Adaptive numerical solution of PDEs. (Q437804): Difference between revisions

The present book gives a thorough overview of some of the most important methods in use nowadays to adaptively solve partial differential equations (PDEs). One of the authors, Peter Deuflhard, founded the Zuse Institute in Berlin where from the beginning in the 1980s adaptive algorithms for PDEs were in the focus. Hence, the book under review draws on first hand expert knowledge. The first chapter gives an introduction to some elementary types of PDEs. The Laplace, the Poisson, and the Helmholtz equation are treated in more detail than the wave equation or the Schrödinger equation. Chapter 2 gives an overview of the modeling with PDEs, namely the role of Maxwell's equations in electrodynamics, Euler's and Navier-Stokes's equations in Fluid Dynamics, and the equations of linear and nonlinear elastomechanics. Chapter 3 discusses finite difference discretisations of the Poisson equation while in the fourth chapter Galerkin methods for this type of PDE are introduced. Here, also spectral methods are discussed. In Chapter 5, some numerical algorithms for the solution of linear elliptic grid equations are given, including direct elimination and iterative solvers like multigrid or hierarchical-basic methods. The first real occurrence of adaptive methods is in Chapter 6 where adaptive hierarchical methods are in the focus. The approach is fairly general and includes a thorough discussion of errors and error indicators. A convergence result is given for a model refinement strategy. Adaptive multigrid methods for linear elliptic problems constitute the content of Chapter 7. A combination of finite element analysis, multigrid methods, and adaptive hierarchical grids allows the treatment of the boundary problems in a modern functional analytic way. The cascadic multigrid method is also presented and explained. In Chapter 8, the view is enlarged to allow for nonlinear elliptic problems and it is here where different Newton methods are discussed. In the last Chapter 9, the focus is on parabolic problems, leading to stiff differential equations. An Appendix gives information on Fourier Analysis, Integral Theorems, Sobolev spaces and the delta distribution. Also given are the locations of software packages in the public domain. The book closes with a bibliography containing 235 items and an index. Due to the focus of the authors own work on elliptic and parabolic equations the book does not contain any further material on adaptive methods for hyperbolic problems although much was done in this area in the past. However, the book is a highly valuable addition to the literature. Readers should have more than only a basic knowledge in numerical analysis to appreciate the contents of this book valuable to students, researchers, and practitioners alike.