Shrinking random \(\beta\)-transformation (Q505041): Difference between revisions

For a real number \(1<\beta<2\), the \(\beta\)-expansion of a number \(x\in\mathcal{A}_\beta=[0,(\beta-1)^{-1}]\) is an expression \[ x=\sum_{i=1}^\infty \frac{a_i}{\beta^i} \] with each \(a_i\in\{0,1\}\). It turns out that almost every \(x\) has infinitely many different \(\beta\)-expansions. A standard way of generating a \(\beta\)-expansion of \(x\) is to encode the orbit of \(x\) for a piecewise linear interval map with respect to the partition of the domain into the interval of monotonicity of the map. Two standard maps allowing to do so are known as the greedy map \(T_0\) and the lazy map \(T_1\). The second author and \textit{C. Kraaikamp} [Ergodic Theory Dyn. Syst. 23, No. 2, 461--479 (2003; Zbl 1035.37006)] considered a superimposition of these two maps, and for each \(x\) in the switching region, where the lazy and greedy map differ, they randomly choose which map to use to determine a unique image of \(x\). A point \((\omega_i)_{i\in\mathbb{N}}\in\mathbb{N}\) can be used to encode the sequence of these choices. Then coding of the orbit of \((\omega,x)\) with respect to the random transformation \(K'(\omega,x)=(\sigma(\omega),T_{\omega_0}(x))\) generates \(\beta\)-expansions and all possible \(\beta\)-expansions can be obtained in such way. In this paper under review only \(\beta=\beta_n\) for some \(n\geq 3\) are considered, where \(1<\beta_n<2\) is the largest positive real number satisfying the equation \[ \beta^n=\beta^{n-2}+\beta^{n-3}+\ldots+\beta+1 . \] The authors define the shrinking random \(\beta\)-transformation \(K\) in a similar way as \(K'\) above, with the only difference that a random choice is made only if \(x\) belongs to a specified subset of the switching region. Then they investigate natural invariant measures for \(K\), and the induced transformation defined by \(K\) on a special subset of the domain. It turns out that both transformations have a unique measure of maximal entropy (i.e., are intrinsically ergodic), but the measure induced from the intrinsically ergodic measure for \(K\) (i.e., the measure of maximal entropy of \(K\)) by inducing on the special subset is not the intrinsically ergodic measure (i.e., the measure of maximal entropy) for the induced system.