External approximation of eigenvalue problems in Banach spaces
Eigenvalue approximation
Superconvergence effect in finite element methods for two-point boundary value problems
Approximate methods for solving two-point boundary value problems are considered. The aim of the paper is to explain superconvergence effect in the methods using finite element spaces. The existence of a class of the methods with the superconvergence property is demonstrated. Detailed proofs of superconvergence are presented for the case of the Galerkin method (due to Douglas and Dupont results) and for some example of external method.
Approximate solving of ill posed problems
The paper has a form of review article. The aim of the paper is to explain the meaning of solution of ill posed problem. Moreover, it is shown that by using an appropriate method the ill posed problem can be reasonably solved also in the case of incexact data. The regularization method and, especially, certin methods of constructing regularization operators are discussed. The method of approximate solving of ill posed problems is illustrated by an example of the 1st kind Fredholm equation.
Review: J. Stoer; Introdution to numeric methods, vol. I; J. Stoer, R. Bulirsch; Introdution to numeric methods, vol. II
The article contains no abstract
On stability of outer inverses
The paper is concerned with stability properties of bounded outer inverses for bounded operators on Banach spaces. We investigate the effect of perturbations of subspaces generating outer inverses for a given operator. We prove that the convergence of respective gaps between subspaces implies the convergence in the norm of outer inverses generated by these subspaces. The notion of outer inverses is illustrated by an example which gives the form of outer inverses for an arbitrary matrix.
Approximation of eigenvalues of nonselfadjoint problems on an infinite interval
The author treats the problem of approximation of the eigenvalues of the nonselfadjoint problem Lu=λu, u(0)=0, u∈L2(0,∞), where L=−d2/dt2+p(x), domL={u∈L2(0,∞):du/dt continuous, d2u/dt2∈L2(0,∞), u(0)=0}. The same question for a selfadjoint problem was answered in the author's recent paper.
Function representation in wavelet base
The paper concerns wavelet bases in the space L2(M). We give a review of elements of the wavelet theory selected from the point of view of numerical analysis applications. Our purpose is to show properties of a function representation in a wavelet basis, among others, to formulate criteria of global and local function smoothness in terms of its wavelet decomposition coefficients.
Approximation of eigenvalues for a class of differential problems on an infinite interval
The author investigates the problem (1) Lu=λu, u(0)=0, u∈L2(0,∞), where L=d2/dt2+a(t)d/dt+b(t) and domL={u∈L2(0,∞):du/dt absolutely continuous, d2u/dt2∈L2(0,∞), u(0)=0}. She proves that under certain conditions it is possible to approximate the spectrum of (1) by means of the spectrum of a suitable problem of the form −u′′+c(t)u=λu, u(0)=0, u′(n)=α(n)u(n).The review of the paper is available at MR0518666.
Approximate methods for solving differential equations on infinite interval
Convergence of approximation methods for eigenvalue problem for two forms
The paper concerns an approximation of an eigenvalue problem for two forms on a Hilbert space . We investigate some approximation methods generated by sequences of forms and defined on a dense subspace of . The proof of convergence of the methods is based on the theory of the external approximation of eigenvalue problems. The general results are applied to Aronszajn’s method.
Superconvergence of external approximation for two-point boundary problems
The superconvergence property of a certain external method for solving two point boundary value problems is established. In the case when piecewise polynomial spaces are applied, it is proved that the convergence rate of the approximate solution at the knot points can exceed the global one.
Error estimates for external approximation of ordinary differential equations and the superconvergence property
A pointwise error estimate and an estimate in norm are obtained for a class of external methods approximating boundary value problems. Dependence of a superconvergence phenomenon on the external approximation method is studied. In this general framework, superconvergence at the knot points for piecewise polynomial external methods is established.
On a certain algorithm of eigenvalue localization for normal operators
Page 1