## Abstract

In this study, we characterize all families of saturated numerical semigroups with multiplicity four. We also present some results about invariants of these semigroups.

Show Summary Details# Open Mathematics

### formerly Central European Journal of Mathematics

# On the saturated numerical semigroups

#### Open Access

## Abstract

## 1 Introduction

## 2 Main results

*
**In view of Proposition 2.1, we find that **S* is saturated a numerical semigroup. □*In view of Proposition 2.1, we have that **S* is a saturated numerical semigroup. □## Acknowledgement

## References

## About the article

Editor-in-Chief: Gianazza, Ugo / Vespri, Vincenzo

1 Issue per year

IMPACT FACTOR 2015: 0.512

SCImago Journal Rank (SJR) 2015: 0.521

Source Normalized Impact per Paper (SNIP) 2015: 1.233

Impact per Publication (IPP) 2015: 0.546

Mathematical Citation Quotient (MCQ) 2015: 0.39

In this study, we characterize all families of saturated numerical semigroups with multiplicity four. We also present some results about invariants of these semigroups.

KeyWords: Saturated numerical semigroup; Frobenius number; Gaps

MSC 2010: 20M14

Let ℕ = {1, 2, ..., *n*, ..} and ℤ be the set of integers. A subset *S* of the set ℕ of nonnegative integers is called a numerical semigroup if it satisfies the following conditions:

- (i)
0 ∈

*S*, - (ii)
*a, b*∈*S*⇒*a*+*b*∈*S*, - (iii)
ℕ \

*S*has a finite number of elements.

Condition (iii) is equivalent to *gcd*(*S*) = 1 (Here, *gcd*(*S*) is the greatest common divisor of the element of *S*).

All numerical semigroups are finitely generated, i.e.
$$S=\u3008{a}_{1},{a}_{2},...{a}_{r}\u3009\phantom{\rule{thinmathspace}{0ex}}=\phantom{\rule{thinmathspace}{0ex}}\{\sum _{k=1}^{t}{c}_{i}{a}_{i},...,{c}_{r}\in \mathbb{N}\}$$
where *a*_{1}*, a*_{2}*, ... ,a _{r}* ∈

We define the following invariants of numerical semigroups: $$F(S)=max\{x\phantom{\rule{thinmathspace}{0ex}}:\phantom{\rule{thinmathspace}{0ex}}x\in \mathbb{Z}\mathrm{\setminus}S\}$$ and $$n(S)\phantom{\rule{thinmathspace}{0ex}}=\phantom{\rule{thinmathspace}{0ex}}\left|\{0,1,2,...,F(S)\}\cap S\right|.$$

*F* ( *S* ) and *n* ( *S* ) are called the Frobenius number of *S* and the number determiner of *S*, respectively.

We can write
$$S=\u3008{a}_{1},{a}_{2},...,{a}_{r}\u3009=\u3008{s}_{0}=0,{s}_{1},{s}_{2},...,{s}_{n-1},{s}_{n}=F(S)+1,\to ...\u3009$$
where *s _{i}* <

The set ℕ \ *S* is the gap of *S*, and the set of gaps of *S* is denoted by *H*(*S*). *g*(*S*) = |*H*(*S*)| is called the genus of *S*. It is clear that *g*(*S*) = *F*(*S*) + 1 − *n*(*S*). An element *x* ∈ *H*(*S*) is called a fundamental gap of *S* if 2*x*, 3*x* ∈ *S*. The set of all the fundamental gaps of *S* is denoted by *FH*(*S*), i.e.
$$FH(S)=\{x\in H(S):2x,3x\in S\}.$$

An element *x* ∈ ℤ is called a Pseudo-Frobenius number of *S* if *x* ∉ *S* and *x* + *s* ∈ *S*, for *s* ∈ *S* \ {0}. We denote by *PF*(*S*) the set of all Pseudo-Frobenius numbers of *S*, i.e.
$$PF(S)=\{x\phantom{\rule{thinmathspace}{0ex}}\in \mathbb{Z}\mathrm{\setminus}S:x+s\in S,\phantom{\rule{thinmathspace}{0ex}}\text{for}\phantom{\rule{thinmathspace}{0ex}}\text{all}\phantom{\rule{thinmathspace}{0ex}}s\phantom{\rule{thinmathspace}{0ex}}\in \phantom{\rule{thinmathspace}{0ex}}S\mathrm{\setminus}\{0\}\}$$
(see [7]). Given a numerical semigroup *S* and *x* ∈ *S* \ {0}, we define the Apery set of *x* in *S* as *Ap*(*S, x*) = {*s* ∈ *S* : *s* − *x* ∉ *S*} (for details see [9]).

