Ergodic properties of the Bolyai-Rényi expansion (Q5935912): Difference between revisions

The Bolyai-Rényi expansion of a number \(x\) is given by \[ T(x)=\begin{cases} x^2+2x&\text{if \(0\leq x<\sqrt 2-1\)}\\ x^2+2x-1&\text{if \(\sqrt 2-1\leq x<\sqrt 3-1\)}\\ x^2+2x-2&\text{if \(\sqrt 3-1\leq x\leq 1\)}\end{cases}, \] and the statistical properties of these expansions are given by the ergodic properties of \(T\). The authors provide very accurate approximate ergodic properties of the Bolyai-Rényi expansion with respect to the absolutely continuous \(T\)-invariant probability measure \(\mu\), for which there is no known formula. Firstly, they compute the corresponding metric entropy \(h(\mu)\) as \(h(\mu)=1.05631307402\), and also give an empirical estimate \(h(\mu)=1.056313074\). Secondly, they compute the frequency of the digits \(0\), \(1\) and \(2\) occurring in almost every expansion, from which the asymptotic arithmetic mean of the digits are derived. Thirdly, they estimate the rate of decay of correlations to be \(\alpha=0.40582\), and prove rigorously that this is within \(0.000076\) of the true value.