Treffer: Multiscale Bases for the Sparse Representation of Boundary Integral Operators on Complex Geometry: Multiscale bases for the sparse representation of boundary integral operators on complex geometry
Title:
Multiscale Bases for the Sparse Representation of Boundary Integral Operators on Complex Geometry: Multiscale bases for the sparse representation of boundary integral operators on complex geometry
Authors:
Source:
SIAM Journal on Scientific Computing. 24:1610-1629
Publisher Information:
Society for Industrial & Applied Mathematics (SIAM), 2003.
Publication Year:
2003
Subject Terms:
numerical examples, Multigrid methods, domain decomposition for boundary value problems involving PDEs, Numerical solutions to overdetermined systems, pseudoinverses, Laplace operator, Helmholtz equation (reduced wave equation), Poisson equation, sparse matrices, fast multipole method, Boundary element methods for boundary value problems involving PDEs, wavelets, hierarchical basis preconditioner, 01 natural sciences, boundary element method, potential theory, Complexity and performance of numerical algorithms, multiscale methods, Integral representations, integral operators, integral equations methods in two dimensions, 0101 mathematics
Document Type:
Fachzeitschrift
Article
File Description:
application/xml
Language:
English
ISSN:
1095-7197
1064-8275
1064-8275
DOI:
10.1137/s1064827500369451
Access URL:
Accession Number:
edsair.doi.dedup.....42169d4d16f82d9bcdcd8ff4f284d57e
Database:
OpenAIRE
Weitere Informationen
A multilevel transform is introduced to represent discretizations of integral operators from potential theory by nearly sparse matrices. The new feature is to construct the basis in a hierarchical decomposition of the three-space and not, in a parameter space of the boundary manifold. This construction leads to sparse representations of the operator even for geometrically complicated domains. It is demonstrated that the numerical costs are essentially equal to performing the fast multipole method. The diagonal blocks of the transformed matrix can be used as an inexpensive preconditioner which is empirically shown to reduce the condition number independent of the mesh size.