If a numerical semigroup *S* satisfies the condition *x* + *y* − *z* ∈ *S*, for every *x, y, z* ∈ *S* such that *x* ≥ *y* ≥ *z*, then *S* is called Arf. If *S* is an Arf numerical semigroup, then *S* has maximal embedding dimension.

The investigation of combinatorial properties of semigroups is very important, because they often occur in applications ([1, 3, 5]) and are related to automata theory (see [4]). A numerical semigroup *S* is saturated if *s* + *c*_{1}*s*_{1} + *c*_{2}*s*_{2} + ... + *c _{k}*

In this study, we show that all families of numerical semigroups with multiplicity four are saturated numerical semigroups; these are numerical semigroups of the form *S* = 〈4, *k, k* + 2, *k* + 3〉, for *k* = 3(*mod*4) and *k* ≥ 7 and *S* = 〈4, *k, k* + *t, k* + *t* + 2〉, for *k* = 2(*mod* 4) and *k* ≥ 6 and *t* an odd integer. We also give the formulae for *F*(*S*), *n*(*S*), *PF*(*S*), *g*(*S*), *H*(*S*) and *FH*(*S*) of these numerical semigroups.

In this section we provide some results for numerical semigroups with multiplicity four; i.e. numerical semigroups of the form *S* = 〈4, *k, k* + 2, *k* + 3〉 (for *k* ≡ 3(*mod* 4) and *k* ≥ 7) and *S* = 〈4, *k, k* + *t, k* + *t* + 2〉, (for *k* ≡ 2(*mod* 4) and *k* ≥ 6 and *t* an odd integer).

**Proposition 2.1:** *([8]). Let S be a numerical semigroup, then the following conditions are equivalent:
*

*(i)**S is a saturated numerical semigroup*.*(ii)**a*+*d*_{S}*(a)*∈*S for all a*∈*S*,*a*> 0*where d*_{S}*(a)*=*gcd*{*x*∈*S*:*x*≤*a*}.*(iii)**a*+*kd*_{S}*(a)*∈*S for all a*∈*S*,*a*> 0*and k*∈ ℕ.

**Theorem 2.2:** *([10]). If S = 〈4,k, k + 1, k + 2 〉 , then S is a saturated numerical semigroup, for k ≡ 1(mod 4) and k ≥ 5.*

**Theorem 2.3:** *Let S* = 〈4, *k, k* + 2, *k* + 3〉 *be numerical semigroup, where k* ≡ 3( *mod* 4) *and k* ≥ 7. *Then S is saturated*.

**Proof:** *Let S = 〈4,k, k + 2, k + 3〉 be numerical semigroup, where k ≡ 3(mod 4) and k ≥ 7. We note that k = 4r + 3, r ≥ 1 and r ∈ ℤ. Thus, we have
$$\begin{array}{c}S=\u30084,k,k+2,k+3\u3009=\{0,4,8,\cdots ,k-7,k-3,k\to \cdots ,\}\\ =\{0,4,8,\cdots ,4r-4,4r,4r+3,\to \cdots \}.\end{array}$$In this case,
*

*(a)**If**a*< 4*r*+ 3, then*d*_{S}*(a)*=*1*. So, we find that*a*+*d*_{S}*(a)*∈*S*since*a*+*d*_{S}*(a)*=*a*+*1*≥*4r*+*4*∈*S*, for all*a*∈*S*,*a*> 0.*(b)**If**a*≥ 4*r*+ 3, then*d*_{S}*(a)*= 4. So, we have*a*+*d*_{S}*(a)*=*a*+ 4 ∈*S*, for all*a*∈*S*,*a*> 0.

