Behavior of alternation points in best rational approximation

The behavior of the equioscillation points (alternants) for the error in best uniform approximation on [–1, 1] by rational functions of degreen is investigated. In general, the points of the alternants need not be dense in [–1, 1], even when approximation by rational functions of degree (m, n) is considered and asymptoticallym/n 1. We show, however, that if more thanO(logn) poles of the approximants stay at a positive distance from [–1, 1], then asymptotic denseness holds, at least for a subsequence. Furthermore, we obtain stronger distribution results when n (0 < 1) poles stay away from [–1, 1]. In the special case when a Markoff function is approximated, the distribution of the equioscillation points is related to the asymptotics for the degree of approximation.The research of this author was supported, in part, by NSF grant DMS 920-3659.     

On the numerical treatment of a small divisor problem

Summary Quasiperiodic solutions of perturbed integrable Hamiltonian equations such as weakly coupled harmonic oscillators can be found by constructing an appropriate coordinate transformation which leads to a small divisor problem. However the numerical difficulties are not merely caused by the small divisors but rather by the appearence of ghost solutions, which appear in any reasonable discretization of the problem. Our numerical treatment, based on a Newton-type iteration, guarantees an approximation of the relevant solution of the nonlinear problem. Numerical solutions are found up to a critical value of the coupling constant, which is much larger than the coupling constants allowed by the existence theory available so far.     

The contraction number of a multigrid method for solving the Poisson equation

Summary The treatment of a multigrid method in the framework of numerical analysis elucidates that regularity of the solution is not necessary for the convergence of the multigrid algorithm but only for fast convergence. For the linear equations which arise from the discretization of the Poisson equation, a convergence factor 0,5 is established independent of the shape of the domain and of the regularity of the solution.Dedicated to Professor Dr.Dr.h.c. Lothar Collatz on the occasion of his 70 th birthday     

On the uniqueness of generalized monosplines of least Lp-norm

The uniqueness of monosplines and perfect splines of leastL _p-norm is treated in the framework of generalized monosplines and total positivity. The analysis is based on the invariance properties of the degree of a certain mapping and on a new composition result for totally positive kernels. For theL _p-case 1<p<∞, uniqueness is shown under the same extra conditions as were previously shown to be needed in theL _p-case. The uniqueness in theL _∞-case is obtained without any restrictions.     

A posteriori error estimators for obstacle problems – another look

We show that a posteriori estimators for the obstacle problem are easily obtained from the theory for linear equations. The theory would be even simpler if the Lagrange multiplier does not have a nonconforming contribution as it has in actual finite element computations.     

Über die Einzugsbereiche der Nullstellen von Polynomen beim Newton-Verfahren

Summary The region of attraction of a zero of a polynomial consists of those points from which the Newton method may be started when this zero is to be computed. The regions of attraction are approximately characterized by studying a continuous analog of the Newton method.

     

A finite element method for nearly incompressible elasticity problems

A finite element method is considered for dealing with nearly incompressible material. In the case of large deformations the nonlinear character of the volumetric contribution has to be taken into account. The proposed mixed method avoids volumetric locking also in this case and is robust for (with being the well-known Lamé constant). Error estimates for the -norm are crucial in the control of the nonlinear terms.

     

On the number of best approximations in certain non-linear families of functions

Ohne Zusammenfassung     

Über den Zusammenhang von Interpolation und diskreterT-Approximation mit rationalen Funktionen

Ohne Zusammenfassung     

Bernstein polynomials and learning theory

When learning processes depend on samples but not on the order of the information in the sample, then the Bernoulli distribution is relevant and Bernstein polynomials enter into the analysis. We derive estimates of the approximation of the entropy function x log x that are sharper than the bounds from Voronovskaja's theorem. In this way we get the correct asymptotics for the Kullback–Leibler distance for an encoding problem.