Documentation

Hironaka.Resolution.Algebraic.Tuning.Sheaf

The maximal coefficient ideal of an ideal sheaf #

Kollár's maximal coefficient ideals [Kol07, Definition 98]: for a smooth variety X, an ideal sheaf I ⊂ 𝒪_X and m = max-ord I, the maximal coefficient ideal of order s of I is W_s(I) := ∑_{c : ∑ (m − j) c_j ≥ s} ∏_{j=0}^m (D^j I)^{c_j}, where D^j I is the j-th derivative ideal sheaf. The same definition on a regular local ring with coordinates is IsLocalRing.RegularCoords.W, with the weight IsLocalRing.wt m c = ∑ (m − j) c_j; Theorem 100 and Corollary 101 (Hironaka.Resolution.Algebraic.Tuning.Assemble, Hironaka.Resolution.Algebraic.Tuning.Corollary101) speak of ideal sheaves, so the definition is given here for ideal sheaves, with the parameter m explicit as in the maximal-contact ideal MC f I m (Kollár's m = max-ord I is a hypothesis where the theorems need it).

Definition 98 for ideal sheaves #

noncomputable def AlgebraicGeometry.Scheme.IdealSheafData.W {X : Scheme} {k : Type u} [Field k] (f : X ⟶ Spec ↧k) (I : X.IdealSheafData) (m s : ℕ) :

The maximal coefficient ideal of order s of the ideal sheaf I [Kol07, Definition 98], with the parameter m (Kollár's max-ord I) explicit: W_s(I) = ∑_{wt_m(e) ≥ s} ∏_{j ≤ m} (D^j I)^{e_j}, the sum over all exponent vectors e : {0, …, m} → ℕ of weight wt_m(e) = ∑ (m − j) e_j ≥ s, where D^j I is the j-th derivative ideal sheaf derivativeIter f j I.

Equations
  • One or more equations did not get rendered due to their size.
Instances For
    theorem AlgebraicGeometry.Scheme.IdealSheafData.W_eq {X : Scheme} {k : Type u} [Field k] (f : X ⟶ Spec ↧k) (I : X.IdealSheafData) (m s : ℕ) :
    W f I m s = ⨆ (e : Fin (m + 1) → ℕ), ⨆ (_ : s ≤ IsLocalRing.wt m e), ∏ j : Fin (m + 1), derivativeIter f (↑j) I ^ e j

    Definition 98 unfolded (definitional).

    theorem AlgebraicGeometry.Scheme.IdealSheafData.prod_le_W {X : Scheme} {k : Type u} [Field k] (f : X ⟶ Spec ↧k) {I : X.IdealSheafData} {m s : ℕ} {e : Fin (m + 1) → ℕ} (he : s ≤ IsLocalRing.wt m e) :
    ∏ j : Fin (m + 1), derivativeIter f (↑j) I ^ e j ≤ W f I m s

    Every generating product of weight ≥ s lies in W_s(I) [Kol07, Definition 98].

    theorem AlgebraicGeometry.Scheme.IdealSheafData.W_le_iff {X : Scheme} {k : Type u} [Field k] (f : X ⟶ Spec ↧k) {I J : X.IdealSheafData} {m s : ℕ} :
    W f I m s ≤ J ↔ ∀ (e : Fin (m + 1) → ℕ), s ≤ IsLocalRing.wt m e → ∏ j : Fin (m + 1), derivativeIter f (↑j) I ^ e j ≤ J

    W_s(I) ⊆ J iff every generating product of weight ≥ s lies in J [Kol07, Definition 98].

    @[simp]
    theorem AlgebraicGeometry.Scheme.IdealSheafData.W_zero {X : Scheme} {k : Type u} [Field k] (f : X ⟶ Spec ↧k) (I : X.IdealSheafData) (m : ℕ) :
    W f I m 0 = ⊤

    W_0(I) = 𝒪_X: the zero exponent vector has weight 0.

    theorem AlgebraicGeometry.Scheme.IdealSheafData.pow_le_W {X : Scheme} {k : Type u} [Field k] (f : X ⟶ Spec ↧k) {I : X.IdealSheafData} {m s c : ℕ} (h : s ≤ m * c) :
    I ^ c ≤ W f I m s

    I^c ⊆ W_s(I) whenever mc ≥ s: the exponent vector e_0 = c has weight mc. This is the inclusion I^{s'} ⊆ W_{ms'}(I) through which the proof of [Kol07, Corollary 101] invokes Theorem 100, and it is used again in the proof of [Kol07, Theorem 103].

    theorem AlgebraicGeometry.Scheme.IdealSheafData.MC_pow_le_W {X : Scheme} {k : Type u} [Field k] (f : X ⟶ Spec ↧k) {m : ℕ} (hm : 1 ≤ m) (s : ℕ) (I : X.IdealSheafData) :
    MC f I m ^ s ≤ W f I m s

    MC(I)^s ⊆ W_s(I) for m ≥ 1 [Kol07, Proposition 99, proof]: the exponent vector e_{m−1} = s has weight s.

    theorem AlgebraicGeometry.Scheme.IdealSheafData.W_one {X : Scheme} {k : Type u} [Field k] (f : X ⟶ Spec ↧k) {m : ℕ} (hm : 1 ≤ m) (I : X.IdealSheafData) :
    W f I m 1 = MC f I m

    W_1(I) = MC(I) for m ≥ 1 Kol07, Proposition 99 (4), proof: "W_1(I) ⊂ ∑_{j<m} D^j(I) = D^{m−1}(I)": ⊇ by MC_pow_le_W; ⊆ because a generating product of weight ≥ 1 has some e_j ≥ 1 with j < m, and so lies in D^j I ⊆ D^{m−1} I.

    theorem AlgebraicGeometry.Scheme.IdealSheafData.W_anti {X : Scheme} {k : Type u} [Field k] (f : X ⟶ Spec ↧k) {I : X.IdealSheafData} {m s t : ℕ} (h : s ≤ t) :
    W f I m t ≤ W f I m s

    W_t(I) ⊆ W_s(I) for s ≤ t [Kol07, Proposition 99 (1)].

    The stalk bridge #

    theorem AlgebraicGeometry.Scheme.IdealSheafData.stalkIdeal_W {X : Scheme} {k : Type u} [Field k] (f : X ⟶ Spec ↧k) [CharZero k] (n : ℕ) [SmoothOfRelativeDimension n f] (I : X.IdealSheafData) (m s : ℕ) (p : ↥X) {n' : ℕ} (c : IsLocalRing.RegularCoords (↑(X.presheaf.stalk p)) n') (hk : c.IsLinearOver k) (hs : c.SpansDerivations k) :
    (W f I m s).stalkIdeal p = c.W m s (I.stalkIdeal p)

    The stalk of the maximal coefficient ideal: for coordinates c on 𝒪_{X,p} that are k-linear and span the k-derivations, (W_s(I))_p = W_{c,s}(I_p), the maximal coefficient ideal of the stalk computed in coordinates (IsLocalRing.RegularCoords.W). The stalk of a sum of products of powers is the sum of the products of the powers of the stalks, and (D^j I)_p = D_c^j(I_p) (stalkIdeal_derivativeIter, derivativeIter_eq_Dpow).