On a Parallel Nonoverlapping Domain Decomposition Method

Xuejun Xu · Institue of Computational Mathematcis, Chinese; Academy of Sciences · 11/04/05

On a Parallel Nonoverlapping Domain Decomposition Method Xuejun Xu

Converted from past_seminars/cam_colloquium/abstracts/XX1104.pdf (PDF). Mathematical notation is approximate — it comes from text extraction, not from the typeset source.

LSEC, Institute of Computational Mathematics, Chinese Academy of Sciences, Beijing, China

Abstract In recent years, a nonoverlapping domain decomposition iterative procedure, which is based on using Robin-type boundary conditions as information transmission conditions on the subdomain interfaces, has been developed and analyzed. The idea of employing Robin-type boundary conditions as interface conditions was first proposed by P.L. Lions. It is known that the convergence rate of this method is 1 − O(h), where h is mesh size. In this talk, the convergence rate will be improved to 1 − O(h1/2 H −1/2 ) by choosing suitable parameter on the interfaces, where H is the subdomain size. Counter examples will be constructed to show that the convergence rate is sharp.

1