A domain decomposition method for 3d elasticity problems:

Abstract: "The symmetric-and-antisymmetric (SAS) domain decomposition method is a special matrix decomposition scheme for certain classes of discretized physical problems. This scheme decomposes the original problem into independent or loosely coupled subproblems by taking advantage of symmetry...

Full description

Saved in:
Bibliographic Details
Main Authors: Chen, Hsin-Chu (Author), Sameh, Ahmed (Author)
Format: Book
Language:English
Published: Urbana, Ill. 1989
Series:Center for Supercomputing Research and Development <Urbana, Ill.>: CSRD report 890
Subjects:
Summary:Abstract: "The symmetric-and-antisymmetric (SAS) domain decomposition method is a special matrix decomposition scheme for certain classes of discretized physical problems. This scheme decomposes the original problem into independent or loosely coupled subproblems by taking advantage of symmetry or partial symmetry. This is achieved by exploiting important properties possessed by two special classes of matrices A and B, A and B [are elements of] C[superscript nxn], which satisfy the relations A = PAP and B = - PBP where P is some symmetric signed permutation matrix
In this paper we present the application of the SAS decomposition method to 3D isotropic elasticity problems using the (preconditioned) conjugate gradient algorithm. The method is first applied to problems with symmetric domains and boundary conditions. We then extend its application to general asymmetric problems by choosing the preconditioner to be the stiffness matrix of some symmetrically discretized substructure. This method is not only superior to classical schemes on sequential computers, but also lends itself to efficient implementation on multiprocessors with several levels of parallelism.
Physical Description:18 S.

There is no print copy available.

Interlibrary loan Place Request Caution: Not in THWS collection!