Mixed indices, disjoint factors, sampling grids and minima
DefinitionPhilipponMultiplicity_CorollariesActual mixed-codimension indices, factor projections, the disjoint-factor condition, real-bound integer sampling grids, quotient ranks, and infima over obstruction subgroups. Empty infima are zero; the source audit records the trivial analytic-subgroup boundary.
Compiled, admission-free definition bundle. Theorems asserting its substantive properties remain open targets in the full-paper goal.
import Definitions.Def_PhilipponMultiplicity_Analytic
import Definitions.Def_PhilipponMultiplicity_Degree
set_option autoImplicit false
open scoped BigOperators
noncomputable section
namespace PhilipponMultiplicity
universe u
variable {K : Type u} [NontriviallyNormedField K]
structure SourceMixedCodimensionIndex (G : EmbeddedGroupProduct K)
(H : AlgebraicSubgroup G) where
exponent : G.FactorIndex → ℕ
bounded : ∀ i, exponent i ≤ (G.factor i).dimension
sum_eq : (∑ i, exponent i) + varietyDimension G H.carrier = G.dimension
def SourceMixedCodimensionIndex.complementIndex {G : EmbeddedGroupProduct K}
{H : AlgebraicSubgroup G} (r : SourceMixedCodimensionIndex G H) : G.FactorIndex → ℕ :=
fun i => (G.factor i).dimension - r.exponent i
def SourceMixedCodimensionIndex.degreeMonomial {G : EmbeddedGroupProduct K}
{H : AlgebraicSubgroup G} (r : SourceMixedCodimensionIndex G H)
(D : G.FactorIndex → ℕ) : ℝ :=
∏ i, (D i : ℝ) ^ r.exponent i
def factorProjection (G : EmbeddedGroupProduct K) (H : AlgebraicSubgroup G)
(i : G.FactorIndex) : Set (G.factor i).Point :=
(fun x : G.Point => x i) '' H.carrier
def factorTopology (E : EmbeddedCommutativeGroup K) : TopologicalSpace E.Point :=
TopologicalSpace.induced (fun x _ => x.val) (projectiveSpace K E.ambientDimension).zariskiTopology
/-- Every algebraic subgroup is a product of algebraic subgroups of the factors. -/
def HasDisjointFactors (G : EmbeddedGroupProduct K) : Prop :=
∀ H : AlgebraicSubgroup G,
(∀ x : G.Point, x ∈ H.carrier ↔ ∀ i, x i ∈ factorProjection G H i) ∧
(∀ i, @IsClosed _ (factorTopology (G.factor i)) (factorProjection G H i))
def factorCodimension (G : EmbeddedGroupProduct K) (H : AlgebraicSubgroup G)
(i : G.FactorIndex) : ℕ :=
(G.factor i).dimension -
(Hilbert.hilbertPolynomial K 1 (fun _ => (G.factor i).ambientDimension)
((projectiveSpace K (G.factor i).ambientDimension).vanishingIdeal
((fun x : (G.factor i).Point => fun _ => x.val) '' factorProjection G H i))).totalDegree
def samplingGrid {G : EmbeddedGroupProduct K} {l : ℕ}
(γ : Fin l → G.Point) (S : ℝ) : Set G.Point :=
{x | ∃ a : Fin l → ℕ, (∀ i, (a i : ℝ) ≤ S) ∧ x = ∑ i, a i • γ i}
def samplingQuotientRank {G : EmbeddedGroupProduct K} {l : ℕ}
(γ : Fin l → G.Point) (H : AlgebraicSubgroup G) : ℕ :=
Module.finrank ℤ (Submodule.span ℤ
(Set.range (fun i => QuotientAddGroup.mk' H.toAddSubgroup (γ i))))
def IsCorollary23Obstruction {G : EmbeddedGroupProduct K} (A : AnalyticSubgroup G)
(r : G.FactorIndex → ℕ) (H : AlgebraicSubgroup G) : Prop :=
(¬ A.carrier ⊆ H.carrier) ∧ ∀ i, r i ≤ factorCodimension G H i
/-- The source minimum, using the natural-number infimum (zero on an empty
family). Nonempty-family and zero-dimensional-image cases need separate audit. -/
def analyticCodimensionMinimum {G : EmbeddedGroupProduct K} (A : AnalyticSubgroup G)
(r : G.FactorIndex → ℕ) : ℕ :=
sInf {k | ∃ H : AlgebraicSubgroup G,
IsCorollary23Obstruction A r H ∧ k = analyticCodimension A H.carrier}
def samplingRankMinimum {G : EmbeddedGroupProduct K} (A : AnalyticSubgroup G)
{l : ℕ} (γ : Fin l → G.Point) (r : G.FactorIndex → ℕ) : ℕ :=
sInf {k | ∃ H : AlgebraicSubgroup G,
IsCorollary23Obstruction A r H ∧ k = samplingQuotientRank γ H}
end PhilipponMultiplicity
Read-back
What the Lean code literally says, in plain math · gpt-6
This bundle assumes any nontrivially normed field , with no additional isometric-isomorphism hypothesis. It uses the following product, analytic, and numerical data. An embedded product consists of a positive number of factors ; each factor is a locally closed subset of , with , carrying an abelian additive group structure. Its addition and negation are required to have local projective polynomial representations: at every source point and for each target block, some open neighborhood in the source projective Zariski topology and some tuple of homogeneous polynomials of a common block multidegree represent that block of the map at every source point in the neighborhood, and the evaluated tuple is nonzero there. The points of are the Cartesian product of the factor carriers, and is its coordinate polynomial ring. In each projective block a nonzero representative is chosen; denotes evaluation at these chosen representatives. A polynomial is multihomogeneous of degree when every monomial in its support has total exponent in block ; the zero polynomial has every such degree. The