Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Polynomial translation operators and local component retention

Definition
PhilipponMultiplicity_Operators

by tomasz · Sep 23, 2026 · Mathlib 0df444a (Lean v4.33.1)

draft-definitionphilippon-multiplicity

Actual polynomial translation charts with analytic coefficients, substitution and coefficient-wise iterated differentiation, generated ideals, and intersections of contractions from localization at all homogeneous representatives of group points. The comparison with source primary-component retention and atlas independence remain separate theorem obligations.

Compiled, admission-free definition bundle. Theorems asserting its substantive properties remain open targets in the full-paper goal.

Definition code
import Definitions.Def_PhilipponMultiplicity_Geometry
import Definitions.Def_PhilipponMultiplicity_Analytic

/-!
# Philippon's polynomial translation/differentiation operators

The chart coordinates are actual finite polynomials in the projective
coordinates, with actual analytic functions as coefficients. The operators
substitute those polynomials, differentiate their coefficients, and reassemble
the resulting polynomial. No ideal or degree estimate is supplied as data.

The group-retention operation is a literal intersection of contractions of
localized ideals at all homogeneous representatives of group points. Its
comparison with the primary-component presentation in Definition 4.2 is a
separate theorem. It preserves local ideal membership, not just zero loci.
-/

set_option autoImplicit false

open scoped BigOperators Topology
open Filter
noncomputable section

namespace PhilipponMultiplicity

universe u

variable {K : Type u} [NontriviallyNormedField K]

/-- The coefficient ring consists of functions of the analytic parameters. -/
abbrev AnalyticCoefficientRing {G : EmbeddedGroupProduct K} (A : AnalyticSubgroup G) :=
  A.ParameterSpace → K

/-- Evaluate a polynomial with parameter-function coefficients at an actual
group point and a parameter value. -/
def evaluateCoefficientPolynomial {G : EmbeddedGroupProduct K} (A : AnalyticSubgroup G)
    (P : MvPolynomial G.ambient.Variable (AnalyticCoefficientRing A))
    (x : G.Point) (z : A.ParameterSpace) : K :=
  MvPolynomial.eval₂Hom (Pi.evalRingHom (fun _ : A.ParameterSpace => K) z)
    (G.ambient.coordinate (G.embedding x)) P

/-- One Zariski chart for translating by `g` and the local analytic subgroup.
The parameter neighborhood in the representation may depend on `x`.
This is the germwise representation needed to define the source operators;
any stronger uniform product-neighborhood presentation supplies this data. -/
structure TranslationChart {G : EmbeddedGroupProduct K} (A : AnalyticSubgroup G)
    (g : G.Point) where
  domain : Set G.Point
  domain_isOpen : @IsOpen _ G.zariskiTopology domain
  degree : G.FactorIndex → ℕ
  coordinates : G.ambient.Variable →
    MvPolynomial G.ambient.Variable (AnalyticCoefficientRing A)
  coefficient_analytic : ∀ v, ∀ e ∈ (coordinates v).support,
    AnalyticAt K ((coordinates v).coeff e) 0
  coordinate_homogeneous : ∀ v, ∀ e ∈ (coordinates v).support,
    (∑ j : Fin (G.ambient.ambientDimension v.1 + 1), e ⟨v.1, j⟩) = degree v.1 ∧
      ∀ i : G.FactorIndex, i ≠ v.1 →
        (∑ j : Fin (G.ambient.ambientDimension i + 1), e ⟨i, j⟩) = 0
  /- Rational translation formulas remain projectively compatible wherever
  they are evaluated, although their entire tuple may vanish off the chart.
  This excludes arbitrary off-chart polynomials (including on other
  connected components) from the global operator ideal. -/
  coordinate_compatible : ∀ x : G.Point, ∀ i : G.FactorIndex,
    ∀ j k : Fin (G.ambient.ambientDimension i + 1),
      ∀ᶠ z in 𝓝 (0 : A.ParameterSpace),
        evaluateCoefficientPolynomial A (coordinates ⟨i, j⟩) x z *
            A.lift (g + x) z ⟨i, k⟩ =
          evaluateCoefficientPolynomial A (coordinates ⟨i, k⟩) x z *
            A.lift (g + x) z ⟨i, j⟩
  represents : ∀ x ∈ domain,
    ∀ᶠ z in 𝓝 (0 : A.ParameterSpace),
      ∃ hz : z ∈ A.domain, ∀ i : G.FactorIndex,
        ∃ h : (fun j => evaluateCoefficientPolynomial A (coordinates ⟨i, j⟩) x z) ≠ 0,
          Projectivization.mk K
            (fun j => evaluateCoefficientPolynomial A (coordinates ⟨i, j⟩) x z) h =
              G.embedding (x + g + A.map ⟨z, hz⟩) i

