Capturing deviation from ergodicity at different scales (Q843015)

Let \((X, \Sigma, \mu)\) be a probability space, \(T: X \to X \) be a measure preserving map. Birkhoff's individual ergodic theorem states that for an \(f\in L_1\) \[ \lim_n \frac1n \sum_{k=0}^{n-1} f(T^k x) = f^* (x) \quad \text{exists a.e., } \] and furthermore, if \(T\) is ergodic, then \(f^* (x) =\int_X f \,d\mu\). In this article, the authors introduce a computational technique that quantifies the deviation of a dynamical system \((X,T)\) from being ergodic. This deviation is defined in terms of \(\| f^* - \int_X f d\mu \|_2^2, \) called the ergodicity defect of \(T. \) The technique essentially consists of the investigation of the effect of scaling changes on the ergodicity defect, and they conduct scaling analysis for several examples of \(T\) using the Haar wavelet and Haar scaling function.