Iterated discretization

DOI10.1080/01630569608816719MaRDI QIDQ4354827zbMATH OpenOpenAlexFDO

Authors Rainer Ansorge

Publication date 22 January 1998

Published in Numerical Functional Analysis and Optimization (Search for Journal in Brave)

Full work available at URL https://doi.org/10.1080/01630569608816719

zbMATH Keywords

convergence analysisfixed point equationiterated discretization methodmetric spacesnormed spacesprojection methodssuperconvergence

Mathematics Subject Classification ID

Nonlinear ordinary differential equations and systems (34A34) Iterative procedures involving nonlinear operators (47J25) Numerical solutions to equations with nonlinear operators (65J15) Numerical methods for initial value problems involving ordinary differential equations (65L05)

Review text

With the paper in hand, the idea of a general iterated discretization is presented, and the question of superconvergence is dealt with, too. The idea is based on the recent development of projection methods for the numerical approximation of integral equations by means of an additional iteration step. These iterated projection methods could be proven to be superconvergent under suitable conditions [cf. \textit{I.-H. Sloan} and \textit{V. Thomée}, J. Integral Equations 9, No. 1, 1-23 (1985; Zbl 0575.65131)]. The author provides a general framework in which this idea is extended to more general discretizations (not only projection methods), and considers firstly the discretization of the fixed point equation \(x=Kx\), \(K:X\to X\), on metric spaces \(X\), \(X_n\) \((n=1,2,\dots)\). The approximate spaces \(X_n\) might not be necessarily subspaces of \(X\), the operator \(K\) is quite general. The usual approximation scheme (representing the relations between the original and discretized problem) is generalized, and the concept of convergence is settled. Secondly, for normed spaces \(X\), \(X_n\) with linear generalized restriction and prolongation operators, and under feasible assumptions on the operator \(K\), superconvergence is analysed. The framework then is applied to the autonomous initial value problem \(\dot x(t)= f(x(t))\), \(0<t\leq 1\), \(x(0)=1\), in the space of piecewise continuous functions on \([0,1]\), equipped with the usual \(L_1\)-norm. Here, the function \(f:\mathbb{R}\to\mathbb{R}\) is assumed to be continuously differentiable.

Recommendations

• Superconvergence of the iterated degenerate kernel method
• Projection and Iterated Projection Methods for Nonlinear Integral equations
• The iterated Galerkin method for linear integro-differential equations
• Superconvergence results of iterated projection methods for linear Volterra integral equations of second kind
• Superconvergence of iterated Galerkin method for a class of nonlinear Fredholm integral equations

Cites work

• A unified framework for the discretization of nonlinear operator equations
• Projection and Iterated Projection Methods for Nonlinear Integral equations

Cited in

(1)

• Growth of Discrete Projection Ghosts Created by Iteration

This page was built for publication: Iterated discretization

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4354827)