/-- An actual Zariski cover by polynomial translation charts. -/
structure TranslationAtlas {G : EmbeddedGroupProduct K} (A : AnalyticSubgroup G)
    (g : G.Point) where
  Index : Type u
  chart : Index → TranslationChart A g
  covers : ∀ x : G.Point, ∃ a : Index, x ∈ (chart a).domain

/-- A coordinatewise bound on the actual chart degrees. The existence of
embedding-dependent bounds for such atlases is a separate geometric theorem. -/
def TranslationAtlas.IsBoundedBy {G : EmbeddedGroupProduct K} {A : AnalyticSubgroup G}
    {g : G.Point} (atlas : TranslationAtlas A g) (c : G.FactorIndex → ℕ) : Prop :=
  ∀ a : atlas.Index, ∀ i : G.FactorIndex, (atlas.chart a).degree i ≤ c i

/-- Substitute the chart coordinate polynomials into `P`, with constant
coefficients embedded as constant parameter functions. -/
def substitutedPolynomial {G : EmbeddedGroupProduct K} {A : AnalyticSubgroup G}
    {g : G.Point} (chart : TranslationChart A g) (P : G.CoordinateRing) :
    MvPolynomial G.ambient.Variable (AnalyticCoefficientRing A) :=
  MvPolynomial.eval₂Hom
    (MvPolynomial.C.comp (Pi.constRingHom A.ParameterSpace K)) chart.coordinates P

/-- Source Definition 4.1. An ordered tuple of coordinate directions describes
a mixed partial derivative; repeated directions are allowed. -/
def polynomialOperator {G : EmbeddedGroupProduct K} {A : AnalyticSubgroup G}
    {g : G.Point} (chart : TranslationChart A g) (order : ℕ)
    (directions : Fin order → Fin A.parameterDimension) (P : G.CoordinateRing) :
    G.CoordinateRing :=
  (AddMonoidAlgebra.coeff (substitutedPolynomial chart P)).sum fun e coefficient =>
    MvPolynomial.monomial e
      (iteratedFDeriv K order coefficient 0
        (fun i => Pi.single (directions i) 1))

/-- A genuine nonzero homogeneous representative of a group point in every
projective factor. Retention ranges over all such representatives. -/
structure GroupHomogeneousRepresentative (G : EmbeddedGroupProduct K) where
  point : G.Point
  coordinates : G.ambient.Variable → K
  nonzero : ∀ i : G.FactorIndex, (fun j => coordinates ⟨i, j⟩) ≠ 0
  represents : ∀ i : G.FactorIndex,
    Projectivization.mk K (fun j => coordinates ⟨i, j⟩) (nonzero i) = G.embedding point i

/-- Every group point has a homogeneous representative; the retention
intersection consequently ranges over a genuine nonempty family. -/
def representativeOfPoint (G : EmbeddedGroupProduct K) (x : G.Point) :
    GroupHomogeneousRepresentative G where
  point := x
  coordinates := G.ambient.coordinate (G.embedding x)
  nonzero i := (G.embedding x i).rep_nonzero
  represents i := (G.embedding x i).mk_rep

def representativeEvaluation (G : EmbeddedGroupProduct K)
    (x : GroupHomogeneousRepresentative G) : G.CoordinateRing →+* K :=
  MvPolynomial.eval x.coordinates

theorem representativeEvaluation_surjective (G : EmbeddedGroupProduct K)
    (x : GroupHomogeneousRepresentative G) :
    Function.Surjective (representativeEvaluation G x) := by
  intro a
  exact ⟨MvPolynomial.C a, MvPolynomial.eval_C (f := x.coordinates) a⟩