ambient Zariski topology is generated by the sets where a multihomogeneous polynomial is nonzero, and has the induced topology. For any , is the ideal generated by all multihomogeneous polynomials vanishing at every point of ; write for the whole group and . An algebraic subgroup means an additive subgroup closed in this topology; connectedness refers to this topology. Translation means . The data consist of a natural number , a real radius , the open ball with its asserted openness and membership , and a map such that and whenever . No injectivity is required. There are coordinate lifts , defined for every and , each analytic at ; for each coordinate admits a formal multilinear power series on the entire radius- ball and each block of is nonzero and represents for every . For every , on some neighborhood of contained in , every block of is nonzero and represents . The set is the additive subgroup generated by , with no topological closure operation. Put and . The natural number is , using natural-number subtraction, and is . The order is the least for which the -fold Fréchet derivative of at is nonzero, or if all these derivatives vanish; the derivative of order zero is the value at . Derivatives use the total Fréchet derivative, which is zero where differentiability fails. All order comparisons take place in . For any ideal and block degree , let be the -dimension of the image of the vector space of multihomogeneous polynomials of degree in . Let be a polynomial agreeing with at every natural vector above some coordinatewise natural threshold, if such a polynomial exists; if none exists, set . The polynomial, when it exists, is unique. Set , with total degree of the zero polynomial defined as , and , where brackets select the total homogeneous part of that degree. This is a rational value, viewed as a real number in the group inequalities. Write and . The number assigned to a factor is the total degree of the corresponding polynomial for its vanishing ideal in its one-block homogeneous coordinate ring, and . None of these definitions supplies a separate existence hypothesis for ; the zero fallback makes both and equal to zero in that case. Evaluation permits zero coordinates of , with . For an algebraic subgroup , a mixed codimension index consists of a natural vector with for every factor and . Its complement vector is using natural subtraction, and its degree monomial at is . The index type may be empty; in particular its sum condition cannot be met if . The coordinate projection of is its set-theoretic image in a factor, and the factor topology is the topology induced by its one-factor projective embedding. The condition on the factors means that every algebraic subgroup , with no connectedness restriction, is exactly the product of its coordinate images , and each is closed in the topology induced on from its one-factor projective Zariski topology. For any algebraic subgroup, put , where is the total degree of the eventual Hilbert polynomial of the homogeneous vanishing ideal of the image in that one-factor projective space, using zero as the polynomial if none exists; the subtraction is truncated at zero. For every , family , and real bound , the sampling grid is the set of with natural coefficients satisfying as real numbers. The sampling quotient rank at an algebraic subgroup is the natural-valued -dimension of the -submodule spanned by the images of the in the additive quotient . An obstruction for means precisely and for every . For a natural vector , consider all algebraic subgroups such that and for every . Connectedness is not part of this condition. Let be the minimum of over this family, and let be the minimum, separately over the same family, of the -rank of the subgroup generated by the images of in . Each minimum is defined to be when the family is empty; the two minima need not be attained by the same . The rank used is the natural-valued dimension of the -span of those quotient images, namely the supremum of cardinalities of -linearly independent subsets converted to a natural number, with value zero if that supremum is infinite; torsion contributes no rank, and in this finitely generated span the rank is finite. The general grid definition does not assume : when and it is empty, whereas for it is for every real . For the grid is , and for nonnegative it always contains . Repeated or zero sampling points are allowed and no freeness, independence, or torsion-freeness hypothesis is present. For every sampling quotient rank is zero. If , there are no obstruction subgroups because every subgroup contains zero, so both minima are zero for every . Powers use ; all factor dimensions and codimensions can be zero. The factor condition concerns every closed subgroup, and the obstruction family also contains potentially disconnected subgroups.