Scaling Limit for the Diffusion Exit Problem

Abstract

The objective of this dissertation is to prove a scaling limit for the exit of a domain problem of a small noise system with underlying hyperbolic dynamics. In this case, Large Deviation kind of estimates fail to provide a complete picture of the dynamics of the system under consideration. This is so because the rate function given by the Large Deviation Principle has several minimizing trajectories hence making them indistinguishable at the exponential level. We propose a pathwise approach based on the theory of normal forms combined with geometrical arguments. We prove a scaling limits and provide the state of the art results in related problems.