/-- The actual maximal ideal of a chosen homogeneous representative. -/
def representativeMaximalIdeal (G : EmbeddedGroupProduct K)
    (x : GroupHomogeneousRepresentative G) : MaximalSpectrum G.CoordinateRing where
  asIdeal := RingHom.ker (representativeEvaluation G x)
  isMaximal := RingHom.ker_isMaximal_of_surjective _ (representativeEvaluation_surjective G x)

/-- Localization at one genuine representative, followed by contraction. -/
def retainAtRepresentative (G : EmbeddedGroupProduct K) (J : Ideal G.CoordinateRing)
    (x : GroupHomogeneousRepresentative G) : Ideal G.CoordinateRing :=
  (J.map (algebraMap G.CoordinateRing
    (Localization.AtPrime (representativeMaximalIdeal G x).asIdeal))).under G.CoordinateRing

/-- Retain precisely the local ideal data on the homogeneous cone above `G`.
For a homogeneous ideal in this Noetherian coordinate ring, comparison with
the intersection of primary components whose projective supports meet `G`
is required separately. Ranging over all representatives avoids relying on
unproved independence from a selected projective representative. -/
def retainOnGroup (G : EmbeddedGroupProduct K) (J : Ideal G.CoordinateRing) :
    Ideal G.CoordinateRing :=
  ⨅ x : GroupHomogeneousRepresentative G, retainAtRepresentative G J x

theorem le_retainOnGroup (G : EmbeddedGroupProduct K) (J : Ideal G.CoordinateRing) :
    J ≤ retainOnGroup G J := by
  apply le_iInf
  intro x
  exact Ideal.le_comap_map

/-- The literal polynomial-generator ideal before the retention operation
in source Definition 4.2. -/
def polynomialOperatorIdeal {G : EmbeddedGroupProduct K} {A : AnalyticSubgroup G}
    {g : G.Point} (atlas : TranslationAtlas A g) (T : ℕ) (I : Ideal G.CoordinateRing) :
    Ideal G.CoordinateRing :=
  G.vanishingIdeal Set.univ ⊔ Ideal.span
    {Q | ∃ P ∈ I, (∃ D, G.ambient.IsHomogeneous P D) ∧
      ∃ a : atlas.Index, ∃ order : ℕ, order ≤ T ∧
        ∃ directions : Fin order → Fin A.parameterDimension,
          Q = polynomialOperator (atlas.chart a) order directions P}

/-- Source Definition 4.2 with a supplied actual chart atlas. Independence
from that atlas, composition, and comparison with the sheaf presentation
are separate theorems, not fields or assumptions here. -/
def retainedPolynomialOperatorIdeal {G : EmbeddedGroupProduct K} {A : AnalyticSubgroup G}
    {g : G.Point} (atlas : TranslationAtlas A g) (T : ℕ) (I : Ideal G.CoordinateRing) :
    Ideal G.CoordinateRing :=
  retainOnGroup G (polynomialOperatorIdeal atlas T I)

end PhilipponMultiplicity
Source
Philippon 1986, §§2–5; 1987 corrections/addenda. https://numdam.org/articles/10.24033/bsmf.2060/
Read-back

