Bücher von Wolfgang Hackbusch

Special numerical techniques are already needed to deal with n x n matrices for large n. Tensor data are of size n x n x...x n=nd, where nd exceeds the computer memory by far. Since standard methods fail, a particular tensor calculus is needed to treat such problems.

This is the soft cover reprint of the very popular hardcover edition. The book offers a simultaneous presentation of the theory and of the numerical treatment of elliptic problems.

This book describes the methods by which tensors can be practically treated and shows how numerical operations can be performed. Applications include problems from quantum chemistry, approximation of multivariate functions, solution of pde and more.

However, the technique of hierarchical matrices makes it possible to store matrices and to perform matrix operations approximately with almost linear cost and a controllable degree of approximation error.

The GAMM Committee for Efficient Numerical Methods for Par­ tial Differential Equations (GAMM-FachausschuB "Effiziente numerische Verfahren fUr partielle Differenzialgleichungen") organizes conferences and seminars on subjects concerning the algorithmic treatment of partial differential equation prob­ lems. The first seminar "Efficient Solution of Elliptic Systems" was followed by a second one held at the University of Kiel from January 17th to January 19th, 1986. The title was "Efficient Numerical Methods in Continuum Mechanics". The equations arising in continuum mechanics have many con­ nections to those of fluid mechanics, but are usually more complex. Therefore, much attention has to be paid to the ef­ ficient discretization, postprocessing and extrapolation. The seminar was attended by 66 scientists from 10 countries. Most of the 21 lectures presented at the seminar treated the discretization of equations in continuum mechanics by finite elements, methods for improving the accuracy of these approx­ imations and the use of boundary elements. Other contribu­ tions presented efficient methods for investigating bifurca­ tions which play an essential role in practical applications. These proceedings contain 11 contributions in alphabetical order. The editors and organizers of the seminar would like to thank the land Schleswig-Holstein and the DFG (Deutsche Forschungs­ gemeinschaft) for their support. Kiel, November 1986 W. Hackbusch K. Witsch v Contents Page K. ERIKSSON, C. JOHNSON, J. LENNBLAD: Optimal error estimates and adaptive time and space step control for linear parabolic problems ....................... .

This book offers a self-contained presentation of aspects of stability in numerical mathematics. It compares and characterizes stability in different subfields of numerical mathematics.

In full multigrid methods for elliptic difference equations one works on a sequence of meshes where a number of pre- and/or postsmoothing steps are performed on each level. As is well known these methods can converge very fast on problems with a smooth solution and a regular mesh, but the rate of convergence can be severely degraded for problems with unisotropy or discontinuous coefficients unless some form of robust smoother is used. Also problems can arise with the increasingly coarser meshes because for some types of discretization methods, coercivity may be lost on coarse meshes and on massively parallel computers the computation cost of transporting information between computer processors devoted to work on various levels of the mesh can dominate the whole computing time. For discussions about some of these problems, see (11). Here we propose a method that uses only two levels of meshes, the fine and the coarse level, respec­ tively, and where the corrector on the coarse level is equal to a new type of preconditioner which uses an algebraic substructuring of the stiffness matrix. It is based on the block matrix tridiagonal structure one gets when the domain is subdivided into strips. This block-tridiagonal form is used to compute an approximate factorization whereby the Schur complements which arise in the recursive factorization are approximated in an indirect way, i. e.

This volume contains 23 contributions to the 10th GAMM-Seminar, which was held in Kiel on the 14-16 of January 1994. The central topics are advanced numerical techniques for solving flow problems. Five papers are devoted to parallel algorithms, a further one to domain decomposition techniques.

