A stronger connection between the Erdős-Burgess and Davenport constants (Q2291713)

Let \(S\) be a semigroup, the Erdős-Burgess constant \(I(S)\) is the smallest positive integer \(k\) such that every sequence over \(S\) of length \(k\) contains a nonempty subsequence whose elements multiply to an idempotent element of \(S\). Idempotents are the elements \(x\) of \(S\) with \(x^2=x\). The Davenport constant \(D(G)\) of a finite multiplicative group \(G\) is the smallest positive integer \(k\) such that any sequence of \(k\) elements of \(G\) contains a nonempty subsequence whose product is \(1\). In 2018, Hao, Wang, and Zhang conjectured that, for any integer \(n>1\), we have \[ I_r(\mathbb{Z}/n \mathbb{Z})=\mathsf{D}((\mathbb{Z}/n \mathbb{Z})^{\times})+\Omega(n)-\omega(n), \] where \(\Omega(n)\) is the total number of primes in the prime factorization of \(n\) (with multiplicity), and \(\omega(n)\) is the number of distinct primes dividing \(n\). The paper under review works on the above connection between \(I(S)\) and \(D(S)\) and confirms the conjecture for \(n\) with at most two prime factors: Let \(n = p^kq^{\ell}\), where \(p\) and \(q\) are distinct primes and \(k\) and \(\ell\) are positive integers. Then \[ I_r(\mathbb{Z}/n\mathbb{Z}) = \mathsf{D}((\mathbb{Z}/n\mathbb{Z})^\times) + (k - 1) + (\ell - 1). \] The authors also confirm the conjecture for \(n = 2p^kq^{\ell}\), where \(p\) and \(q\) are distinct odd primes. They generalize this connection to both unique factorization domains and Dedekind domains and give some other interesting remarks at the end of the paper.