a rank minimized \mathcal{H}^{2} -matrix from multi-level FMM to represent electrically large surface integral equations (SIEs). The rank of the new \mathcal{H}^{2} -matrix is determined based on prescribed accuracy and is found to be much smaller than the original FMM's rank. The resultant rank-minimized \mathcal{H}^{2} -matrix can be used to accelerate both iterative and direct solutions. Numerical experiments on large-scale surface IE-based scattering analysis having millions of unknowns on a single CPU core have demonstrated its accuracy and efficiency.
ISSN:1947-1491
DOI:10.1109/IEEECONF35879.2020.9329651