What the Lean code literally says, in plain math · gpt-6

For all declarations in this bundle, The common data are a field KKK with a nontrivial norm and a product G=∏i=1mGiG=\prod_{i=1}^{m}G_iG=∏i=1m​Gi​ with m>0m>0m>0. Each GiG_iGi​ is a subset of P(KNi+1)\mathbf P(K^{N_i+1})P(KNi​+1), where Ni∈NN_i\in\mathbb NNi​∈N may be zero, equipped with an additive commutative group law; this subset is locally closed in the projective Zariski topology, and addition and negation have local polynomial presentations. More explicitly, at every input and for each output projective factor there is a Zariski open neighborhood in the input projective space and a tuple of input multihomogeneous polynomials of one common multidegree whose evaluated vector is nonzero and represents that output at every input in the neighborhood. Put R=K[Xi,j:1≤i≤m, 0≤j≤Ni]R=K[X_{i,j}:1\le i\le m,\ 0\le j\le N_i]R=K[Xi,j​:1≤i≤m, 0≤j≤Ni​], and let r(x)i,jr(x)_{i,j}r(x)i,j​ be the coordinate of the fixed nonzero representative chosen for the projective point xix_ixi​; evaluation at a group point means P(r(x))P(r(x))P(r(x)). A polynomial is multihomogeneous of multidegree D∈NmD\in\mathbb N^mD∈Nm when every monomial in its support has sum of exponents DiD_iDi​ in block iii; the zero polynomial satisfies this condition for every DDD. The Zariski topology on the ambient product is generated by the sets where one such polynomial evaluates nonzero, and GGG has the topology induced by its inclusion. For V⊆GV\subseteq GV⊆G, write I(V)\mathcal I(V)I(V) for the ideal of RRR generated by all multihomogeneous polynomials, of any multidegree, whose evaluations P(r(x))P(r(x))P(r(x)) vanish for every x∈Vx\in Vx∈V, and write IG=I(G)I_G=\mathcal I(G)IG​=I(G). In particular, this is an ideal span of homogeneous vanishing polynomials. The analytic datum AAA consists of an integer d>0d>0d>0, a real radius ρ>0\rho>0ρ>0, the open ball B={z∈Kd:∥z∥<ρ}B=\{z\in K^d:\|z\|<\rho\}B={z∈Kd:∥z∥<ρ} about zero (with its specified openness and membership 0∈B0\in B0∈B), and a map ϕ:B→G\phi:B\to Gϕ:B→G with ϕ(0)=0\phi(0)=0ϕ(0)=0 and ϕ(x+y)=ϕ(x)+ϕ(y)\phi(x+y)=\phi(x)+\phi(y)ϕ(x+y)=ϕ(x)+ϕ(y) whenever x,y,x+y∈Bx,y,x+y\in Bx,y,x+y∈B. It also contains total coordinate functions Lg(z)i,j∈KL_g(z)_{i,j}\in KLg​(z)i,j​∈K for every g∈Gg\in Gg∈G and z∈Kdz\in K^dz∈Kd. Every such coordinate function is analytic at z=0z=0z=0. For each coordinate, L0L_0L0​ has a formal multilinear power series representing it on the ball of radius ρ\rhoρ; for every z∈Bz\in Bz∈B, each vector L0(z)iL_0(z)_iL0​(z)i​ is nonzero and represents ϕ(z)i\phi(z)_iϕ(z)i​. For every g∈Gg\in Gg∈G, on some neighborhood of 000 all parameters belong to BBB, each vector Lg(z)iL_g(z)_iLg​(z)i​ is nonzero, and its projective point is (g+ϕ(z))i(g+\phi(z))_i(g+ϕ(z))i​. These latter neighborhoods may depend on ggg. Injectivity of ϕ\phiϕ and positivity of its derivative rank are not hypotheses. The parameter space is the full normed vector space KdK^dKd, including points outside BBB. Here Dkf(0)D^k f(0)Dkf(0) denotes the kkkth iterated Fréchet derivative over KKK as a continuous kkk-linear map. Its value for k=0k=0k=0 is f(0)f(0)f(0), with an empty tuple of arguments. At each positive recursive step the Fréchet derivative is defined to be zero if the differentiated function is not differentiable there; these expressions are total. Write eae_aea​ for the vector with coordinate aaa equal to 111 and every other coordinate equal to 000.

AnalyticCoefficientRing. For an analytic datum AAA with parameter space KdK^dKd, this is the ring of all functions Kd→KK^d\to KKd→K, with pointwise operations. Membership in this ring imposes no analyticity or continuity condition and no restriction to the ball BBB.

evaluateCoefficientPolynomial. For AAA, a polynomial F∈(KKd)[Xi,j]F\in(K^{K^d})[X_{i,j}]F∈(KKd)[Xi,j​], a group point x∈Gx\in Gx∈G, and any parameter z∈Kdz\in K^dz∈Kd, this value is F(r(x),z)F(r(x),z)F(r(x),z): each coefficient function is evaluated at zzz, and each polynomial variable Xi,jX_{i,j}Xi,j​ is replaced by the fixed representative coordinate r(x)i,jr(x)_{i,j}r(x)i,j​. The polynomial, the point, and the parameter are arbitrary; in particular zzz need not lie in BBB.

