Polynomials appear related to numerical semigroups in several ways. One of them is through their associated generating function (or Hilbert series), and another via value semigroups of a curve; and curves might be defined by polynomials. In this chapter we present several functions to compute the polynomial and Hilbert series associated to a numerical semigroup, and to calculate the respective numerical semigroups given a set of defining polynomials.
Let S be a numerical semigroup. The Hilbert series or generating function associated to S is H_S(x)=∑_s∈ Sx^s (actually it is the Hilbert function of the ring K[S] with K a field). See for instance [Mor14].
‣ NumericalSemigroupPolynomial( s, x ) | ( function ) |
s is a numerical semigroups and x a variable (or a value to evaluate in). The output is the polynomial 1+(x-1)∑_s∈ N∖ S x^s, which equals (1-x)H_S(x).
gap> x:=X(Rationals,"x");; gap> s:=NumericalSemigroup(5,7,9);; gap> NumericalSemigroupPolynomial(s,x); x^14-x^13+x^12-x^11+x^9-x^8+x^7-x^6+x^5-x+1
‣ IsNumericalSemigroupPolynomial( f ) | ( function ) |
f is a polynomial in one variable. The output is true if there exists a numerical semigroup S such that f equals (1-x)H_S(x), that is, the polynomial associated to S (false otherwise).
gap> x:=X(Rationals,"x");; gap> s:=NumericalSemigroup(5,6,7,8);; gap> f:=NumericalSemigroupPolynomial(s,x); x^10-x^9+x^5-x+1 gap> IsNumericalSemigroupPolynomial(f); true
‣ NumericalSemigroupFromNumericalSemigroupPolynomial( f ) | ( function ) |
f is a polynomial associated to a numerical semigroup (otherwise yields error). The output is the numerical semigroup S such that f equals (1-x)H_S(x).
gap> x:=X(Rationals,"x");; gap> s:=NumericalSemigroup(5,6,7,8);; gap> f:=NumericalSemigroupPolynomial(s,x); x^10-x^9+x^5-x+1 gap> NumericalSemigroupFromNumericalSemigroupPolynomial(f)=s; true
‣ HilbertSeriesOfNumericalSemigroup( s, x ) | ( function ) |
s is a numerical semigroup and x a variable (or a value to evaluate in). The output is the series ∑_s∈ S x^s. The series is given as a rational function.
gap> x:=X(Rationals,"x");; gap> s:=NumericalSemigroup(5,7,9);; gap> HilbertSeriesOfNumericalSemigroup(s,x); (x^14-x^13+x^12-x^11+x^9-x^8+x^7-x^6+x^5-x+1)/(-x+1)
‣ GraeffePolynomial( p ) | ( function ) |
p is a polynomial. Computes the Graeffe polynomial of p. Needed to test if p is a cyclotomic polynomial (see [BD89]).
gap> x:=Indeterminate(Rationals,1);; SetName(x,"x"); gap> GraeffePolynomial(x^2-1); x^2-2*x+1
‣ IsCyclotomicPolynomial( p ) | ( function ) |
p is a polynomial. Detects if p is a cyclotomic polynomial using the procedure given in [BD89].
gap> CyclotomicPolynomial(Rationals,3); x^2+x+1 gap> IsCyclotomicPolynomial(last); true
‣ IsKroneckerPolynomial( p ) | ( function ) |
p is a polynomial. Detects if p is a Kronecker polynomial, that is, a monic polynomial with integer coefficients having all its roots in the unit circumference, or equivalently, a product of cyclotomic polynomials. The current implementation has been done with A. Herrera-Poyatos, following [BD89].
gap> x:=X(Rationals,"x");; gap> s:=NumericalSemigroup(3,5,7);; gap> t:=NumericalSemigroup(4,6,9);; gap> p:=NumericalSemigroupPolynomial(s,x); x^5-x^4+x^3-x+1 gap> q:=NumericalSemigroupPolynomial(t,x); x^12-x^11+x^8-x^7+x^6-x^5+x^4-x+1 gap> IsKroneckerPolynomial(p); false gap> IsKroneckerPolynomial(q); true
‣ IsCyclotomicNumericalSemigroup( s ) | ( function ) |
s is a numerical semigroup. Detects if the polynomial associated to s is a Kronecker polynomial.
gap> l:=CompleteIntersectionNumericalSemigroupsWithFrobeniusNumber(21);; gap> ForAll(l,IsCyclotomicNumericalSemigroup); true
‣ CyclotomicExponentSequence( s, k ) | ( operation ) |
s is a numerical semigroup and k is a positive integer. Outputs the list of the first k elements of the cyclotomic exponent sequence of s (see [CGM16]).
