On the maximal cardinality of half-factorial sets in cyclic groups (Q2575647)

For a subset \(G_0\) of an abelian group \(G\) the block monoid over \(G_0\), denoted by \({\mathcal B}(G_0)\), is defined as the monoid of all zero-sum sequences in \(G_0\). It is called half-factorial if for every irreducible blocks \(a_1,\dots,a_n,b_1, \dots,b_m\in{\mathcal B}(G_0)\), equality \(a_1 \dots a_n=b_1\dots b_m\) implies \(n=m\). Let \(\mu(G)\) denote the maximal cardinality of \(G_0\subset G\) for which \({\mathcal B}(G_0)\) is half-factorial. This constant appears in a natural way in the quantitative theory of factorization in algebraic number fields as observed by \textit{W. Narkiewicz} [Colloq. Math. 42, 319--330 (1979; Zbl 0514.12004)] who among others, asked for a formula for \(\mu(G)\) for a finite abelian group \(G\) in terms of its basic invariants. In general the problem remains unsolved up to date, even for cyclic groups. It is known that \(\mu(\mathbb{Z}/n \mathbb{Z})\leq\tau(n)\), where \(\tau(n)\) denotes the number of divisors of \(n\) [\textit{J. Śliwa}, Colloq. Math. 46, 123--130 (1982; Zbl 0514.12005)]. The present authors concentrate on the size of \(\mu\mathbb{Z}/n\mathbb{Z})\) and prove several results comparing it with the divisor function \(\tau(n)\). The reader is directed to the original paper for the full details, here is just a selection of proved facts. For every \(n\) one has \(\mu(\mathbb{Z}/n\mathbb{Z})\geq 1+\frac 12\tau(n)\). For two fixed positive integers \(k,n\), \(k>1\), we have \(\lim_{\nu\to\infty}\mu(\mathbb{Z}/k^\nu n\mathbb{Z})/\tau(k^\nu n)=1\). If \(n\) has at most two prime powers then \(\mu(\mathbb{Z}/n\mathbb{Z})=\tau(n)\). If \(n\) is a product of three distinct primes then \(\mu(\mathbb{Z}/n\mathbb{Z})\in\{\tau(n)-1,\tau(n) \}\). If \(n\) is a product of four distinct primes then \(\tau(n)-4\leq \mu(\mathbb{Z}/n \mathbb{Z})\leq\tau(n)\).