论文信息 - CONJUGATE-SYMPLECTICITY OF LINEAR MULTISTEP METHODS

CONJUGATE-SYMPLECTICITY OF LINEAR MULTISTEP METHODS

For the numerical treatment of Hamiltonian differential equations, symplectic integrators are the most suitable choice, and methods that are conjugate to a symplectic integrator share the same good long-time behavior. This note characterizes linear multistep methods whose underlying one-step method is conjugate to a symplectic integrator. The boundedness of parasitic solution components is not addressed.

Ernst Hairer | E. Hairer

[1] Ander Murua,et al. An Algebraic Approach to Invariant Preserving Integators: The Case of Quadratic and Hamiltonian Invariants , 2006, Numerische Mathematik.

[2] E. Hairer,et al. Geometric Numerical Integration: Structure Preserving Algorithms for Ordinary Differential Equations , 2004 .

[3] Ernst Hairer,et al. Symmetric multistep methods over long times , 2004, Numerische Mathematik.

[4] Ernst Hairer,et al. Backward error analysis for multistep methods , 1999, Numerische Mathematik.

[5] Yifa Tang,et al. The symplecticity of multi-step methods , 1993 .

[6] Dai,et al. NON-EXISTENCE OF CONJUGATE-SYMPLECTIC MULTI-STEP METHODS OF ODD ORDER , 2007 .