Non uniform expansion and additive noise imply random horseshoe
Jeroen S. W. Lamb, Giuseppe Tenaglia, Dmitry Turaev
TL;DR
This work extends the random-dynamical-systems theory of Smale horseshoes to non-uniformly expanding maps with diffusive additive noise by proving the density of random horseshoes in the support of an ergodic stationary measure when all Lyapunov exponents are positive. The authors combine large-deviation estimates for Lyapunov-type averages with annealed control of large-ball dynamics and a refined Young-time framework to build a random Young tower, from which a κ-horseshoe is extracted. Key contributions include an annealed large-deviation scheme via a perturbed transfer operator, a density result for random horseshoes under mild regularity assumptions on the critical set, and a constructive horseshoe scheme relying on independent noise and stopping-time arguments. The findings generalize previous random-horseshoe results beyond predominantly expanding or uniformly hyperbolic settings, with implications for chaotic structure in stochastic dynamics. All mathematical notation is presented with explicit delimiters to ensure precise interpretation.
Abstract
We propose a notion of random horseshoe and prove density of random horseshoes for non uniformly expanding random dynamical systems with additive noise