**Theorem 2.4:** *Let S* = 〈4,*k, k* + *t, k* + *t* + 2〉 *be numerical semigroup, where k* ≡ 2(*mod* 4), *k* ≥ 6, *and t is an odd integer. Then S is saturated*.

**Proof:** *It is trivial that gcd {4,k, k + t, k + t + 2} = 1 since k is even and t is an odd integer. If we put k = 4r + 2, r ≥ 1 and r ∈ ℤ, then we have
$$\begin{array}{c}S=\u30084,k,k+t,k+t+2\u3009=\{0,4,8,...,k-6,k-2,k,k+2,...,k+t-3,k+t-1,\to ...,\}\\ =\{0,4,8,...,4r-4,4r,4r+2,4r+4,...,4r+t-1,4r+t+1,\to ...,\}.\end{array}$$In this case,
*

*(i)**If**a*> 4*r*+*t*+ 1, then*d*_{S}*(a)*= 1. So, we obtain*a*+*d*_{S}*(a)*∈*S*from the inequality*a*+*d*_{S}*(a)*=*a*+ 1 ≥*4r*+*t*+ 2 ∈*S*, for all*a*∈*S*,*a*> 0.*(ii)**If 4**r*≤*a*≤ 4*r*+*t*+ 1, then*d*_{S}*(a)*= 2. So, we obtain*a*+*d*_{S}*(a)*=*a*+ 2 ∈*S*from the inequality*4r*≤*a*≤ 4*r*+*t*+ 1, for all*a*∈*S*,*a*> 0.*(iii)**If**a*< 4*r*, then*d*_{S}*(a)*= 4. So, we obtain*a*+*d*_{S}*(a)*=*a*+ 4 ∈*S*since*a*+ 4*r*< 4*r*+ 4, for all*a*∈*S*,*a*> 0.

**Proposition 2.5:** *([6]). Let S be a numerical semigroup minimally generated by {n_{1} < n_{2} < ... < n_{r}}. Then S has maximal embedding dimension if and only if Ap(S, n_{1}) = {0,n_{2}, n_{3},...,n_{r}}.*

**Corollary 2.6:** *([6]). Let S be a numerical semigroup minimally generated by {n_{1} < n_{2} < ... < n_{r}}. Then the following conditions are true:
*

*(1)**If S has maximal embedding dimension, then F*(*S*) =*n*−_{r}*n*_{1}.*(2)**S has maximal embedding dimension if and only if*$$g(S)=\frac{{n}_{2}+{n}_{3}+...+{n}_{r}}{{n}_{1}}-\frac{{n}_{1}-1}{2}.$$

**Theorem 2.7:** *If S* = 〈4, *k k* + 2, *k* + 3〉 *is a numerical semigroup, where k* ≡ 3(*mod* 4) *and k* ≡ 7. *Then we obtain following equalities*:

*(a)**F*(*S*) =*k*− 1,*(b)**$g\left(S\right)\phantom{\rule{thinmathspace}{0ex}}=\phantom{\rule{thinmathspace}{0ex}}\frac{3k\phantom{\rule{thinmathspace}{0ex}}-\phantom{\rule{thinmathspace}{0ex}}1}{4}$,**(c)**PF*(*S*) = {*k*− 4,*k*− 2,*k*− 1},*(d)**$n\left(S\right)\phantom{\rule{thinmathspace}{0ex}}=\frac{k\phantom{\rule{thinmathspace}{0ex}}+\phantom{\rule{thinmathspace}{0ex}}1}{4}$,**(e)**H*(*S*) = {1, 2, 3, 5, 6, 7, ... ,*k*− 6,*k*− 5,*5k*− 4,*k*− 2,*k*− 1}.

