Investigation of exponential dichotomy of linear Itô stochastic systems with random initial data by using quadratic forms

Abstract

We study conditions for the mean-square exponential dichotomy of linear Itô stochastic systems. We prove that a sufficient condition for exponential dichotomy is the existence of a quadratic form whose derivative along the solutions of a system is negative definite. The converse theorem is also proved.