Invariant measure selection by noise

Etienne Pardoux
Université d'Aix-Marseille

We consider a toy example of a 3D dynamical system with two invariant quantities,
which we perturb with both noise and dissipation, so that the ODE becomes a hypoelliptic
diffusion, which has unique invariant probability measure. We precise exactly which invariant
measure of the dynamical system is the unique limit of the invariant measures of the stochastic
systems, as both noise and dissipation vanish. We thus prove much stronger results than are currently
available concerning (S)PDEs.
This is joint work with Jonathan Mattingly (Duke Univ.)

Presentation (PDF File)

Back to Rough Paths: Theory and Applications