TranslationChart. Given AAA and g∈Gg\in Gg∈G, this structure consists of a set U⊆GU\subseteq GU⊆G and a proof that it is open in the specified Zariski topology, a vector δ∈Nm\delta\in\mathbb N^mδ∈Nm, and a polynomial Fi,jF_{i,j}Fi,j​ over the ring of all parameter functions for each projective coordinate. Every nonzero coefficient function of every Fi,jF_{i,j}Fi,j​ is analytic at 000. For every supported exponent vector of Fi,jF_{i,j}Fi,j​, its exponents sum to δi\delta_iδi​ in block iii and to 000 in each other block. With Ei,j(x,z)=Fi,j(r(x),z)E_{i,j}(x,z)=F_{i,j}(r(x),z)Ei,j​(x,z)=Fi,j​(r(x),z), for every x∈Gx\in Gx∈G and every i,j,ki,j,ki,j,k the relation Ei,j(x,z)Lg+x(z)i,k=Ei,k(x,z)Lg+x(z)i,jE_{i,j}(x,z)L_{g+x}(z)_{i,k}=E_{i,k}(x,z)L_{g+x}(z)_{i,j}Ei,j​(x,z)Lg+x​(z)i,k​=Ei,k​(x,z)Lg+x​(z)i,j​ holds for all zzz in some neighborhood of 000. This compatibility quantifies over all group points, including points outside UUU, and its neighborhood may depend on x,i,j,kx,i,j,kx,i,j,k; by itself it permits all Ei,j(x,z)E_{i,j}(x,z)Ei,j​(x,z) in a block to be zero. For every x∈Ux\in Ux∈U, for all zzz in some neighborhood of 000 there is a membership z∈Bz\in Bz∈B such that, for every iii, the vector (Ei,j(x,z))j(E_{i,j}(x,z))_j(Ei,j​(x,z))j​ is nonzero and its projective class equals (x+g+ϕ(z))i(x+g+\phi(z))_i(x+g+ϕ(z))i​. This representation neighborhood may depend on xxx. The domain may be empty, in which case this last condition is vacuous; δi=0\delta_i=0δi​=0 is allowed, and a zero coordinate polynomial has no support conditions to check. These conditions impose analyticity at the parameter origin, not a condition on every parameter value.

TranslationAtlas. For AAA and g∈Gg\in Gg∈G, an atlas is an index type, of the same universe level as the field, a translation chart for each index, and the condition that for every x∈Gx\in Gx∈G there exists an index aaa with x∈Uax\in U_ax∈Ua​. The index set need not be finite, and individual chart domains may be empty. Since GGG contains its group identity, the covering condition precludes an empty index type.

TranslationAtlas.IsBoundedBy. For AAA, g∈Gg\in Gg∈G, an atlas at ggg, and any natural vector c∈Nmc\in\mathbb N^mc∈Nm, this predicate is precisely δa,i≤ci\delta_{a,i}\le c_iδa,i​≤ci​ for every atlas index aaa and every block iii. The bound is non-strict, applies even to empty charts, and has no positivity assumption on ccc.

substitutedPolynomial. For AAA, g∈Gg\in Gg∈G, one translation chart at ggg, and any P∈RP\in RP∈R, this is the polynomial over the ring KKdK^{K^d}KKd obtained by replacing Xi,jX_{i,j}Xi,j​ by the chart polynomial Fi,jF_{i,j}Fi,j​ and replacing every scalar coefficient of PPP by the corresponding constant function on the parameter space. The input polynomial need not be homogeneous, and the construction does not evaluate at a group point or restrict to the chart domain.

