No label defined (Q3765140)
From MaRDI portal
!
This is the item page for this Wikibase entity, intended for internal use and editing purposes. Please use the normal view instead:
scientific article; zbMATH DE number 4022513
| Language | Label | Description | Also known as |
|---|---|---|---|
| default for all languages | No label defined |
||
| English | No label defined |
scientific article; zbMATH DE number 4022513 |
Statements
1984
0 references
singularly perturbed boundary value problems
0 references
multiple shooting
0 references
decoupling
0 references
power series expansions
0 references
small parameter
0 references
truncation error
0 references
pathfinder
0 references
0.8394407629966736
0 references
0.8275412321090698
0 references
0.7963708639144897
0 references
[For the entire collection see Zbl 0626.00027.] It has become almost folklore that numerical methods based on initial- value integration techniques cannot work well for singularly perturbed boundary value problems, or at least that they cannot be efficient. We show that this is not quite correct, at least if such a method - like multiple shooting - is part of a sensible global strategy. We consider two types of approach. In the first one we utilize some kind of decoupling of the dynamics, such that the slow mode part of the solution can be computed separately. Often this can be done quite conveniently by employing analytical methods, like power series expansions in the small parameter \(\epsilon\). The resulting separated smooth system is solved by repeated superposition, i.e. multiple shooting. Clearly the accuracy obtained is the ``worst'' of the truncation error of the expansion and the tolerance set for the numerical integrator needed for the latter algorithm. A different approach is based on the use of a so called ``pathfinder'', which defines the grid for the integration of the variational equations. The crucial idea here is to use special integrations for this ``pathfinder'', which, together with a monitoring of the Wronskians makes it possible to obtain relaxed stepsizes away from the layers. The main emphasis will be on the first method, which is the most complete one.
0 references