Dimension of invariant measures for maps with exponent zero

From MaRDI portal





Let \(\mu\) be a probability measure on [0,1] and let \(d\geq 0\). If the limit \[ \lim_{\eta \to 0}\frac{\log \mu ([x-\eta,x+\eta])}{\log \eta}=d \] \(\mu\)-a.e., then it is said that the dimension of \(\mu\) is d. It is well-known that then the Hausdorff dimension of \(\mu\) is also d. For any \(r\in N\) the authors give an example of a \(C^ r\)-map of \([0,1]\) into itself with an f-invariant non-atomic ergodic probability measure \(\mu\) such that the entropy of \(([0,1],f,\mu)\) is 0 and the dimension of \(\mu\) does not exist. The authors also prove that there exist a \(C^{\infty}\)-map of \([0,1]\) into itself and an f-invariant non-atomic ergodic probability measure with dimension, entropy and Lyapunov exponent all 0. This is proved by considering exactly the example from the second author [Ergodic Theory Dyn. Syst. 2, 405-415 (1982; Zbl 0522.58032)]. The third result concerns maps lying in the stable manifold of Feigenbaum's fixed point. Let f be such a map and let \(\mu\) be a unique f-invariant non-atomic ergodic probability. The entropy of \(([0,1],f,\mu)\) is 0. The authors give the formula for the dimension of \(\mu\). A similar result has been obtained independently by \textit{E. B. Vul}, \textit{Ya. G. Sinaj} and \textit{K. M. Khanin} [Usp. Mat. Nauk 39, No. 3(237), 3-37 (1984; Zbl 0561.58033)]. From this formula it follows that the dimension of \(\mu\) is positive and the Lyapunov exponent of f is 0.




Cited in
(29)








This page was built for publication: Dimension of invariant measures for maps with exponent zero

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q3748515)