Conjugate-symplecticity properties of Euler--Maclaurin methods and their implementation on the Infinity Computer
For researchers in geometric numerical integration, it provides a theoretical justification for using multi-derivative methods despite non-symplecticity, with a practical implementation strategy.
The paper shows that Euler-Maclaurin methods for Hamiltonian systems are conjugate-symplectic up to order p+2, enabling their use in geometric integration, and proposes efficient derivative computation via the Infinity Computer.
Multi-derivative one-step methods based upon Euler-Maclaurin integration formulae are considered for the solution of canonical Hamiltonian dynamical systems. Despite the negative result that simplecticity may not be attained by any multi-derivative Runge--Kutta methods, we show that the Euler-MacLaurin method of order p is conjugate-symplectic up to order p+2. This feature entitles them to play a role in the context of geometric integration and, to make their implementation competitive with the existing integrators, we explore the possibility of computing the underlying higher order derivatives with the aid of the Infinity Computer.