**Proof:** *We have Ap(S, 4) = {0, k, k − 2, k − 3} since S has maximal embedding dimension. Thus,
*

*(a)**We have**F*(*S*) = (*k*+ 3) − 4 =*k*− 1 from Corollary 2.6 (1).*(b)**We obtain $g\left(S\right)\phantom{\rule{thinmathspace}{0ex}}=\frac{k\phantom{\rule{thinmathspace}{0ex}}+\phantom{\rule{thinmathspace}{0ex}}k\phantom{\rule{thinmathspace}{0ex}}+\phantom{\rule{thinmathspace}{0ex}}2\phantom{\rule{thinmathspace}{0ex}}+\phantom{\rule{thinmathspace}{0ex}}k\phantom{\rule{thinmathspace}{0ex}}+\phantom{\rule{thinmathspace}{0ex}}3}{4}\phantom{\rule{thinmathspace}{0ex}}-\phantom{\rule{thinmathspace}{0ex}}\frac{4\phantom{\rule{thinmathspace}{0ex}}-\phantom{\rule{thinmathspace}{0ex}}1}{2}\phantom{\rule{thinmathspace}{0ex}}=\phantom{\rule{thinmathspace}{0ex}}\frac{3k\phantom{\rule{thinmathspace}{0ex}}-\phantom{\rule{thinmathspace}{0ex}}1}{4}$ from Corollary 2.6 (2).**(c)**It is obvious that**PF*(*S*) = {*k*− 4,*k*+*2*− 4,*k*+ 3 − 4}. So we find*PF*(*S*) = {*k*− 4,*k*− 2,*k*− 1}.*(d)**We have $n\left(S\right)\phantom{\rule{thinmathspace}{0ex}}=\phantom{\rule{thinmathspace}{0ex}}\left(k\phantom{\rule{thinmathspace}{0ex}}-\phantom{\rule{thinmathspace}{0ex}}1\right)\phantom{\rule{thinmathspace}{0ex}}+\phantom{\rule{thinmathspace}{0ex}}1\phantom{\rule{thinmathspace}{0ex}}-\phantom{\rule{thinmathspace}{0ex}}\frac{3k\phantom{\rule{thinmathspace}{0ex}}-\phantom{\rule{thinmathspace}{0ex}}1}{4}\phantom{\rule{thinmathspace}{0ex}}=\phantom{\rule{thinmathspace}{0ex}}\frac{k\phantom{\rule{thinmathspace}{0ex}}+\phantom{\rule{thinmathspace}{0ex}}1}{4}$ from**g*(*S*) =*F*(*S*) + 1 −*n*(*S*).*(e)**We find that**H*(*S*) = {1, 2, 3, 5, 6, 7, ... , 4*r*− 3, 4*r*− 2, 4*r*− 1, 4*r*+ 1, 4*r*+*2*} = {1, 2, 3, 5, 6, 7, ... ,*k*− 6,*k*− 5,*k*− 4,*k*− 2,*k*− 1} from the equality*S*= < 4,*k*,*k*+ 2,*k*+ 3 > {0, 4, 8, ... ,*k*− 7,*k*− 3,*k*, → ... ,} = {0, 4, 8, ... , 4*r*− 4, 4*r*, 4*r*− 3, → ... ,}.

**Theorem 2.8:** *Let S*= 〈4, *k*, *k* + *t, k* + *t* + *2*〉 *be a numerical semigroup, where k* ≡ 2(*mod* 4), *k* ≡ 6*, and t is an odd integer. Then, we have following equalities*:

*(a)**F*(*S*) =*k*+*t*− 2,*(b)**$g\left(S\right)\phantom{\rule{thinmathspace}{0ex}}\phantom{\rule{thinmathspace}{0ex}}=\phantom{\rule{thinmathspace}{0ex}}\phantom{\rule{thinmathspace}{0ex}}\frac{3k\phantom{\rule{thinmathspace}{0ex}}+\phantom{\rule{thinmathspace}{0ex}}2t\phantom{\rule{thinmathspace}{0ex}}-\phantom{\rule{thinmathspace}{0ex}}4}{4}$,**(c)**PF*(*S*) = {*k*− 4,*k*+*t*− 4,*k*+*t*− 2},*(d)**$n\left(S\right)\phantom{\rule{thinmathspace}{0ex}}\phantom{\rule{thinmathspace}{0ex}}=\phantom{\rule{thinmathspace}{0ex}}\phantom{\rule{thinmathspace}{0ex}}\frac{k\phantom{\rule{thinmathspace}{0ex}}+\phantom{\rule{thinmathspace}{0ex}}2t}{4}$,**(e)**H*(*S*) = {1, 2, 3, 5, 6, 7, ... ,*k*− 5,*k*− 4,*k*− 3,*k*− 1,*k*− 1,*k*+ 3, ... ,*k*+*t*− 2}.

