A characterization of energy-preserving methods and the construction of parallel integrators for Hamiltonian systems
arXiv:1505.0253754 citationsh-index: 40
Originality Incremental advance
AI Analysis
Provides a theoretical framework and practical construction for efficient high-order energy-preserving integration, relevant to computational physics and numerical analysis.
The authors characterize energy-preserving methods for Hamiltonian systems and construct high-order parallel integrators with computational cost comparable to second-order methods.
High order energy-preserving methods for Hamiltonian systems are presented. For this aim, an energy-preserving condition of continuous stage Runge--Kutta methods is proved. Order conditions are simplified and parallelizable conditions are also given. The computational cost of our high order methods is comparable to that of the average vector field method of order two.