Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

The regular chain bound (DALAG Th. 8.4, finite level): for P0,…,Pr−1∈FDP_0, \dots, P_{r-1} \in F_DP0​,…,Pr−1​∈FD​ regular in turn modulo III, 2r(r−1)/2urhI+(P)(uD)≤hI(2ruD)2^{r(r-1)/2} u^r h_{I+(P)}(uD) \le h_I(2^r uD)2r(r−1)/2urhI+(P)​(uD)≤hI​(2ruD)

Proved
DiazModulus.boxHilb_chain_bound

by carlok · Oct 29, 2026 · Mathlib 0df444a (Lean v4.33.1)

commutative-algebradiaz-modulus-leanhilbert-functiontranscendence

Let KKK be a field, SSS a commutative KKK-algebra, BBB a type, and FFF a family of KKK-subspaces FTF_TFT​ of SSS indexed by T∈NBT \in \mathbb{N}^BT∈NB (a filtration, on which only the hypotheses listed are assumed). Assume that FFF is monotone (T≤T′T \le T'T≤T′ coordinatewise implies FT⊆FT′F_T \subseteq F_{T'}FT​⊆FT′​), that FFF is multiplicative (a∈FTa \in F_Ta∈FT​ and b∈FUb \in F_Ub∈FU​ imply ab∈FT+Uab \in F_{T+U}ab∈FT+U​) and that every FTF_TFT​ is finite-dimensional. Let III be an ideal of SSS, D∈NBD \in \mathbb{N}^BD∈NB, r∈Nr \in \mathbb{N}r∈N and P0,…,Pr−1∈FDP_0, \dots, P_{r-1} \in F_DP0​,…,Pr−1​∈FD​ such that each PkP_kPk​ is a nonzerodivisor on S/(I+(Pj:j<k))S/(I + (P_j : j < k))S/(I+(Pj​:j<k)). Then for every u∈Nu \in \mathbb{N}u∈N, 2r(r−1)/2urhI+(P0,…,Pr−1)(uD)≤hI(2ruD)2^{r(r-1)/2} u^r h_{I + (P_0, \dots, P_{r-1})}(uD) \le h_I(2^r u D)2r(r−1)/2urhI+(P0​,…,Pr−1​)​(uD)≤hI​(2ruD), where uDuDuD is the coordinatewise multiple. For an ideal III of SSS, hI(T)h_I(T)hI​(T) is boxHilb F I T, the dimension over KKK of the image of FTF_TFT​ in S/IS/IS/I (DALAG's H(I;T)H(I; T)H(I;T)). The names boxHilb, blockBox, torusBox, torusIdeal, torusList, torusReduced, boxD, torusCoeffPoly, torusPoint and UnmixedHeight are those of the published Definition DiazHilbertBox (module Definitions.Def_DiazHilbertBox).

Proof. Put Ik=I+(Pj:j<k)I_k = I + (P_j : j < k)Ik​=I+(Pj​:j<k) and fk(v)=hIk(vD)f_k(v) = h_{I_k}(vD)fk​(v)=hIk​​(vD). Each fk+1f_{k+1}fk+1​ is monotone (DiazModulus.boxHilb_mono), and for k<rk < rk<r, DiazModulus.boxHilb_sup_span_add_le with T=vDT = vDT=vD gives fk+1(v+1)+fk(v)≤fk(v+1)f_{k+1}(v+1) + f_k(v) \le f_k(v+1)fk+1​(v+1)+fk​(v)≤fk​(v+1), since (v+1)D=vD+D(v+1)D = vD + D(v+1)D=vD+D. The iterated telescope DiazModulus.nat_telescope_iter with j=rj = rj=r gives 2r(r−1)/2urfr(u)≤f0(2ru)2^{r(r-1)/2} u^r f_r(u) \le f_0(2^r u)2r(r−1)/2urfr​(u)≤f0​(2ru), which is the claim.

Novelty. Not found in the sources read in this form; routine. A finite-level form of the chain of DALAG Th. 8.4 (built with Lemma 8.5), which compares leading coefficients of Hilbert polynomials instead.

Preamble
import Definitions.Def_DiazHilbertBox
Formal statement
namespace DiazModulus

theorem boxHilb_chain_bound {K S B : Type*} [Field K] [CommRing S] [Algebra K S] (F : (B → ℕ) → Submodule K S)
    (hFmono : Monotone F)
    (hFmul : ∀ {Tₑ Uₑ : B → ℕ} {aₑ bₑ : S}, aₑ ∈ F Tₑ → bₑ ∈ F Uₑ → aₑ * bₑ ∈ F (Tₑ + Uₑ))
    (hFfin : ∀ Tₑ, FiniteDimensional K ↥(F Tₑ)) (I : Ideal S) (D : B → ℕ) (r : ℕ)
    (P : Fin r → S) (hP : ∀ k, P k ∈ F D)
    (hreg : ∀ k : Fin r, IsSMulRegular (S ⧸ (I ⊔ Ideal.span (P '' {j | j < k}))) (P k)) (u : ℕ) :
    2 ^ (r * (r - 1) / 2) * u ^ r * boxHilb F (I ⊔ Ideal.span (Set.range P)) (u • D)
      ≤ boxHilb F I ((2 ^ r * u) • D) := by
  sorry

end DiazModulus
Source
Finite-level form of M. Waldschmidt, Diophantine Approximation on Linear Algebraic Groups, Grundlehren der mathematischen Wissenschaften 326, Springer, 2000, Th. 8.4 and its proof through Lemma 8.5 (§8.2.3, pp. 236–238; Chapter 8 is by D. Roy). Not found in the sources read in this form; routine. Uses DiazModulus.boxHilb_mono, DiazModulus.boxHilb_sup_span_add_le and DiazModulus.nat_telescope_iter. C1.1 of the Diaz modulus mission. Formal proof: Diaz modulus mission, 29 October 2026 (C. Perassi).

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