MOUGIN, Mr Thibault (2023) Filtering and Symplectical Time-Integration PRE - Research Project, ENSTA.



Hamiltonian systems are dynamical systems used to model many physical systems, and that require to employ numerical integration methods providing energy conserving solutions : symplectic schemes. A recent development in convolutional filters for numerical methods has shown interesting results in increasing the smoothness of solutions of discontinuous Galerkin methods and this idea may have a promising application in this Hamiltonian systems context. This is what we try to experiment with in this study. We encounter convergence improvements for specific filters and symplectic methods. Computation times can be improved and an optimization problem is raised to further improve convergence of schemes.

Item Type:Thesis (PRE - Research Project)
Uncontrolled Keywords:Hamiltonian systems, symplectic integrators, post processing, convolutional filters, accuracy extraction
Subjects:Mathematics and Applications
ID Code:9566
Deposited By:Thibault MOUGIN
Deposited On:25 août 2023 14:08
Dernière modification:25 août 2023 14:08

Repository Staff Only: item control page