Generalizations of Alladi's formula for arithmetical semigroups

The fact that the Riemann zeta function has a pole at \(s = 1\) implies that \(\sum_{n \geq 1} \mu(n)/n = 0\). Rearrange this as \(-\sum_{n \geq 2} \mu(n)/n = 1\), then restrict to \(n\) whose smallest prime divisor \(p(n)\) is congruent to \(\ell\) modulo \(k\), where \(k\) and \(\ell\) are coprime positive integers. \textit{K. Alladi} [J. Number Theory 9, 436--451 (1977; Zbl 0359.10036)] showed this restricted sum is \[-\sum_{\substack{n\geq 2\\ p(n) \equiv \ell \mod k}}\frac{\mu(n)}{n} = \frac{1}{\varphi(k)}.\] This article generalizes this formula to arithmetic(al) semigroups, which are free countably-generated commutative semigroups equipped with a norm function that generalize the set of natural numbers under multiplication. The arithmetic semigroups considered in this article are assumed to have the property that the number \(N(X)\) of elements with norm less than any real number \(X\) is finite and satisfies \(N(X) = cX + O(X^\eta)\) for some \(c>0\) and \(0\leq \eta < 1\) (Axiom A). Arithmetic semigroups with this property satisfy a form of the prime number theorem. The authors also consider a variant of Axiom A called Axiom A\#. Details of the proof are given in Section 2 for arithmetic semigroups satisfying Axiom A\#; presumably the proofs are similar for semigroups satisfying Axiom A, and the authors say this is the case without giving the details. The proof comes down to showing that two formulas (24), (25) are equivalent. One input to demonstrating this equivalence is a formula derived from Möbius inversion (Lemma 2.2), and the other inputs are some elementary bounds on the sum of the Möbius function over subsets of elements with a given degree (Lemmas 2.4 and 2.5). This already suffices to generalize Alladi's formula to arithmetic semigroups. Some work is then invested to further generalize their formula. It was shown by [\textit{B. Wang}, J. Number Theory 221, 232--246 (2021; Zbl 1464.11098)] that the Möbius function \(\mu\) in Alladi's formula can be replaced with a convolution \(\mu \ast a\) (in the sense of Dirichlet) with any arithmetic function \(a\) satisfying a certain boundedness condition, where the right-hand side is unchanged. As shown by [\textit{M. Kural} et al., Arch. Math. 115, No. 1, 53--66 (2020; Zbl 1446.11202)], the divisor \(p(n)\) in Alladi's formula may be allowed to range over any set of primes with natural density \(\delta\), where the right-hand side is replaced with \(\delta\). The main result of the article shows that these generalizations also hold in the arithmetic semigroup setting. In Section 4, the authors give background on some naturally occurring arithmetic semigroups before stating their formula in each case.

Recommendations

Cites work

Cited in

(5)

This page was built for publication: Generalizations of Alladi's formula for arithmetical semigroups

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2155877)