Bracelet monoids and numerical semigroups (Q300874): Difference between revisions

Let \(n_1,\ldots, n_p\) be positive integers. A submonoid \(M\) of \((\mathbb{N},+)\) is an \((n_1,\ldots, n_p)\)-bracelet if \(a+b+\{n_1,\ldots, n_p\} \subseteq M\) for every \(a,b \in M\setminus \{0\}\). For example, \(M\) is an \((n)\)-bracelet if and only if \(n+ (M\setminus \{0\})\) is a semigroup, that is, the translate of this semigroup by a positive integer is also a semigroup. An \((n_1,\ldots, n_p)\)-bracelet \(M\) is called a numerical \((n_1,\ldots,n_p)\)-bracelet if \(\text{gcd}(M) = 1\). A numerical semigroup is a submonoid of \((\mathbb{N},+)\) with finite complement. One can check that \((n_1,\ldots, n_p)\)-bracelets fit into the study of patterns on numerical semigroups [\textit{M. Bras-Amorós} et al., Int. J. Algebra Comput. 23, No. 6, 1469--1483 (2013; Zbl 1291.20057)]. The authors give a characterization of the smallest \((n_1,\ldots,n_p)\)-bracelet containing a finite set \(X\) of positive integers. They show that the intersection of any two numerical \((n_1,\ldots, n_p)\)-bracelets is a numerical \((n_1,\ldots, n_p)\)-bracelet. More precisely, the set of numerical \((n_1,\ldots, n_p)\)-bracelets forms a Frobenius variety, which means that one can organize this set in a tree rooted at \(\mathbb{N}\) [\textit{J. C. Rosales}, Houston J. Math. 34, No. 2, 339--348 (2008; Zbl 1198.20051)]. This tree gives a recursive way to understand this set. They also show that a submonoid \(M\) of \((\mathbb{N},+)\) is an \((n_1,\ldots, n_p)\)-bracelet if and only if it is an intersection of numerical \((n_1,\ldots, n_p)\)-bracelets. An \((n_1,\ldots, n_p)\)-bracelet is indecomposable if it cannot be expressed as an intersection of \((n_1,\ldots, n_p)\)-bracelets that properly contain it. The authors give a criertion for when a numerical \((n_1,\ldots, n_p)\)-bracelet is indecomposable, which leads to an algorithm to check this condition.