Minimal zero-sum sequences over \([[-m, n]]\) (Q2312892)

Let \(G\) be an abelian group and let \(X\subset G\) be a nonempty finite subset. A sequence over \(X\) is a finite unordered sequence with terms from \(X\) and repetition is allowed. We say \(S\) is zero-sum if all terms of \(S\) sums zero. If \(S\) is a zero-sum sequence but any nonempty proper subsequence of \(S\) is not a zero-sum sequence, \(S\) is called a minimal zero-sum sequence. The Davenport constant \(\mathsf D(X)\) is defined as \[\mathsf D(X)=\sup\{|S|\colon S \text{ is a minimal zero-sum sequence over }G\}\,,\] which is the most studied invariant in zero-sum theory. Since \(X\) is finite, we know \(\mathsf D(X)\) is finite. The authors study a special case: \(G=\mathbb Z\) and \(X=\{-m, -(m-1), \ldots, -1,0,1,\ldots, n-1,n \}\) is a discrete interval, where \(m,n\) are both positive integers. By showing a structure result of long minimal zero-sum sequences over \(X\) (see Theorem 3.8), they obtain that \(\mathsf D(X)=\sup_{x,y\in \mathbb N\text{ with } 1\le x\le m, 1\le y\le n }\frac{x+y}{\gcd{x,y}}\) for all but finite pairs \((m,n)\in \mathbb N^2\).