Irrational numbers associated to sequences without geometric progressions (Q2926265)

Given integers \(s\geq2\), \(k\geq3\), let \(g_k^{(s)}(n)\) denote the cardinality of the largest subset of \(\{1,2,\dots,n\}\) that contains no geometric progression of length \(k\) whose common ratio is a power of \(s\). Similarly, let \(r_k(\ell)\) denote the cardinality of the largest subset of \(\{0,1,\dots,\ell\}\) not containing an arithmetic progression of length \(k\), and \(r_k^{-1}(\ell)=\{\ell\in{\mathbb N} : r_k(\ell)=m\}\). The author prove that the limit \(\lim_{n\to\infty} g_k^{(s)}(n)/n\) exists and that it converges to \((s-1)\sum_{m=1}^\infty s^{-\min(r_k^{-1})(m)}=((s-1)^2/s)\sum_{\ell=1}^\infty r_k(\ell)/s^\ell\), where this common value is an irrational number. In the proof Szemerédi theorem is used. The irrationality statement is a special case of the following observation (Theorem 2): Let \((a_n)_{n=1}^\infty\) be an unbounded sequence of positive integers such that \(0\leq a_n-a_{n-1}\leq s-1\) for all \(n\geq2\). If \(a_n=o(n)\), then, for every integer \(s\geq 2\), the real number \(\sum_{n=1}^\infty a_n/s^n\) is irrational. Authors posed some open problems, e.g. whether the irrational number is transcendental, or whether Szemerédi theorem can be avoided.