Domain-Decomposition Approach to Krylov Subspace Iteration
Journal article, 2016

Krylov subspace iterative techniques consist of finding the solution of a scattering problem as a linear combination of "generating vectors" obtained through successive matrix-vector multiplications. This letter extends this approach to domain-decomposition. Here, on each subdomain, a subspace is obtained by constructing the segments of each generating vector associated with the subdomain and by weighting these segments independently, which provides more degrees of freedom. The method is tested for scattering by a sphere and a rectangular plate, as well as radiation from connected arrays with strongly coupled antenna elements. It is shown that substantial computational savings can be obtained for the sphere and the array. This opens up new perspectives for faster solutions of multiscaled problems.

antennas

Generalized Minimal Residual Method (GMRES)

Characteristic Basis Functions Method (CBFM)

Telecommunications

Krylov subspace iteration

Engineering

connected arrays

domain-decomposition

linear-systems

Author

Oleg Iupikov

Chalmers, Signals and Systems, Communication, Antennas and Optical Networks

C. Craeye

Universite catholique de Louvain

Rob Maaskant

Chalmers, Signals and Systems, Communication, Antennas and Optical Networks

Marianna Ivashina

Chalmers, Signals and Systems, Communication, Antennas and Optical Networks

IEEE Antennas and Wireless Propagation Letters

1536-1225 (ISSN) 15485757 (eISSN)

Vol. 15 1414-1417 7362146

Areas of Advance

Information and Communication Technology

Subject Categories

Telecommunications

DOI

10.1109/lawp.2015.2511195

More information

Latest update

3/8/2018 1