Multistep methods on manifolds
Bateman equationdifferential algebraic equationdifferential equation on manifolderror analysismultistep methodpreservation of invariantsRunge-Kutta schemeTaylor method
Implicit ordinary differential equations, differential-algebraic equations (34A09) Discretization methods and integrators (symplectic, variational, geometric, etc.) for dynamical systems (37M15) Multistep, Runge-Kutta and extrapolation methods for ordinary differential equations (65L06) Error bounds for numerical methods for ordinary differential equations (65L70) Numerical methods for differential-algebraic equations (65L80) Numerical methods for Hamiltonian systems including symplectic integrators (65P10)
The author discusses the preservation of invariants of an ordinary differential equation in a numerical integration. He considers methods of Taylor type, i.e. methods of order \(p\) where the leading term in the error analysis is given by the \((p+1)\)th derivative of the solution. Many classical numerical schemes are of this type, in particular multistep methods; a notable exception are Runge-Kutta schemes. It is first shown that an invariant can be preserved by a method of Taylor type only, if it satisfies the Bateman equation, then that this implies that the invariant must be linear.
- Preserving algebraic invariants with Runge-Kutta methods
- Manifold topological multi-resolution analysis method
- Multi-parameter surfaces of analytic centers and long-step surface-following interior point methods
- Numerical Methods on (and off) Manifolds
- Multistep methods integrating ordinary differential equations on manifolds
- Time symmetry and high-order Magnus methods
- A double Fourier sphere method for \(d\)-dimensional manifolds
- Functional equivariance and conservation laws in numerical integration
This page was built for publication: Multistep methods on manifolds
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2713135)