Position vectors of numerical semigroups (Q896222)

A numerical semigroup \(S\) is s subsemigroup of \(\mathbb{N}_0\) containing \(0\) and such that \(\mathbb{N}_0 \setminus S\) is finite, where \(\mathbb{N}_0\) is the set of non negative integers. Given \(n \in S\setminus \{0\}\), the Apéry set of \(S\) with respect to \(n\) is the set Ap\((S,n)=\{\omega \in S : \omega-n \notin S\}\) and represents a very important tool in numerical semigroup theory. Let \(S=\{\lambda_0, \lambda_1, \lambda_2, \dots \}\) be a numerical semigroup containing \(n \neq 0\) such that \(\lambda_i < \lambda_j\) whenever \(i < j\) and let Ap\((S,n)=\{\lambda_0, \lambda_{x_1}, \dots, \lambda_{x_{n-1}}\}\). In this paper, the authors introduce the \textit{position vector} \(\text{pv}_n(S)\) of \(S\) with respect to \(n\) as the \((n-1)\)-tuple \((x_1, x_2-x_1,x_3-x_2, \dots, x_{n-1}-x_{n-2})\). They prove that if \(S\) and \(T\) are two numerical semigroups containing \(n\), then \(\text{pv}_n(S)=\text{pv}_n(T)\) if and only if \(S=T\); moreover, they show a method to re-construct \(S\) from a position vector. Since not every \((n-1)\)-tuple of positive integers is the position vector of a numerical semigroup, the definition is generalized to numerical sets, that are subsets of \(\mathbb N_0\) containing \(0\) and with finite complement in \(\mathbb{N}_0\). In this case, it is proved that there is a one-to-one correspondence between the numerical sets closed under addition by \(n\) and the elements of \(\mathbb{N}^{n-1}\). Finally, the problem of determining which position vectors correspond to numerical semigroups is also considered, see Theorem 3.5. In particular, an explicit answer is given for \(n \leq 5\).