**Proof:** *We have Ap(S, 4) = {0, k, k + t, k + t + 2} since S has maximal embedding dimension. Thus,
*

*(a)**We have**F*(*S*) = (*k*+*t*+ 2) − 4 =*k*+*t*+ 2 from Corollary 2.6 (1).*(b)**We obtain $g\left(S\right)\phantom{\rule{thinmathspace}{0ex}}\phantom{\rule{thinmathspace}{0ex}}=\phantom{\rule{thinmathspace}{0ex}}\phantom{\rule{thinmathspace}{0ex}}\frac{k\phantom{\rule{thinmathspace}{0ex}}+\phantom{\rule{thinmathspace}{0ex}}k\phantom{\rule{thinmathspace}{0ex}}+\phantom{\rule{thinmathspace}{0ex}}t\phantom{\rule{thinmathspace}{0ex}}+\phantom{\rule{thinmathspace}{0ex}}k\phantom{\rule{thinmathspace}{0ex}}+\phantom{\rule{thinmathspace}{0ex}}t\phantom{\rule{thinmathspace}{0ex}}+\phantom{\rule{thinmathspace}{0ex}}2}{4}\phantom{\rule{thinmathspace}{0ex}}\phantom{\rule{thinmathspace}{0ex}}-\phantom{\rule{thinmathspace}{0ex}}\phantom{\rule{thinmathspace}{0ex}}\frac{4\phantom{\rule{thinmathspace}{0ex}}-\phantom{\rule{thinmathspace}{0ex}}1}{2}\phantom{\rule{thinmathspace}{0ex}}\phantom{\rule{thinmathspace}{0ex}}=\phantom{\rule{thinmathspace}{0ex}}\phantom{\rule{thinmathspace}{0ex}}\frac{3k\phantom{\rule{thinmathspace}{0ex}}+\phantom{\rule{thinmathspace}{0ex}}2t\phantom{\rule{thinmathspace}{0ex}}-\phantom{\rule{thinmathspace}{0ex}}4}{4}$ from Corollary 2.6 (2).**(c)**It is obvious that**PF*(*S*) = {*k*− 4,*k*+*t*− 4,*k*+*t*− 2 − 4}. So, we find*PF*(*S*) = {*k*− 4,*k*+*t*− 4,*k*+*t*− 2}.*(d)**We have $n(S)=(k+t-2)+1-\frac{3k+2t-4}{4}=\frac{k+2t}{4}$ from**g*(*S*) =*F*(*S*) + 1 −*n*(*S*).*(e)**We observe that**H*(*S*) = {1, 2, 3, 5, 6, 7, ... , 4*r*− 3, 4*r*− 2, 4*r*− 1, 4*r*− 1, 4*r*+ 3, 4*r*+ 5, ... , 4*r*+*t*} = {1, 2, 3, 5, 6, 7, ... ,*k*− 5,*k*− 4,*k*− 3,*k*− 1,*k*+ 1,*k*+ 3, ... ,*k*+*t*+ 2} since*S*= 〈4,*k, k*+*t, k*+*t*+ 2〉 = {0, 4, 8, ... ,*k*− 6,*k*− 2,*k, k*+ 2, ... ,*k*+*t*− 3,*k*+*t*−*1*, → ... ,} = {0, 4, 8, ... ,4*r*− 4, 4*r*, 4*r*+ 2, 4*r*+ 4, ... , 4*r*−*t*− 1, 4*r*+*t*+*1*, → ... ,}.

