This study presents two domain decomposition formulations combined with the Preconditioned Conjugate Gradient (PCG) method for solving large-scale linear structural problems on workstation clusters. The first approach, which is a subdomain-by-subdomain PCG algorithm, operates on the global stiffness matrix performing the dominant matrix-vector operations locally, in the subdomain level. A diagonal preconditioner is implemented, as well as an additive polynomial preconditioner expressed by a truncated Neumann series. In the second approach a dual domain decomposition formulation is implemented as proposed by Farhat and Roux (FETI method); the domain is partitioned into a set of totally disconnected subdomains using Lagrange multipliers and the resulting interface problem is handled by a preconditioned conjugate projected gradient algorithm. The two domain decomposition formulations are applied using the master/slave model for distributed computing and numerical tests are performed on ethernet-networked workstations running the message passing software PVM.

purchase the full-text of this paper (price £20)

go to the previous paper
go to the next paper
return to the table of contents
return to the book description
purchase this book (price £66 +P&P)