Premium
A multiblock generalized forward‐backward method
Author(s) -
Pino M. R.,
Obelleiro F.,
Rodríguez J. L.,
Burkholder R. J.
Publication year - 2001
Publication title -
radio science
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.371
H-Index - 84
eISSN - 1944-799X
pISSN - 0048-6604
DOI - 10.1029/1999rs002292
Subject(s) - generalization , position (finance) , reduction (mathematics) , convergence (economics) , algorithm , mathematics , computer science , relaxation (psychology) , matrix (chemical analysis) , matrix decomposition , set (abstract data type) , factorization , mathematical optimization , mathematical analysis , eigenvalues and eigenvectors , geometry , physics , psychology , social psychology , materials science , finance , quantum mechanics , economics , composite material , programming language , economic growth
In a previous work, the generalized forward‐backward (GFB) method has been proposed to obtain the scattering from targets on rough ocean‐like surfaces. In this paper a generalization of the GFB method is presented which allows larger targets and multiple targets to be handled efficiently. The solution is based on a multiblock resolution of each target which is divided into a set of small subregions (blocks). The solution is obtained via a standard method of moments matrix factorization for the small blocks, combined with a conventional forward‐backward iterative procedure to account both for the interactions between blocks and for the regions without obstacles. A relaxation parameter is also introduced to improve convergence. The proposed method provides a general formulation completely independent of the number, the size and the position of the obstacles over the sea surface and allows a significant reduction in the computational and storage costs.