On the \(r\)-continued fraction expansions of reals (Q2329276)

One of the types of generalised continued fractions is \(r\)-continued fraction. An expression of the form \[ a_0+\cfrac{r}{a_1+\cfrac{r}{a_2+\genfrac{}{}{0pt}{0}{}{ \ddots} }} \] is called an \(r\)--continued fraction, where \(r\in \mathbb{R}, r>1, \) is some fixed real number, \(a_0\in \mathbb{Z} \), and all of the coefficients \(a_k \in \mathbb{N}, k \geq 1 \). The \(r\)-continued fraction map \(T_r: [0,1) \to [0,1) \) is defined by \[ T_r(x)=\frac{r}{x}-\Big\lfloor \frac{r}{x}\Big\rfloor \quad \text{if}\; x\neq 0 \; \text{and} \; T_r(0)=0. \] Let \(x \in [0,1)\) and given setting \(x_0=x, x_1=T_r(x), x_2=T^2_r(x), \dots \) and setting \[ a_n=a_n(x)=\begin{cases} \big\lfloor \frac{r}{x_{n-1}}\big\rfloor, & if \; x_{n-1}\neq 0,\\ \infty, & if \; x_{n-1}=0, \end{cases} \quad n\geq 1. \] Then has \[ x=\cfrac{r}{a_1+\cfrac{r}{a_2+\genfrac{}{}{0pt}{0}{}{\ddots \genfrac{}{}{0pt}{0}{}{+\cfrac{r}{a_n+x_n}}}}}=[a_1,a_2,\dots,a_n+x_n]_r, \; n\geq 1. \] If \(x \in \mathbb{R}, \) then \(x=a_0+x',\) where \(x'\in [0,1), a_0=\big\lfloor x\big\rfloor\) and \(x=[a_0;a_1,a_2,\dots]_r\). Theorem. For each \(r>1\), every real number \(x\) has an \(r\)-continued fraction expansion. Theorem. Let \(a_0 \in \mathbb{Z} \) and let \((a_n)_{n\geq 1}\) be a sequence of positive integer. Then there exist a real number \(x\) for which \(\lim\limits_{n \to \infty}[a_0;a_1,a_2,\dots,a_n]_r =x.\) The space of one-sided sequences of digits of the \(r\)-continued fraction expansion considered in the third section of the paper. The alternating lexicographic order on this space defined and shown that a sequence in this space represents the \(r\)-continued fraction expansion of a real number \(x \in [0,1)\) iff the sequence and all of its shifts are alternating lexicographically less than the sequence obtained from the \(r\)-continued fraction expansion of 1. The article also proves that the \(r\)-continued fraction transformation has an ergodic absolutely continuous invariant measure that is equivalent to Lebesgue measure.