The sequence will be truncated if the semigroup is cyclotomic and k is bigger than the last nonzero element in its sequence.
gap> s:=NumericalSemigroup(3,4);; gap> CyclotomicExponentSequence(s,20); [ 1, 0, -1, -1, 0, 0, 0, 0, 0, 0, 0, 1 ] gap> s:=NumericalSemigroup(3,5,7);; gap> CyclotomicExponentSequence(s,20); [ 1, 0, -1, 0, -1, 0, -1, 0, 0, 1, 0, 1, 0, 1, 0, 0, -1, 0, -1, 0 ]
‣ WittCoefficients( p, k ) | ( operation ) |
p is a univariate polynomial with integer coefficientas and p(1)=1. Then p(x)=∏_n≥ 0(1-x^n)^e_n, for some integers e_n. The output is the list [e_1,...,e_k], and it is computed by using [CGHM19].
gap> s:=NumericalSemigroup(3,4);; gap> x:=Indeterminate(Rationals,"x");; gap> p:=NumericalSemigroupPolynomial(s,x);; gap> WittCoefficients(p,20); [ 1, 0, -1, -1, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0 ]
The difference with this example and the one in CyclotomicExponentSequence (10.1-9) is that in that case a cyclotomic check is performed that slows down the process.
‣ IsSelfReciprocalUnivariatePolynomial( p ) | ( function ) |
p is a univariate polynomial. Detects if p is selfreciprocal. A numerical semigroup is symmetric if and only if it is selfreciprocal, [Mor14]. The current implementation is due to A. Herrera-Poyatos.
gap> l:=IrreducibleNumericalSemigroupsWithFrobeniusNumber(13);; gap> x:=X(Rationals,"x");; gap> ForAll(l, s-> > IsSelfReciprocalUnivariatePolynomial(NumericalSemigroupPolynomial(s,x))); true
Let f(x,y)∈ K[x,y], with K an algebraically closed field of characteristic zero. Let f(x,y)=y^n+a_1(x)y^n-1+dots+a_n(x) be a nonzero polynomial of K[x][y]. After possibly a change of variables, we may assume that, that deg_x(a_i(x))≤ i-1 for all i∈{1,..., n}. For g∈ K[x,y] that is not a multiple of f, define mathrmint(f,g)=dim_ K frac K[x,y](f,g). If f has one place at infinity, then the set {mathrmint(f,g)∣ g∈ K[x,y]∖(f)} is a free numerical semigroup (and thus a complete intersection).
‣ SemigroupOfValuesOfPlaneCurveWithSinglePlaceAtInfinity( f ) | ( function ) |
f is a polynomial in the variables X(Rationals,1) and X(Rationals,2). Computes the semigroup {mathrmint(f,g)∣ g∈ K[x,y]∖(f)}, where mathrmint(f,g)=dim_ K ( K[x,y]/(f,g)). The algorithm checks if f has one place at infinity. If the extra argument "all" is given, then the output is the δ-sequence and approximate roots of f. The method is explained in [AG16a].
gap> x:=Indeterminate(Rationals,1);; SetName(x,"x"); gap> y:=Indeterminate(Rationals,2);; SetName(y,"y"); gap> f:=((y^3-x^2)^2-x*y^2)^4-(y^3-x^2);; gap> SemigroupOfValuesOfPlaneCurveWithSinglePlaceAtInfinity(f,"all"); [ [ 24, 16, 28, 7 ], [ y, y^3-x^2, y^6-2*x^2*y^3+x^4-x*y^2 ] ]
‣ IsDeltaSequence( l ) | ( function ) |
l is a list of positive integers. Assume that l equals a_0,a_1,dots,a_h. Then l is a δ-sequence if gcd(a_0,..., a_h)=1, ⟨ a_0,⋯, a_s⟩ is free, a_kD_k > a_k+1D_k+1 and a_0> a_1 > D_2 > D_3 > ... > D_h+1, where D_1=a_0, D_k=gcd(D_k-1,a_k-1).
