Orbits of \(N\)-expansions with a finite set of digits (Q2143271)

The authors study more deeply the so-called \(N\)-expansions (a type of continued fractions) introduced in [\textit{C. Kraaikamp} and \textit{N. Langeveld}, J. Math. Anal. Appl. 454, No. 1, 106--126 (2017; Zbl 1366.65002)].\\ These expansions are a subclass of the new continued fractions introduced by \textit{E. B. Burger} et al. [J. Number Theory 128, No. 1, 144--153 (2008; Zbl 1210.11078)].\\ In the introduction (\S1) they give definitions and some historical background material. \\ The case \(N\in\mathbb{N}_{\geq 2}\) and \(\alpha\in\mathbb{R}\) such that \(0<\alpha<\sqrt{N}-1\) is considered, where \(N_{\geq k}\) is the set of positive integers \(n\geq k\) (\(N_1=\mathbb{N}\)). Define \(I_{\alpha}=[\alpha,\alpha+1]\) and \(I_{\alpha}^{-}=[\alpha,\alpha+1)\). The \(N\)-expansion map \(T_{\alpha}\,:\,I_{\alpha}\rightarrow I_{\alpha}^{-} \hbox{or }I_{\alpha}\) ias now given by \[T_{\alpha}(x)=\frac{N}{x}-d(x),\] with \(d\,:\,I_{\alpha}\rightarrow \mathbb{N}\) defined by \[d(x)= \lfloor \frac{N}{x}-\alpha\rfloor-1\hbox{ if }x\in(\alpha,\alpha+1]\hbox{ or both }x=\alpha,\,\frac{N}{\alpha}-\alpha\in\mathbb{Z},\] and \[d(\alpha)= \lfloor \frac{N}{\alpha}-x\alpha\rfloor-1\hbox{ if }\frac{N}{x}-x\in\mathbb{Z}.\] Fixing \(\alpha\in (0,\sqrt{N}-1)\), take \(x\in I_{\alpha}\) and for \(n\in\mathbb{N}\) \[d_n=d_n(x)=d(T_n^{\alpha -1}(x)).\] The continued fraction appearing is denoted by \[x=[d_1,d_2,d_3,\ldots]_{N,\alpha}\] (i.e., an infinite expansion for \(x\in I_{\alpha}\)).\\ \textit{Cylinder sets of rank \(1\)} are \(\Delta=\{x\in I_{\alpha}\,:\,d(x)=i\}\) with \(d_{\min}\leq i\leq d_{\max}\) where \(d_{\max}=d(\alpha)\) is the largest and \(d_{\min}=d(\alpha +1)\) is the smallest partial quotient given \(N\) and \(\alpha\). In each cylinder set the map \(T_{\alpha}\) has one \textit{fixed point} only. The main object of this paper is to study \(\{T_{\alpha}^n(x)\}_{n\geq 0}\) for \(x\in I_{\alpha}\), called \textit{the orbit of \(x\) under \(T_{\alpha}\)}; specifically special subsets that will be denoted by \textit{gaps}: Definition 1. A maximal interval \((a,b)\subset I_{\alpha}\) is called a \textit{gap} of \(T_{\alpha}\) if for almost every \(x\in I_{\alpha}\) there is an \(x_0\in\mathbb{N}\) with \(T_{\alpha}^n(x)\not\in (a,b)\) for all \(n\geq n_0\).\\ (`almost every' : excluded fixed points and their pre-images). Now a cylinder set is called \textit{gapless} if there are no gaps with non-empty intersection with the set of cylinders. After these preliminaries, the main results follow `illuminated' by a host of very nice computer generated pictures (`a host of golden daffodils' to quote William Wordsworth); they will be indicated shortly below. Theorem 1. The interval \(I_{\alpha}\) consists of \(m\) full cylinder sets, with \(m\in\mathbb{N}_{\geq 2}\), if and only if there is a positive integer \(k\) such that \[\alpha=k,\ n+mk(k+1),\ d(\alpha)(m-1)(k+1).\] Theorem 2. Let \(N\in\mathbb{N}_{\geq 2}\). Then there is a unique absolutely continuous invariant probability measure \(\mu_{\alpha}\) such thet \(T_{\alpha}\) is ergodic with respect to \(\mu_{\alpha}\). Theorem 3. Let \(N\in\mathbb{N}_{\geq 3}\) and let \(0<\alpha\leq \sqrt{N}-1\). Let \(|T_{\alpha}'(x)|>2\) for all \(x\in I_{\alpha}\). Then \(I_{\alpha}\) contains no gaps. Theorem 4. Let \(I_{\alpha}\) consist of five cylinder sets or more. Then \(I_{\alpha}\) has no gaps. Theorem 5. Let \(I_{\alpha}=\Delta_d\cup \Delta{d-1}\). If \(T_{\alpha}(\alpha)\geq f_{d-1}\) and \(T_{\alpha})\alpha +1)\leq f_d\), then \(I_{\alpha}\) is gapless. Theorem 6. Let \(I_{\alpha}=\Delta_d\cup\cdots\cup \Delta_{d-m}\) with \(m\in\{2,3\}\). Then \(I_{\alpha}\) is gapless if \[T_{\alpha}(\alpha)\geq f_{d-1}\hbox{ or }T_{\alpha}(\alpha +1)\leq f_{d-m+1}.\] Theorem 7. Let \(N\in\mathbb{N}_{\geq 2}\) and \(I_{\alpha}=\Delta_d\cup\Delta_{d-1}\cup\Delta_{d-2}\cup\Delta_{d-3}\). Then there is a gap in \(I_{\alpha}\) if and only if \(N=2k^2+2k-i\), with \(k>1\) and \(i\in\{1,2,3\}\). Moreover, if there is a gap in \(I_{\alpha}\), the gap contains \(f_{d-1}\) and \(f_{d-2}\), while \(\Delta_d\) and \(\Delta_{d-3}\) are gapless. The paper concludes with a section References (\(13\) items).\\ A nicely written paper.