Dynamic Domain Decomposition.pdfVIP

2010, Vol.15 No.1, 016-020 Article ID 1007-1202(2010)01-0016-05 DOI 10.1007/s11859-010-0104-1 Dynamic Domain Decomposition Method and Its Application on Nonlinear Problem □ GONG Rongfang1 2 2 0 Introduction , JIANG Ke , SUN Lelin 1. Department of Mathematics, Nanjing University of Aeronautics and Astronautics, Nanjing 210016, Jiangsu, China; It is well known that a normal domain decomposi- 2. School of Mathematics and Statistics, Wuhan University, tion method divides an original physical domain into a Wuhan 430072, Hubei, China certain number, m say, of smaller sub-domains and there- fore a large problem is divided correspondingly into m © Wuhan University and Springer-Verlag Berlin Heidelberg 2010 sub-problems (see Refs. [1-5] and some related refer- Abstract: In this paper, domain decomposition method (DDM) ences therein). Though such a strategy provides a basic for numerical solutions of mathematical physics equations is im- frame for a wide variety of parallel algorithms, it may be proved into dynamic domain decomposition method (DDDM). inconvenient in many applications. We assume that, for The main feature of the DDDM is that the number, shape and instance, there are m divided sub-problems and m proc- volume of the sub-domains are all flexibly changeable during the essors being used to solve them on a one-to-one corre- iterations, so it suits well to be implemented on a reconfigurable