11 The GAMM Committee for Efficient Numerical Methods for Partial 11 Differential Equations organises workshops on subjects concerning the algorithmic treatment of partial differential equations. The topics are discretisation methods like the finite element and the boundary element method for various types of applications in structural and fluid mechanics. Particular attention is devoted to the advanced solution methods. The series of such workshops was continued in 1991, January 25- 27, with the 7th Kiel-Seminar on the special topic 11 11 Numerical techniques for boundary element methods at the Christian-Albrechts-University of Kiel. The seminar was attended by 57 scientists from 8 countries. The list of topics contained applications of the boundary element method (BEM) to various problems of practical interest, algo­ rithmic aspects of the BEM (coupling with finite element method, parallelisation), convergence analysis, and in particular the treatment of the numerical integration. In six contributions the quadrature of weakly singular, Cauchy singular, and hypersingular integrals is analysed. 11 11 The editor thanks the DFG-Schwerpunkt Randelementmethoden for its support. He also likes to express his gratitude to all persons involved in the organisation of the seminar.

The most frequently used method for the numerical integration of parabolic differential equa­ tions is the method of lines, where one first uses a discretization of space derivatives by finite differences or finite elements and then uses some time-stepping method for the the solution of resulting system of ordinary differential equations. Such methods are, at least conceptually, easy to perform. However, they can be expensive if steep gradients occur in the solution, stability must be controlled, and the global error control can be troublesome. This paper considers a simultaneaus discretization of space and time variables for a one-dimensional parabolic equation on a relatively long time interval, called 'time-slab'. The discretization is repeated or adjusted for following 'time-slabs' using continuous finite element approximations. In such a method we utilize the efficiency of finite elements by choosing a finite element mesh in the time-space domain where the finite element mesh has been adjusted to steep gradients of the solution both with respect to the space and the time variables. In this way we solve all the difficulties with the classical approach since stability, discretization error estimates and global error control are automatically satisfied. Such a method has been discussed previously in [3] and [4]. The related boundary value techniques or global time integration for systems of ordinary differential equations have been discussed in several papers, see [12] and the references quoted therein.

Bei der Diskretisierung von Randwertaufgaben und Integralgleichungen entstehen große, eventuell auch voll besetzte Matrizen. Es wird eine neuartige Methode dargestellt, die es erstmals erlaubt, derartige Matrizen nicht nur effizient zu speichern, sondern auch alle Matrixoperationen einschließlich der Matrixinversion bzw. der Dreieckszerlegung approximative durchzuführen. Anwendungen findet diese Technik nicht nur bei der Lösung großer Gleichungssysteme, sondern auch bei Matrixgleichungen und der Berechnung von Matrixfunktionen.

The GAMM Committee for "Efficient Numerical Methods for Partial Differential Equations" organizes seminars and workshops on subjects concerning the algorithmic treatment of partial differential equations. The topics are discretisation methods like the finite element and the boundary element method for various type of applications in structural and fluid mechanics. Particular attention is devoted to the advanced solution methods. The series of such seminars was continued in 1995, January 20-22, with the 11th Kiel-Seminar on the special topic Numerical Treatment of Coupled Systems at the Christian-Albrechts-University of Kiel. The seminar was attended by 100 scientist from 9 countries. 23 lectures were given, including two survey lectures. Different kinds of couplings are considered in this volume. The coupling of different components may occur in the physical model. On the other hand, a coupling of subsystems can be generated by the numerical solution technique. General examples of the latter kind are the domain decomposition (see p. 128) or subspace decomposition (p. 117). The local defect correction method couples different discretizations of the same problem in order to improve the results, although the basic linear system to be solved remains unchanged (p. 47). In general, the aim of the numerical coupling is to make use of (efficient) subsystem solvers (p. 1). The combination of different discretization techniques is mentioned on page 59.

The theory of integral equations has been an active research field for many years and is based on analysis, function theory, and functional analysis.

Presents the description of the state of the modern iterative techniques together with systematic analysis. This book discusses classical methods, semi-iterative techniques, incomplete decompositions, conjugate gradient methods, multigrid methods and domain decomposition techniques.

The theory of integral equations has been an active research field for many years and is based on analysis, function theory, and functional analysis.

Multi-grid methods are the most efficient tools for solving elliptic boundary value problems. One section describes special applications (convection-diffusion equations, singular perturbation problems, eigenvalue problems, etc.).