Monotone iterative methods for boundary value problems

Authors M. Cherpion, C. De Coster, Patrick Habets

Publication date 26 January 2003

Published in Differential and Integral Equations (Search for Journal in Brave)

Dirichlet conditions lower solution monotone iterative method phi-Laplacian problem second order upper solution

Theoretical approximation of solutions to ordinary differential equations (34A45) Nonlinear boundary value problems for ordinary differential equations (34B15) Numerical solution of boundary value problems involving ordinary differential equations (65L10)

Review text

The authors study the boundary value problem \[ -(\varphi \circ u')'=f(t,u,u'), \quad u(a)=u(b)=0, \tag{1} \] where \(-\infty <a<b<\infty ,\) \(f\) is a Carathéodory function and \(\varphi :\mathbb{R}\mapsto \mathbb{R}\) is an increasing homeomorphism with \(\varphi (0)=0.\) Furthermore, it is assumed that (1) admits a well ordered pair of upper and lower functions \(\alpha ,\beta \) and there are \(C>0,\) \(k\in {\roman C}([0,\infty),(0,\infty)),\) \(\psi \in L^p[a,b],\) \(p\in 0,\infty ]\) such that \(f\) satisfies a Nagumo-type condition and a certain additional assumption (\(C\) is ``admissible for \(\varphi ,\) \((\alpha ,\beta)\) and \((k,\psi)\)). First, the authors show that under these assumptions problem (1) has a maximal and a minimal solution in \([\alpha ,\beta ]\) and the estimate \(\|u'\|_{\infty }<C\) holds for any solution \(u\) to (1) such that \(\alpha (t)\leq u(t)\leq \beta (t)\) on \([a,b]\) (i.e. \(u\in [\alpha ,\beta ]\)). Then they consider two approximation schemes for (1) introduced by E. Picard. It is worth mentioning that the paper contains also a detailed historical survey of the subject.

Recommendations

Cited in

(35)

This page was built for publication: Monotone iterative methods for boundary value problems

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q1854074)