scientific article; zbMATH DE number 3366985
Let (i) \(f\) be an arithmetic function defined on \(\mathbb N= \{1,2,3,\ldots\}\) with \(f(1) =0\), \[ g(a,b) = f(ab) f(a) - f(b)\qquad \forall a,b\in\mathbb N, \tag{\text{ii}} \] \[ d(u) =f(u+1) - f(u) \qquad \forall u\in U\subseteq \mathbb N, \tag{\text{iii}} \] \[ U(x) = \left(\sum_{u\le x} 1\right) \biggr/x,\text{ where }u\in U\subseteq \mathbb N,\text{ and }x \text{ is real and positive}. \tag{\text{iv}} \] Then \(f\) is called additive, if \((a,b)=1 \Rightarrow g(a,b) =0\); and \(f\) is called completely additive, if \(g(a,b) \equiv 0\). \textit{P. Erdős} [Ann. Math. (2) 47, 1--20 (1946; Zbl 0061.07902); Rend. Sem. Mat. Fis. Milano 27, 45--49 (1958; Zbl 0081.04205)] proved that \(f\) is additive and \(\lim_{n\to\infty} d(n) = 0\), \(n\in\mathbb N \Rightarrow f=c\cdot \ln\), where the value of \(c\) is independent of \(n\). Further he conjectured that \(\lim_{x\to\infty} U(x) =1\) and \(\lim_{u\to\infty} d(u) =0\) imply the same conclusion, provided \(f\) is additive. In the present booklet, the author investigates closely the same problem by different methods, establishes sufficient conditions for \(f\), which is additive, to be completely additive and establishes that Erdős' conjecture is true. Refinements and variations of the problem, also associated with an analogous conjecture of Erdős, have since been published by the author [cf. e.g., Ist. Lombardo Accad. Sci. Lett., Rend., Sez. A 103, 931--938 (1969; Zbl 0209.35102); Atti Accad. Naz. Lincei, VIII. Ser., Rend., Cl. Sci. Fis. Mat. Nat. 48, 10--13 (1970; Zbl 0213.33103)] as also by several other mathematicians [cf. e.g. \textit{E. Wirsing}, Symp. Math., Roma 4, 45--57 (1970; Zbl 0223.10036)].
This page was built for publication:
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q5639754)