High-order energy-preserving methods for stochastic Poisson systems

Abstract

A family of explicit parametric stochastic Runge-Kutta methods for stochastic Poisson systems is developed. The methods are based on perturbed collocation methods with truncated random variables and are energy-preserving. Under certain conditions, the truncation does not change the convergence order. More exactly, the methods retain the mean-square convergence order of the original stochastic Runge-Kutta method. Numerical examples show the efficiency of the methods constructed.