Grimm's conjecture and smooth numbers (Q1940078): Difference between revisions
From MaRDI portal
Set profile property. |
Set OpenAlex properties. |
||
Property / OpenAlex ID | |||
Property / OpenAlex ID: W2026636419 / rank | |||
Normal rank |
Revision as of 19:37, 19 March 2024
scientific article
Language | Label | Description | Also known as |
---|---|---|---|
English | Grimm's conjecture and smooth numbers |
scientific article |
Statements
Grimm's conjecture and smooth numbers (English)
0 references
5 March 2013
0 references
Grimm's conjecture states that for positive integers \(n\) and \(k\), if none of \(n+1,\ldots,n+k\) is prime number, then there exist distinct primes \(p_1,\ldots,p_k\) such that \(p_i\mid (n+i)\) simultaneously for \(1\leq i\leq k\). Denote by \(g(n)\) the maximal positive integer \(k\) such that there are distinct primes \(p_1,\ldots,p_k\) with \(p_j\mid (n+j)\) for \(1\leq j\leq k\). Using methods from transcendental number theory, \textit{K. Ramachandra} et al. [J. Reine Angew. Math. 273, 109--124 (1975; Zbl 0297.10031)] established a lower bound \(g(n)\geq c\left(\frac{\log n}{\log\log n}\right)^3\) for \(n>3\) and an absolute constant \(c>0\). In the paper under review, the authors establish upper bounds for \(g(n)\), either unconditionally or assuming a conjecture on smooth numbers in short interval. In detail, they obtain Theorem 1. (i) There exists an \(\alpha<\frac12\) such that \(g(n)<n^{\alpha}\) for sufficiently large \(n\). (ii) For \(\varepsilon>0\), we have \(\#\{n\leq X: g(n)\geq n^{\varepsilon}\}\ll_{\varepsilon} X\exp(-(\log X)^{\frac13-\varepsilon})\). Denote \(\Psi(x,y)=\#\{n\in{\mathbb Z}^+\bigcap [1,x]: p~\,\text{prime},\,p\mid n\Rightarrow p\leq y\}\). Theorem 2. Assume for \(\varepsilon>0\) we have \(\Psi(x+x^{\varepsilon},x^{\varepsilon})-\Psi(x,x^{\varepsilon})\gg x^{\varepsilon}\) for sufficiently large \(x\), then \(g(n)<n^{\varepsilon}\) for large \(n\). A weaker version of Grimm's conjecture is also studied. For an integer \(\nu>1\), denote by \(\omega(\nu)\) the number of distinct primes factors of \(\nu\) with convention \(\omega(1)=0\). It is conjectured that, if \(n+1,n+2,\ldots,n+k\) are all composite numbers, then \(\omega\left(\prod_{i=1}^k(n+i)\right)\geq k\). Denote \(g_1(n)\) the maximal positive integer \(k\) such that \(\omega\left(\prod_{i=1}^{\ell}(n+i)\right)\geq\ell\) for all \(1\leq\ell\leq k\). The authors obtain Theorem 3. There exists a \(\gamma\) with \(0<\gamma<\frac12\) such that \(g(n)\leq g_1(n)<n^{\gamma}\) for large \(n\). Theorem 4. Assume that there exist \(0<\alpha<\frac12\) and \(\delta>0\) such that \[ \sum_{j\leq m^{\alpha}}\left\{\pi\left(\frac{m+m^{\alpha}}{j}\right)-\pi\left(\frac{m}{j}\right)\right\} \geq\delta m^{\alpha} \] for large \(m\). Then \(g_1(n)<n^{\gamma}\) with \(\gamma=\max\left(\alpha,\frac{1-\delta(1-\alpha)}{2-\delta}\right)<\frac12\) for large \(n\).
0 references
Grimm's conjecuture
0 references
smooth numbers
0 references