Documentation

LeanPool.CCDN.Hilbert.FilteredHilbertPolynomial

Existence, uniqueness and positive integral multiplicity of the filtered Hilbert polynomial of a polynomial quotient, in arbitrary finite dimension. These results construct the polynomial from standard exponents.

def CCDN.standardExponentVectors {K : Type u_1} {σ : Type u_2} [Field K] [Fintype σ] (I : Ideal (MvPolynomial σ K)) (m : MonomialOrder σ) :
Set (σ → ℕ)

The exponent vectors, as functions σ → ℕ, of the standard monomials of I with respect to the monomial order m: those that are not leading monomials of elements of I.

Equations
Instances For

    The standard exponent vectors of an ideal form a lower set.

    A proper ideal has a standard exponent vector for the degree-lexicographic order.

    The affine Hilbert function of I at N equals the number of standard exponent vectors of I, for the degree-lexicographic order, of total degree at most N.

    theorem CCDN.filteredCoordinateSpace_eventually_polynomial {K : Type u_1} {σ : Type u_2} [Field K] [Finite σ] [LinearOrder σ] (I : Ideal (MvPolynomial σ K)) :
    ∃ (P : Polynomial ℚ) (N : ℕ), ∀ n ≥ N, Polynomial.eval (↑n) P = ↑(Module.finrank K ↥(filteredCoordinateSpace I n))

    Eventual polynomiality of the filtered coordinate-space dimension.

    theorem CCDN.polynomial_eq_of_eventual_nat_eval_eq (P Q : Polynomial ℚ) (h : ∃ (N : ℕ), ∀ n ≥ N, Polynomial.eval (↑n) P = Polynomial.eval (↑n) Q) :
    P = Q

    Two rational polynomials that agree at all sufficiently large natural numbers are equal.

    noncomputable def CCDN.filteredHilbertPolynomial {K : Type u_1} {σ : Type u_2} [Field K] [Fintype σ] [LinearOrder σ] (I : Ideal (MvPolynomial σ K)) :

    The uniquely determined filtered Hilbert polynomial.

    Equations
    Instances For
      structure CCDN.HasAffineDimensionDegree {K : Type u_1} {σ : Type u_2} [Field K] [Fintype σ] [LinearOrder σ] (I : Ideal (MvPolynomial σ K)) (m d : ℕ) :

      The filtered Hilbert polynomial of I has degree m and leading coefficient d / m!: the affine variety cut out by I has dimension m and degree d.

      Instances For
        @[reducible, inline]
        abbrev CCDN.HasProjectiveDimensionDegree {K : Type u_1} {σ : Type u_2} [Field K] [Fintype σ] [LinearOrder σ] (I : Ideal (MvPolynomial σ K)) (m d : ℕ) :

        The cone over I has filtered Hilbert polynomial of degree m + 1 and leading coefficient d / (m + 1)!. For a homogeneous ideal I this says that the projective variety cut out by I has dimension m and degree d; it is the affine dimension and degree of the cone.

        Equations
        Instances For

          Dimension and degree only depend on the filtered Hilbert polynomial.

          theorem CCDN.filteredHilbertPolynomial_spec {K : Type u_1} {σ : Type u_2} [Field K] [Fintype σ] [LinearOrder σ] (I : Ideal (MvPolynomial σ K)) :
          ∃ (N : ℕ), ∀ n ≥ N, Polynomial.eval (↑n) (filteredHilbertPolynomial I) = ↑(Module.finrank K ↥(filteredCoordinateSpace I n))

          The filtered Hilbert polynomial of I agrees at every sufficiently large n with the dimension of the image in K[X] ⧸ I of the polynomials of total degree at most n.

          theorem CCDN.filteredHilbertPolynomial_eq_of_eventual {K : Type u_1} {σ : Type u_2} [Field K] [Fintype σ] [LinearOrder σ] (I : Ideal (MvPolynomial σ K)) (P : Polynomial ℚ) (hP : ∃ (N : ℕ), ∀ n ≥ N, Polynomial.eval (↑n) P = ↑(Module.finrank K ↥(filteredCoordinateSpace I n))) :

          Any rational polynomial that eventually agrees with the affine Hilbert function of I is the filtered Hilbert polynomial of I.

          Proper polynomial quotients have a positive integer factorial-normalized leading coefficient.

          theorem CCDN.exists_hasAffineDimensionDegree {K : Type u_1} {σ : Type u_2} [Field K] [Fintype σ] [LinearOrder σ] (I : Ideal (MvPolynomial σ K)) (hI : I ≠ ⊤) {m : ℕ} (hdim : (filteredHilbertPolynomial I).natDegree = m) :
          ∃ (d : ℕ), 0 < d ∧ HasAffineDimensionDegree I m d

          A proper ideal whose filtered Hilbert polynomial has degree m has dimension m and some positive degree.