A path integral method for coarse-graining noise in stochastic differential equations with multiple time scales (Q617539): Difference between revisions

The authors propose a method to derive from a given stochastic differential equation, corresponding to a stochastically perturbed ordinary differential equation with multiple time scales, a new stochastic differential equation describing the system's evolution on slow time scales. For this purpose, the authors start from the path integral representation of the stochastic system and apply a multi-scale expansion to the path integral kernel of the corresponding Lagrangian. The deterministic term in the stochastic differential equation is assumed to be periodic in time, so as to have a natural separation between the fast time scale (describing the evolution within a particular period) and the slow scale (describing the evolution over many periods). The method is applied to a system that arises in the study of random dispersion fluctuations in dispersion-managed fiber-optic communications.