---

＜電子ブック＞
A Parallel Multilevel Partition of Unity Method for Elliptic Partial Differential Equations / by Marc Alexander Schweitzer
(Lecture Notes in Computational Science and Engineering. ISSN:21977100 ; 29)

版	1st ed. 2003.
出版者	(Berlin, Heidelberg : Springer Berlin Heidelberg : Imprint: Springer)
出版年	2003
大きさ	VI, 200 p. 15 illus : online resource
著者標目	*Schweitzer, Marc Alexander author SpringerLink (Online service)
件　名	LCSH:Mathematical analysis LCSH:Mathematics—Data processing LCSH:Mathematical physics LCSH:Differential equations LCSH:Engineering mathematics LCSH:Engineering—Data processing FREE:Analysis FREE:Computational Mathematics and Numerical Analysis FREE:Theoretical, Mathematical and Computational Physics FREE:Differential Equations FREE:Mathematical and Computational Engineering Applications
一般注記	1 Introduction -- 2 Partition of Unity Method -- 2.1 Construction of a Partition of Unity Space -- 2.2 Properties -- 2.3 Basic Convergence Theory -- 3 Treatment of Elliptic Equations -- 3.1 Galerkin Discretization -- 3.2 Boundary Conditions -- 3.3 Numerical Results -- 4 Multilevel Solution of the Resulting Linear System -- 4.1 Multilevel Iterative Solvers -- 4.2 Multilevel Partition of Unity Method -- 4.3 Numerical Results -- 5 Tree Partition of Unity Method -- 5.1 Single Level Cover Construction -- 5.2 Construction of a Sequence of PUM Spaces -- 5.3 Numerical Results -- 6 Parallelization and Implementational Details -- 6.1 Parallel Data Structures -- 6.2 Parallel Tree Partition of Unity Method -- 6.3 Numerical Results -- 7 Concluding Remarks -- Treatment of other Types of Equations -- A.1 Parabolic Equations -- A.2 Hyperbolic Equations -- Transformation of Keys -- Color Plates -- References The numerical treatment of partial differential equations with meshfree discretization techniques has been a very active research area in recent years. Up to now, however, meshfree methods have been in an early experimental stage and were not competitive due to the lack of efficient iterative solvers and numerical quadrature. This volume now presents an efficient parallel implementation of a meshfree method, namely the partition of unity method (PUM). A general numerical integration scheme is presented for the efficient assembly of the stiffness matrix as well as an optimal multilevel solver for the arising linear system. Furthermore, detailed information on the parallel implementation of the method on distributed memory computers is provided and numerical results are presented in two and three space dimensions with linear, higher order and augmented approximation spaces with up to 42 million degrees of freedom HTTP:URL=https://doi.org/10.1007/978-3-642-59325-3

 類似資料