On the maximal order of numbers in the ``factorisatio numerorum'' problem (Q880071): Difference between revisions

Let \(m(n)\) be the number of ordered factorizations of \(n\geq 1\) into factors greater than \(1\). The average order of \(m(n)\) was determined by \textit{L. Kalmár} [Mat. Fyz. Lapok 38, 1--15 (1931; Zbl 0003.24604; JFM 57.1366.01)], who showed that, as \(x \rightarrow \infty\), \[ \sum_{n\leq x} m(n) = \varphi \, x^\rho (1+o(1)), \] where \(\rho\) is the positive real solution of \(\zeta(\rho)=2\), and \(\varphi= -1/\rho \zeta'(\rho)\). In the current paper, the authors make significant progress toward determining the maximal order of \(m(n)\). They show that, for \(\varepsilon >0\), \[ m(n) < \frac{n^\rho}{\exp((\log n)^{1/\rho}/(\log \log n)^{1+\varepsilon})} \] holds for all \(n>n_0\). On the other hand they prove that there is a constant \(c>0\) such that \[ m(n) > \frac{n^\rho}{\exp(c(\log n)^{1/\rho}/(\log \log n)^{1/\rho})} \] holds for infinitely many positive \(n\). The last section contains a useful survey of previous results on \(m(n)\), as well as several interesting propositions dealing with arithmetical properties of \(m(n)\). One of these results is that \(m(n)=n\) for infinitely many \(n\); another one states that \(m(n)\) is odd if and only if \(n\) is square-free.