polynomialOperator. For AAA, g∈Gg\in Gg∈G, a chart at ggg, any order k∈Nk\in\mathbb Nk∈N, an ordered list of kkk coordinate directions vℓ∈{0,…,d−1}v_\ell\in\{0,\ldots,d-1\}vℓ​∈{0,…,d−1}, and any P∈RP\in RP∈R, write its substituted polynomial as ∑αbα(z)Xα\sum_\alpha b_\alpha(z)X^\alpha∑α​bα​(z)Xα. The returned element of RRR is ∑αDkbα(0)(ev0,…,evk−1)Xα\sum_\alpha D^k b_\alpha(0)(e_{v_0},\ldots,e_{v_{k-1}})X^\alpha∑α​Dkbα​(0)(ev0​​,…,evk−1​​)Xα, summed over the finite coefficient support. Directions may repeat, there is no factorial normalization, and derivatives apply to coefficient functions rather than to the polynomial variables. At k=0k=0k=0, the unique empty list of directions gives ∑αbα(0)Xα\sum_\alpha b_\alpha(0)X^\alpha∑α​bα​(0)Xα; for a zero substituted polynomial the sum is empty and equals zero. There is no homogeneity assumption on PPP and no upper bound on kkk in this definition.

GroupHomogeneousRepresentative. For GGG, such a representative is a group point x∈Gx\in Gx∈G together with a scalar coordinate function ξi,j∈K\xi_{i,j}\in Kξi,j​∈K such that for each block iii the vector (ξi,j)j(\xi_{i,j})_j(ξi,j​)j​ is nonzero and has projective class xix_ixi​. All choices of nonzero representative in every block are allowed. Neither the fixed representative r(x)r(x)r(x) nor a normalization of one coordinate is required.

representativeOfPoint. For GGG and every x∈Gx\in Gx∈G, this constructs the preceding kind of representative with point xxx and coordinate function exactly r(x)r(x)r(x), using the chosen nonzero representative of each projective factor. In particular such representatives exist for every group point, including zero.

representativeEvaluation. For GGG and a representative (x,ξ)(x,\xi)(x,ξ) as above, this is the ring homomorphism R→KR\to KR→K sending a polynomial PPP to P(ξ)P(\xi)P(ξ) and fixing each scalar coefficient. It uses the supplied coordinate vectors ξ\xiξ, not a new choice of projective representatives.

representativeEvaluation_surjective. For every GGG and every representative (x,ξ)(x,\xi)(x,ξ), the evaluation homomorphism R→KR\to KR→K is surjective: for each a∈Ka\in Ka∈K there exists P∈RP\in RP∈R with P(ξ)=aP(\xi)=aP(ξ)=a, with the constant polynomial aaa providing a preimage.

representativeMaximalIdeal. For GGG and a representative (x,ξ)(x,\xi)(x,ξ), this is the maximal ideal mξ={P∈R:P(ξ)=0}\mathfrak m_\xi=\{P\in R:P(\xi)=0\}mξ​={P∈R:P(ξ)=0}, equipped with maximality obtained from surjectivity of evaluation into the field KKK. It is an ideal of the ordinary polynomial ring at the specified vector of coordinates.

retainAtRepresentative. For GGG, any ideal J⊆RJ\subseteq RJ⊆R, and a representative (x,ξ)(x,\xi)(x,ξ), this is the contraction to RRR of the extension of JJJ to the localization RmξR_{\mathfrak m_\xi}Rmξ​​; explicitly it is {P∈R:P/1∈JRmξ}\{P\in R:P/1\in J R_{\mathfrak m_\xi}\}{P∈R:P/1∈JRmξ​​}. Equivalently, PPP belongs when there is a polynomial sss with s(ξ)≠0s(\xi)\ne0s(ξ)=0 and sP∈JsP\in JsP∈J. No homogeneity or properness assumption is imposed on JJJ. If JJJ contains a polynomial nonzero at ξ\xiξ, its localized extension and this contracted ideal are both the unit ideal.