Every δ-sequence generates a numerical semigroup that is the semigroup of values of a plane curve with one place at infinity.
gap> IsDeltaSequence([24,16,28,7]); true
‣ DeltaSequencesWithFrobeniusNumber( f ) | ( function ) |
f is an integer. Computes the set of all δ-sequences generating numerical semigroups with Frobenius number f.
gap> DeltaSequencesWithFrobeniusNumber(21); [ [ 8, 6, 11 ], [ 10, 4, 15 ], [ 12, 8, 6, 11 ], [ 14, 4, 11 ], [ 15, 10, 4 ], [ 23, 2 ] ]
‣ CurveAssociatedToDeltaSequence( l ) | ( function ) |
l is a δ-sequence. Computes a curve in the variables X(Rationals,1) and X(Rationals,2) whose semigroup of values is generated by the l.
gap> CurveAssociatedToDeltaSequence([24,16,28,7]); y^24-8*x^2*y^21+28*x^4*y^18-56*x^6*y^15-4*x*y^20+70*x^8*y^12+24*x^3*y^17-56*x^\ 10*y^9-60*x^5*y^14+28*x^12*y^6+80*x^7*y^11+6*x^2*y^16-8*x^14*y^3-60*x^9*y^8-24\ *x^4*y^13+x^16+24*x^11*y^5+36*x^6*y^10-4*x^13*y^2-24*x^8*y^7-4*x^3*y^12+6*x^10\ *y^4+8*x^5*y^9-4*x^7*y^6+x^4*y^8-y^3+x^2 gap> SemigroupOfValuesOfPlaneCurveWithSinglePlaceAtInfinity(last,"all"); [ [ 24, 16, 28, 7 ], [ y, y^3-x^2, y^6-2*x^2*y^3+x^4-x*y^2 ] ]
‣ SemigroupOfValuesOfPlaneCurve( f ) | ( function ) |
f is a polynomial in the variables X(Rationals,1) and X(Rationals,2). The singular package is mandatory. Either by loading it prior to numerical semigroups or by using NumSgpsUseSingular(). If f is irreducible, computes the semigroup {mathrmint(f,g)∣ g∈ K[x,y]∖(f)}, where mathrmint(f,g)=dim_ K ( K[[x,y]]/(f,g)). If it has two components, the output is the value semigroup in two variables, and thus a good semigroup. If there are more components, then the output is that of semigroup in the alexpoly singular library.
gap> x:=X(Rationals,"x");; y:=X(Rationals,"y");; gap> f:= y^4-2*x^3*y^2-4*x^5*y+x^6-x^7; -x^7+x^6-4*x^5*y-2*x^3*y^2+y^4 gap> SemigroupOfValuesOfPlaneCurve(f); <Numerical semigroup with 3 generators> gap> MinimalGenerators(last); [ 4, 6, 13 ] gap> f:=(y^4-2*x^3*y^2-4*x^5*y+x^6-x^7)*(y^2-x^3);; gap> SemigroupOfValuesOfPlaneCurve(f); <Good semigroup> gap> MinimalGenerators(last); [ [ 4, 2 ], [ 6, 3 ], [ 13, 15 ], [ 29, 13 ] ]
‣ SemigroupOfValuesOfCurve_Local( arg ) | ( function ) |
The function admits one or two parameters. In any case, the first is a list of polynomials pols. And the second can be the string "basis" or an integer val.
If only one argument is given, the output is the semigroup of all possible orders of K[[pols]] provided that K[[x]]/K[[pols]] has finite length. If the second argument "basis" is given, then the output is a (reduced) basis of the algebra K[[pols]] such that the orders of the basis elements generate minimally the semigroup of orders of K[[pols]]. If an integer val is the second argument, then the output is a polynomial in K[[pols]] with order val (fail if there is no such polynomial, that is, val is not in the semigroup of values).