**Example 2.9:** *Consider the numerical semigroup S*= 〈4, *k, k* + 2, *k* + 3〉. *If we put k* = 15 *then we have that S* = 〈4, *k, k* + 2, *k* + 3〉 = 〈4, 15, 17, 18〉 = {0, 4, 8, 12, 15, → ... ,} *is saturated. Hence, we find that*

*(a)**F*(*S*) =*k*− 1 = 15 − 1 = 14,*(b)**$g(S)=\frac{3k-1}{4}=\frac{45-1}{4}=11$,**(c)**PF*(*S*) = {*k*− 4,*k*− 2,*k*− 1} = {11, 13, 14},*(d)**$n(S)=\frac{k+1}{4}=\frac{15+1}{4}=4$,**(e)**H*(*S*) = {1, 2, 3, 5, 6, 7, ... ,*k*− 6,*k*− 5,*k*− 4,*k*− 2,*k*− 1} = {1, 2, 3, 5, 6, 7, 9, 10, 11, 13, 14}*and Ap*(*S*, 4) = {0,*k, k*+ 2,*k*+ 3} = {0, 15, 17, 18}.

**Example 2.10:** *Consider the numerical semigroup S*= 〈4, *k, k* + *t, k* + *t* + 2〉. *If we put k* = 14 *and t* = *13 then we find that S*= 〈4, *k, k* + *t, k* + *t* + 2〉 = 〈4, 14, 27, 29〉 = {0, 4, 8, 12, 14, 16, 18, 20, 22, 24, 26, → ... ,} *is saturated. Thus, we observe that*

*(a)**F*(*S*) =*k*+*t*− 2 = 14 + 13 − 2 = 25,*(b)**$g(S)=\frac{3k+2t-4}{4}=\frac{64}{4}=16$,**(c)**PF*(*S*) = {*k*− 4,*k*+*t*− 4,*k*+*t*− 2} = {10, 23, 25},*(d)**$n(S)=\frac{k+2t}{4}=\frac{40}{4}=10$,**(e)**H*(*S*) = {1, 2, 3, 5, 6, 7, ... ,*k*− 5,*k*− 4,*k*+ 3,*k*− 1,*k*− 1,*k*+3, ... ,*k*−*t*− 2} = {1, 2, 3, 5, 6, 7, 9, 10, 11, 13, 15, 17, 19, 21, 23, 25}*and Ap*(*S*, 4) = {0,*k, k*+*t, k*+*t*+ 2} = {0, 14, 27, 29}.

The authors thank the anonymous referee for his/her remarks which helped them to improve the presentation of the paper.

- [1]
Abawajy J., Kelarev A.V., Chowdhury M., Power graphs: a survey, Electronic J. Graph Theory and Applications, 2013, 1(2), 125-147

- [2]
Barucci V, Dobbs D.E., Fontana M., Maximality properties in numerical semigroups and applications to one-dimensional analytically irreducible local domains,Memoirs of American Mathematical Society 1997, 125(598), 1-77

- [3]
Kelarev A.V., Combinatorial properties of sequences in groups and semigroups, "Combinatorics, Complexity and Logic", Discrete Mathematics and Theoretical Computer Science, 1996, 289-298

- [4]
Kelarev A.V., Graph Algebras and Automata, Marcel Dekker, New York, 2003

- [5]
Kelarev A.V., Ryan J., Yearwood J., Cayley graphs as classifiers for data mining:The influence of asymmetries, Discrete Mathematics, 2009, 309(17), 5360-5369

- [6]
Rosales J.C., Garcia-Sánchez P.A., Numerical Semigroups, Springer, 2009

- [7]
Rosales J.C., Garcia-Sánchez P.A., Garcia-Garcia J.I., Jimenez Madrid J.A., Fundamental gaps in numerical semigroups, Journal Pure and Applied Algebra, 2004, 189, 301-313

- [8]
Rosales J.C., Garcia-Sánchez P.A., Garcia-Garcia J.I., Branco M.B., Saturated numerical semigroups, Houston J.of Math. 2004, 30, 321-330

- [9]
Rosales J.C., Numerical Semigroups with Apery sets of unique expression, Journal of Algebra, 2000, 226, 479-487

- [10]
Suer M., Ilhan S., On a family of saturated numerical semigroup with multiplicity four, Turkish Journal of Math., (in press), DOI:10.3906

**Received**: 2016-08-04

**Accepted**: 2016-09-23

**Published Online**: 2016-11-03

**Published in Print**: 2016-01-01

**Citation Information: **Open Mathematics, ISSN (Online) 2391-5455, DOI: https://doi.org/10.1515/math-2016-0074. Export Citation

## Comments (0)