Numerical Semigroups with a e = 2 g + 1 a_{e}=2g+1
Michael Hellus, Reinhold Hübl, Anton Rechenauer · 2026
Read Numerical Semigroups with a e = 2 g + 1 a_{e}=2g+1 by Michael Hellus, Reinhold Hübl, Anton Rechenauer (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2604-21948.txt.
Introduction
Hellus, Hübl, and Rechenauer discuss numerical semigroups with a generator as large as possible: 2g plus 1, where g counts the semigroup's genus. These number sets are close to symmetric ones and have marked symmetry traits. They settle Wilf's question in the affirmative, as do some others derived from them.
Catalog record
| CREATOR | Michael Hellus, Reinhold Hübl, Anton Rechenauer |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 25603 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2604.21948
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Numerical Semigroups with a e = 2 g + 1 a_{e}=2g+1 Michael Hellus, Reinhold Hübl, Anton Rechenauer Abstract . This article discusses numerical semigroups having a generator which is as large as possible. This turns out to be 2 g + 1 2g+1 , where g g is the genus of the semigroup. We will show that these semigroups are closely related to symmetric semigroups and have interesting symmetry properties themselves. Furthermore we will show that Wilf’s question has a positive answer for these semigroups and some semigroups derived thereof. 2020 Mathematics Subject Classification 11D07, 13H10, 13F20, 20M14 Dedicated to the memory of Rolf Waldi 1. Introduction Numerical semigroups have been studied intensely since their first appearance in [ 10 ] , nevertheless many questions about them are still wide open. One particularly intriguing problem is a question of Wilf concerning the ratio of gaps and sporadic elements of such a semigroup (cf. [ 11 ] ): For each such numerical semigroup S let F(S) be its Frobenius element, i.e. the largest integer not in S , let g(S) be its genus, i.e. the number of positive integers not in S (called the gaps of S ) and let e(S) be the number of minimal generators of S (the embedding dimension of S ). Wilf’s Question . For each numerical semigroup S\neq\mathbb{N} , is it true that \frac{g(S)}{F(S)+1}\leq\frac{e(S)-1}{e(S)}\,\,? If \sigma(S) denotes the number of integers n<F(S) that are in S (called the sporadic elements of S ), then the question may be reformulated as Wilf’s Question . For each numerical semigroup S\neq\mathbb{N} , is it true that F(S)+1\leq e(S)\cdot\sigma(S)\,\,? As F(S)+1=\sigma(S)+g(S) , this provides a relation between the number of gaps and the number of sporadic elements of a numerical semigroup. Though this question has a positive answer in a number of cases, it is still wide open in general (see [ 2 ] for a survey). In this note we study semigroups having a generator “as large as possible” which turns out to be 2\cdot g(S)+1 . We show that these classes of semigroups have interesting symmetry properties and are closely related to symmetric semigroups, though with the exception of the multiplicity 2 case they are not symmetric themselves. We also show that for these semigroups and some semigroups derived from them the answer to Wilf’s question is positive. Numerical semigroups in general are closely related to curve singularities as they naturally appear as value semigroups of unibranch curve singularities. The semigroups studied here bear a close relationship to the classification of curve singularities by parametrization as curve singularities whose value semigroups satisfy a_{e}=2g+1 show that the bounds on the \mathcal{L} –determinacy, obtained in [ 6 ] , are sharp in general. Parts of the research for this paper were done while the second author was visiting Purdue University. He expresses his gratitude to this institution for its support and hospitality. 2. Maximal possible generator Throughout this paper S=\langle a_{1},\ldots,a_{e}\rangle will be a numerical semigroup with a_{1}<a_{2}<\cdots<a_{e} , embedding dimension e (so that a_{1} , \ldots , a_{e} are the minimal generators of S ) and multiplicity m=a_{1} . By g=g(S) we denote its genus and by f=F(S) its Frobenius element. Lemma 2.1 . In the above situation a_{e}\leq 2g+1 Proof. It suffices to show that any \lambda\in\mathbb{N} with \lambda\geq 2\cdot g+2 is contained in S_{+}+S_{+} . Let t=\lfloor\frac{\lambda}{2}\rfloor . Then t\geq g+1 , and we have t pairwise distinct presentations \begin{array}[]{l c l c l}\lambda&=&1&+&(\lambda-1)\\ \lambda&=&2&+&(\lambda-2)\\ &\vdots&&&\\ \lambda&=&t&+&(\lambda-t)\\ \end{array} As t>g for at least one equation \lambda=i+(\lambda-i) neither i nor \lambda-i can be gaps of S , i.e. both i and \lambda-i are in S_{+} , proving the lemma. ∎ Proposition 2.2 . Let S=\langle a_{1},\ldots,a_{e}\rangle be a numerical semigroup of genus g . Then the following are equivalent: (1) a_{e}=2g+1 . (2) The assignment \varphi given by \varphi(n)=2g+1-n induces a well defined bijection \varphi:\{1,\ldots,2g\}\cap S\longrightarrow\{1,\ldots,2g\}\setminus S . Proof. First assume a_{e}=2g+1 . We get g equations \begin{array}[]{l c l c l}a_{e}&=&1&+&(a_{e}-1)\\ a_{e}&=&2&+&(a_{e}-2)\\ &\vdots&&&\\ a_{e}&=&g&+&(a_{e}-g)\\ \end{array} As a_{e}\notin S_{+}+S_{+} in each equation at least one summand must be not in S . As there are only g gaps, the existence and bijectivity of \varphi follows. Assume now that a bijection \varphi:\{1,\ldots,2g\}\cap S\longrightarrow\{1,\ldots,2g\}\setminus S is given by the assignment \varphi . Then for each 1\leq n\leq g in the representation 2g+1=n+(2g+1-n) exactly one summand is in S and one is not. Thus 2g+1\notin S_{+}+S_{+} . On the other hand, all g gaps of S are contained in \{1,\ldots,2g\} by the bijectivity of \varphi , hence 2g+1\in S . Thus it is a minimal generator of S , and by Lemma 2.1 necessarily the last one. ∎ Proposition 2.3 . Let S=\langle a_{1},\ldots,a_{e}\rangle be a numerical semigroup of genus g . Then a_{e}=2g+1 if and only if S^{\prime}=S\setminus\{a_{e}\} is a symmetric semigroup. Moreover in case a_{e}=2g+1 we have: If e=2 , then a_{1}=2 and S^{\prime}=\langle 2,a_{e}+2\rangle and if e>2 , then S^{\prime}=\langle a_{1},\ldots,a_{e-1}\rangle . In both cases we have m(S^{\prime})=m(S) . Proof. Assume a_{e}=2g+1 . As a_{e} is a minimal generator of S , the set S^{\prime}=S\setminus\{a_{e}\} certainly is a numerical semigroup with g(S^{\prime})=g+1 and F(S^{\prime})=2g(S^{\prime})-1=2g+1 . Hence g(S^{\prime})=\frac{F(S^{\prime})+1}{2} , implying that S^{\prime} is symmetric. Conversely suppose the S^{\prime}=S\setminus\{a_{e}\} is symmetric. Then F(S^{\prime})=2g(S^{\prime})-1=2g+1 . If a_{e}\neq F(S^{\prime}) , then 2g+1=F(S^{\prime})=F(S)\leq 2g , a contradiction. If e=2 , S=\langle m,l\rangle with l\geq m+1 , then 2g+1=(m-1)\cdot(l-1)+1 by the formula for the Frobenius element in numerical semigroups generated by two elements, hence 2g+1>l except in case m=2 in which case 2g+1=l and S^{\prime}=\langle a_{1},l+2\rangle=\langle 2,a_{e}+2\rangle , and in case e>2 the additional statement is a consequence of the following lemma. ∎ Lemma 2.4 . Let S^{\prime} be a symmetric numerical semigroup with multiplicity m\geq 3 . Then S^{\prime} is generated by its elements less than F(S^{\prime}) . Proof. As S^{\prime} is symmetric, F(S^{\prime})+1=2g(S^{\prime}) . Thus by Lemma 2.1 any element s>F(S^{\prime})+2 cannot be part of a minimal generating set of S^{\prime} . But by symmetry (and as m>2 ) F(S^{\prime})+1-m=F(S^{\prime})-(m-1)\in S^{\prime} and F(S^{\prime})+2-m=F(S^{\prime})-(m-2)\in S^{\prime} , hence all generators are smaller than F(S^{\prime}) . ∎ Remark 2.5 . The symmetric numerical semigroup S^{\prime}\coloneqq S\setminus\{a_{e}\} in Proposition 2.3 is derived from an arbitrarily given numerical semigroup S with a_{e}=2g+1 (notation from above). In fact this construction establishes a one-to-one correspondence between numerical semigroups S having a_{e}=2g+1 and symmetric numerical semigroups S^{\prime} , the inverse operation is given by S\coloneqq S^{\prime}\cup\{F(S^{\prime})\} , where F(S^{\prime}) denotes the Frobenius number of S^{\prime} . This is obviuos by Proposition 2.3 and Lemma 2.4 . Corollary 2.6 . Let S=\langle a_{1},\ldots,a_{e}\rangle be a numerical semigroup with multiplicity m and with a_{e}=2g+1 . Then F(S)=a_{e}-m=2g+1-m Let S=\langle a_{1},\ldots,a_{e}\rangle be a numerical semigroup with multiplicity m=a_{1} and n\in\mathbb{N} be a positive integer. The set \operatorname{RG}(n,S)=\{L\in\{1,\ldots,n-1\}|L\notin S,n-L\notin S\} is called the set of reflected gaps of S with respect to n . By \operatorname{Ap}(S) we denote the Apéry–set of S , i.e. \operatorname{Ap}(S)=\{s\in S\,|\,\,s-m\notin S\} Note that \operatorname{Ap}(S) contains exactly m elements, that \operatorname{Ap}(S)\cup\{m\} contains all minimal generators of S and that for each Pseudo–Frobenius element p of S we have p+m\in\operatorname{Ap}(S) . Recall that an integer p is called Pseudo–Frobenius element of S if p\notin S but p+s\in S