2019
DOI: 10.1090/mcom/3494
|View full text |Cite
|
Sign up to set email alerts
|

Stability of Semi-Lagrangian schemes of arbitrary odd degree under constant and variable advection speed

Abstract: The equivalence between semi-Lagrangian and Lagrange-Galerkin schemes has been proved in [9, 10] for the case of centered Lagrange interpolation of odd degree p ≤ 13. We generalize this result to an arbitrary odd degree, for both the case of constant-and variable-coefficient equations. In addition, we prove that the same holds for spline interpolations.

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
4
1

Citation Types

0
7
0

Year Published

2020
2020
2024
2024

Publication Types

Select...
5
2

Relationship

3
4

Authors

Journals

citations
Cited by 8 publications
(7 citation statements)
references
References 13 publications
0
7
0
Order By: Relevance
“…Proof Following [21,22], we sketch the arguments to prove (34) for the cases of symmetric Lagrange and splines interpolation. In these cases, the method can be interpreted as Lagrange--Galerkin schemes with area-weighting.…”
Section: Stabilitymentioning
confidence: 99%
See 1 more Smart Citation
“…Proof Following [21,22], we sketch the arguments to prove (34) for the cases of symmetric Lagrange and splines interpolation. In these cases, the method can be interpreted as Lagrange--Galerkin schemes with area-weighting.…”
Section: Stabilitymentioning
confidence: 99%
“…For a formal definition of the basis functions ψ j in the case of symmetric Lagrange and spline interpolation, we refer the reader to [21,22]. While these two cases allow for a complete theory, at least in one space dimension, in the numerical tests with unstructured grids we will also use P 2 interpolants, for which a first attempt of stability analysis in presented in [23].…”
Section: Stabilitymentioning
confidence: 99%
“…Note also that, in SL schemes (see, e.g., the discussion of this point in [7]), it is usually required for stability reasons that characteristics passing through neighbouring nodes do not cross. In practice, X ∆ (x i , t n+1 ; t n ) and X ∆ (x k , t n+1 ; t n ) must always have a positive distance, so that, using (4) and the reverse triangular inequality,…”
Section: Element Search By Barycentric Walkmentioning
confidence: 99%
“…The following proposition implies stability for the linear part of the scheme with respect to the 2-norm. For a formal definition of the basis functions ψ j we refer the reader to [17], [18].…”
Section: Stabilitymentioning
confidence: 99%
“…We will therefore focus on the case of high-order interpolations, for which the norm in (30) should be understood as 2-norm. Following [17], [18], we sketch the arguments to prove (30) for the cases of symmetric Lagrange and splines interpolation, which can be interpreted as Lagrange-Galerkin schemes with area-weighting. First, we make explicit the dependence of the points z k on x and ∆t.…”
Section: Stabilitymentioning
confidence: 99%