Variable time step integration with symplectic methods (Q1372703): Difference between revisions
From MaRDI portal
Changed an Item |
Set profile property. |
||
Property / MaRDI profile type | |||
Property / MaRDI profile type: MaRDI publication profile / rank | |||
Normal rank |
Revision as of 03:07, 5 March 2024
scientific article
Language | Label | Description | Also known as |
---|---|---|---|
English | Variable time step integration with symplectic methods |
scientific article |
Statements
Variable time step integration with symplectic methods (English)
0 references
12 March 1998
0 references
It is a fundamental problem in the symplectic integration of Hamiltonian systems that the favorable properties of such methods get lost, if one uses variable stepsizes. The author proposes to perform a time reparametrization in the Hamiltonian system instead of changing the stepsize during the numerical integration. A naive transformation of the time variable will not preserve the Hamiltonian character of the system. But the author shows how one can define for given initial data a new Hamiltonian with a transformed time variable such that the solutions of the two Hamiltonian systems for these special initial data coincide. A backward error analysis of the approach is given. The author discusses several concrete choices of time reparametrization and demonstrates their performances in numerical experiments with the Kepler problem.
0 references
symplectic integration
0 references
Hamiltonian system
0 references
variable stepsize
0 references
backward error analysis
0 references
time reparametrization
0 references
performance
0 references
numerical experiment
0 references
Kepler problem
0 references