The method is explained in [AGM17].
gap> x:=Indeterminate(Rationals,"x");; gap> SemigroupOfValuesOfCurve_Local([x^4,x^6+x^7,x^13]); <Numerical semigroup with 4 generators> gap> MinimalGeneratingSystem(last); [ 4, 6, 13, 15 ] gap> SemigroupOfValuesOfCurve_Local([x^4,x^6+x^7,x^13], "basis"); [ x^4, x^7+x^6, x^13, x^15 ] gap> SemigroupOfValuesOfCurve_Local([x^4,x^6+x^7,x^13], 20); x^20
‣ SemigroupOfValuesOfCurve_Global( arg ) | ( function ) |
The function admits one or two parameters. In any case, the first is a list of polynomials pols. And the second can be the string "basis" or an integer val.
If only one argument is given, the output is the semigroup of all possible degrees of K[pols] provided that K[x]/K[pols] has finite length. If the second argument "basis" is given, then the output is a (reduced) basis of the algebra K[pols] such that the degrees of the basis elements generate minimally the semigroup of degrees of K[pols]. If an integer val is the second argument, then the output is a polynomial in K[pols] with degree val (fail if there is no such polynomial, that is, val is not in the semigroup of values).
The method is explained in [AGM17].
gap> x:=Indeterminate(Rationals,"x");; gap> SemigroupOfValuesOfCurve_Global([x^4,x^6+x^7,x^13]); <Numerical semigroup with 3 generators> gap> MinimalGeneratingSystem(last); [ 4, 7, 13 ] gap> SemigroupOfValuesOfCurve_Global([x^4,x^6+x^7,x^13],"basis"); [ x^4, x^7+x^6, x^13 ] gap> SemigroupOfValuesOfCurve_Global([x^4,x^6+x^7,x^13],12); x^12 gap> SemigroupOfValuesOfCurve_Global([x^4,x^6+x^7,x^13],6); fail
‣ GeneratorsModule_Global( A, M ) | ( function ) |
A and M are lists of polynomials in the same variable. The output is a basis of the ideal M K[A], that is, a set F such that deg(F) generates the ideal deg(M K[A]) of deg(K[A]), where deg stands for degree. The method is explained in [AAG19].
gap> t:=Indeterminate(Rationals,"t");; gap> A:=[t^6+t,t^4];; gap> M:=[t^3,t^4];; gap> GeneratorsModule_Global(A,M); [ t^3, t^4, t^5, t^6 ]
‣ GeneratorsKahlerDifferentials( A ) | ( function ) |
A is a list of polynomials in the same variable. The output is GeneratorsModule_Global(A,M), with M the set of derivatives of the elements in A.
gap> t:=Indeterminate(Rationals,"t");; gap> GeneratorsKahlerDifferentials([t^3,t^4]); [ t^2, t^3 ]
‣ IsMonomialNumericalSemigroup( S ) | ( property ) |
S is a numerical semigroup. Tests whether S a monomial numerical semigroup.
Let R a Noetherian ring such that K ⊆ R ⊆ K[[t]], K is a field of characteristic zero, the algebraic closure of R is K[[t]], and the conductor (R : K[[t]]) is not zero. If v : K((t))-> Z is the natural valuation for K((t)), then v(R) is a numerical semigroup.
Let S be a numerical semigroup minimally generated by {n_1,...,n_e}. The semigroup ring associated to S is K[[S]]=K[[t^n_1,...,t^n_e]]. A ring is called a semigroup ring if it is of the form K[[S]], for some numerical semigroup S. We say that S is a monomial numerical semigroup if for any R as above with v(R)=S, R is a semigroup ring. See [Mic02] for details.
gap> IsMonomialNumericalSemigroup(NumericalSemigroup(4,6,7)); true gap> IsMonomialNumericalSemigroup(NumericalSemigroup(4,6,11)); false
‣ LegendrianGenericNumericalSemigroup( n, m ) | ( function ) |
n and m are coprime integers with m≥ 2n+1. The output is the semigroup of a generic element in the class of irreducible Legendrian singularities with equisingularity equal to the topological type of y^n=x^m, as described in Section 5 of [AN09].
gap> s:=LegendrianGenericNumericalSemigroup(5,11);; gap> SmallElements(s); [ 0, 5, 6, 10, 11, 12, 13, 15 ]
generated by GAPDoc2HTML