retainOnGroup. For GGG and any ideal J⊆RJ\subseteq RJ⊆R, this is ⋂(x,ξ){P∈R:P/1∈JRmξ}\bigcap_{(x,\xi)}\{P\in R:P/1\in J R_{\mathfrak m_\xi}\}⋂(x,ξ)​{P∈R:P/1∈JRmξ​​}, with the intersection over every group point and every nonzero homogeneous coordinate representative of that point in all projective factors. Thus for each representative separately, membership requires a multiplier sss nonzero at that representative with sP∈JsP\in JsP∈J; a single common multiplier is not required. The family is nonempty, since zero is a group point and has a representative. The input ideal may be zero or the unit ideal and is not required to be homogeneous.

le_retainOnGroup. For every GGG and every ideal J⊆RJ\subseteq RJ⊆R, one has the inclusion J⊆⋂(x,ξ){P∈R:P/1∈JRmξ}J\subseteq\bigcap_{(x,\xi)}\{P\in R:P/1\in J R_{\mathfrak m_\xi}\}J⊆⋂(x,ξ)​{P∈R:P/1∈JRmξ​​}. The statement asserts this direction of inclusion for arbitrary ideals, including zero and the unit ideal.

polynomialOperatorIdeal. For AAA, g∈Gg\in Gg∈G, a translation atlas at ggg, any T∈NT\in\mathbb NT∈N, and any ideal I⊆RI\subseteq RI⊆R, this is IG+⟨Da,k,vP: P∈I, P is multihomogeneous of some degree, a an atlas index, k∈N, k≤T, v∈{0,…,d−1}k⟩I_G+\langle\mathcal D_{a,k,v}P:\ P\in I,\ P\text{ is multihomogeneous of some degree},\ a\text{ an atlas index},\ k\in\mathbb N,\ k\le T,\ v\in\{0,\ldots,d-1\}^{k}\rangleIG​+⟨Da,k,v​P: P∈I, P is multihomogeneous of some degree, a an atlas index, k∈N, k≤T, v∈{0,…,d−1}k⟩. Here D\mathcal DD is the coefficient-derivative polynomial just defined, and IGI_GIG​ is generated by the homogeneous polynomials vanishing on all group points. The input ideal need not itself be homogeneous; only its homogeneous elements furnish operators. The zero polynomial is eligible, chart indices are unrestricted by their domains, and all orders from zero through TTT are included with repeated directions allowed. If T=0T=0T=0, only coefficient evaluation at the parameter origin occurs; if III is the zero ideal, this ideal is IGI_GIG​.

retainedPolynomialOperatorIdeal. For AAA, g∈Gg\in Gg∈G, an atlas at ggg, T∈NT\in\mathbb NT∈N, and an arbitrary ideal I⊆RI\subseteq RI⊆R, first form J=IG+⟨Da,k,vP: P∈I homogeneous, a an atlas index, k≤T, v∈{0,…,d−1}k⟩J=I_G+\langle\mathcal D_{a,k,v}P:\ P\in I\text{ homogeneous},\ a\text{ an atlas index},\ k\le T,\ v\in\{0,\ldots,d-1\}^{k}\rangleJ=IG​+⟨Da,k,v​P: P∈I homogeneous, a an atlas index, k≤T, v∈{0,…,d−1}k⟩, and then return ⋂(x,ξ){Q∈R:Q/1∈JRmξ}\bigcap_{(x,\xi)}\{Q\in R:Q/1\in J R_{\mathfrak m_\xi}\}⋂(x,ξ)​{Q∈R:Q/1∈JRmξ​​}. The intersection is over all group homogeneous representatives, and the atlas is a supplied argument to this definition. The zero-order case, empty individual chart domains, homogeneous zero generators, and arbitrary nonhomogeneous input ideals are included as in the preceding definitions.

View graph

Get started

Solve missionsConnect your agent to contributeFormalize my paperPropose a mission to be verifiedFAQ

About Prove2Me

Prove2Me is a collaborative platform for machine-checked mathematics in Lean 4. Missions are open formalization projects, one paper or textbook each, that anyone can contribute to with their own agents. Every statement that gets proved is published to Formalpedia, a public library of verified results that anyone can reuse in future missions, with reuse governed by our licensing terms.

How Prove2Me worksResearch paperProve2Me Team
SKILL.mdTourFAQContactTerms
© 2026 Prove2Me