Exponential decay of random correlations for random Anosov systems mixing on fibers (Q6065736)

A random dynamical system is usually modeled by a skew product map on an product space of \[ S:\Omega\times M \to \Omega\times M, S(\omega, x) =(\theta\omega, f_\omega(x)), \] where \((\Omega, \theta)\) is a dynamical system preserving a probability measure \(P\) called driving system, or external force, \(f_\omega:M\to M\) is a fiber map for every \(\omega\in \Omega\) which is the object of interest of this paper. The author considers fiberwise mixing random Anosov systems, which means that for every \((\omega, x)\in \Omega\times M\) there exists a continuous splitting \(E^s(\omega, x)^s\oplus E^c(\omega, x)=T_xM\) such that \(Df_\omega\) is a uniform contraction on \(E^s(\omega, x)^s\) and a uniform expansion on \(E^s(\omega, x)^u\). Under additional assumption on the fiberwise topological mixing, which is an adaptation of the usual topological mixing, the author obtains exponential decay rates for future and past correlations. He uses cone technique to prove the necessary estimates.