Numerical techniques for multi-scale dynamical systems with stochastic effects (Q2583446)

The paper considers numerical methods for multi-scale stochastic dynamical systems. Stochastic systems evolving on widely separated time-scales are very common in many areas of application. They represent a challenge when simulating solutions of the respective stochastic differential equations. Numerical methods for two classes of such systems are considered. These depend on the time-scale under which the evolution of the slow variable is sought. When considering rather short time intervals, the slow variables satisfy ordinary differential equations. Fluctuations become important also for slow variables when long time horizons are considered. The numerical methods proposed use simulations of appropriate auxiliary systems.