Uniqueness of SRB measures for stochastic differential equations

Abstract

For a dissipation stochastic differential equation (abbreviated as SDE), exponential ergodicity has been investigated a lot in a wide range of settings. In this paper, we prove that the Markovian measure $μ$ determined by the unique stationary measure is also the unique invariant measure which satisfies the entropy formula. Consequently, the existence of an SRB measure is completely determined by the sign of the sample top Lyapunov exponent of $μ$: if $χ_μ^1>0$, then $μ$ is the unique SRB measure, whereas if $χ_μ^1\leq0$, then the SDE admits no SRB measure. As an application, we establish a dichotomy of SRB measures for the 2D additive-noise stochastic Hopf normal form within a shear parameter domain.