Speaker：Chuchu Chen（Chinese Academy of Sciences）

Time：2022-05-18, 10:00

Venue：Tencent Meeting ID：978449682（No Password）

Abstract:

A large number of numerical experiments show that stochastic symplectic methods possess excellent long-time computational stability, when applied to stochastic Hamiltonian systems. In this talk, we attempt to give an explanation on the superiority of stochastic symplectic methods by means of large deviations principle. We prove that stochastic symplectic methods are able to asymptotically preserve the large deviations principles for some important observables of the exact solution, while non-symplectic ones do not. This indicates that stochastic symplectic methods are able to preserve the asymptotic results on rare event probabilities of the original system, and may provide an effective approach to approximating rate functions of large deviations principles.