Multifrontal method preconditioned sparse-matrix/canonical grid algorithm for fast analysis of microstrip structure

Author

Zhuang, W. ; Feng, X.P. ; Mo, L. ; Chen, R.S.

Author_Institution

Dept. of Commun. Eng., Nanjing Univ. of Sci. & Technol., China

Volume

4

fYear

2005

fDate

4-7 Dec. 2005

Abstract

In this paper, the microstrip structure is analyzed in terms of the mixed potential integral equation (MPIE) by using of the Rao-Wilton-Glisson (RWG) function expansion. The integral equation is solved by the sparse-matrix/canonical grid (SMCG) method with fast Fourier transforms technique (FFT) to accelerate the matrix-vector multiplication. During the iterative process, multifrontal method is employed to precondition the matrix equations for the purpose of enhancing the computational efficiency of the conjugate-gradient (CG) algorithm. The numerical calculations show that the proposed preconditioned SMCG algorithm can converge more than 30 times faster than the conventional one for the analysis of microstrip structures.

Keywords

antenna theory; conjugate gradient methods; fast Fourier transforms; integral equations; microstrip antenna arrays; planar antenna arrays; sparse matrices; Rao-Wilton-Glisson function expansion; canonical grid algorithm; conjugate-gradient algorithm; fast Fourier transforms; iterative process; matrix preconditioning; matrix-vector multiplication; microstrip structures; mixed potential integral equation; multifrontal method; preconditioned sparse-matrix; Acceleration; Algorithm design and analysis; Character generation; Computational efficiency; Fast Fourier transforms; Integral equations; Iterative algorithms; Iterative methods; Microstrip; Transmission line matrix methods;

fLanguage

English

Publisher

ieee

Conference_Titel

Microwave Conference Proceedings, 2005. APMC 2005. Asia-Pacific Conference Proceedings

Print_ISBN

0-7803-9433-X

Type

conf

DOI

10.1109/APMC.2005.1606759

Filename

1606759