Stability of finite difference schemes approximation for hyperbolic boundary value problems in an interval - Archive ouverte HAL Access content directly
Journal Articles Mathematics of Computation Year : 2022

Stability of finite difference schemes approximation for hyperbolic boundary value problems in an interval

Abstract

In this article we are interested in the stability of finite difference schemes approximation for hyperbolic boundary value problems defined on the interval [0, 1]. The seminal work of [B. Gustafsson and Sundstrom., 1972], mainly devoted to the half-line, gives a necessary and sufficient invertibility condition ensuring the stability of the scheme, the so-called discrete uniform Kreiss-Lopatinskii condition. An interesting point is that this condition is a discretized version of the one imposed in the continuous setting to ensure the strong well-posedness of the hyperbolic boundary value problem. However as pointed in [B. Gustafsson and Sundstrom., 1972] and as soon as several boundary conditions are concerned the solution of the scheme may develop an exponential growth with respect to the discrete time variable. The question addressed here is to characterize the schemes having this growth or not. This is made under new invertibility conditions which are discretized versions of the ones preventing the exponential growth in time of the solution to continuous hyperbolic boundary value problems in the strip studied in [Benoit, a]. In some sense it shows that this continuous to discrete extension of the characterization occurs in the interval like in the half-line.
Fichier principal
Vignette du fichier
schema_bande1.pdf (531.28 Ko) Télécharger le fichier
Origin : Files produced by the author(s)
Loading...

Dates and versions

hal-02887577 , version 1 (02-07-2020)

Identifiers

Cite

Antoine Benoit. Stability of finite difference schemes approximation for hyperbolic boundary value problems in an interval. Mathematics of Computation, 2022, 91, ⟨10.1090/mcom/3698⟩. ⟨hal-02887577⟩
31 View
87 Download

Altmetric

Share

Gmail Facebook X LinkedIn More