Gaps in nonsymmetric numerical semigroups. (Q5962245)

A numerical semigroup \(S\) is a submonoid of \(\mathbb{N}\) with finite complement. We call the complement \(\Delta(S):=\mathbb{N}\setminus S\) the set of gaps of \(S\). The largest gap \(F(S)\) is called the Frobenius number of \(S\). The authors distinguish between two types of gaps, \(g\)-gaps and \(h\)-gaps, defined by, \(\Delta_{\mathcal G}(S)=\{g\in\Delta(S)\mid F(S)-g\in S\}\) and \(\Delta_{\mathcal H}(S)=\{g\in\Delta(S)\mid F(S)-g\in\Delta(S)\}\), respectively. The bulk of this paper is dedicated to studying \(h\)-gaps, particularly the minimal and maximal \(h\)-gaps of \(S\), for various families of numerical semigroups. Symmetric numerical semigroups are exactly those with no \(h\)-gaps, and pseudosymmetric semigroups are those with a unique \(h\)-gap. The first part of this paper uses techniques of commutative algebra to study the generating functions \(\sum_{s\in\Delta_{\mathcal G}(S)}z^s\) and \(\sum_{s\in\Delta_{\mathcal H}(S)}z^s\). The paper then studies \(h\)-gaps of different types of semigroups that have received attention recently: those with maximal embedding dimension, those with maximal or almost maximal length, and those with embedding dimension three. In this last setting the authors use properties of the Johnson matrix and earlier results of \textit{L. G. Fel}, [Funct. Anal. Other Math. 1, No. 2, 119-157 (2006; Zbl 1194.20058)], to prove their results.