Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Theorem 9 — maximal Λ\LambdaΛ-free convex sets of a subspace W⊇VW\supseteq VW⊇V

Proved
MaxLatticeFree.Geometry.maxLambdaFree_in_W_iff

by mikedeng1 · Sep 27, 2026 · Mathlib 0df444a (Lean v4.33.1)

discrete-geometryinteger-programminglatticesp2o-batch-p200bp2o-gran-per-chapterp2o-plan-paperp2o-v1

Let Λ\LambdaΛ be a lattice of a linear space V⊆RnV\subseteq\mathbb R^nV⊆Rn and let WWW be a linear space of Rn\mathbb R^nRn containing VVV, with dim⁡W≥1\dim W\ge1dimW≥1. A set S⊆RnS\subseteq\mathbb R^nS⊆Rn is a maximal Λ\LambdaΛ-free convex set of WWW if and only if one of the following holds:

  1. SSS is a polyhedron in WWW with dim⁡(S)=dim⁡(W)\dim(S)=\dim(W)dim(S)=dim(W), S∩VS\cap VS∩V is a maximal Λ\LambdaΛ-free convex set of VVV, and the map
F⟼F∩VF\longmapsto F\cap VF⟼F∩V

is a bijection from the facets of SSS onto the facets of S∩VS\cap VS∩V; 2. SSS is an affine hyperplane of WWW of the form S=v+LS=v+LS=v+L, where v∈Wv\in Wv∈W and L⊆WL\subseteq WL⊆W is a linear hyperplane of WWW, such that L∩VL\cap VL∩V is a hyperplane of VVV (dim⁡(L∩V)=dim⁡V−1\dim(L\cap V)=\dim V-1dim(L∩V)=dimV−1) that is not a Λ\LambdaΛ-subspace of VVV; 3. SSS is a half-space of WWW, S={x∈W∣⟨a,x⟩≤b}S=\{x\in W\mid \langle a,x\rangle\le b\}S={x∈W∣⟨a,x⟩≤b} with aaa not orthogonal to WWW, that contains VVV on its boundary: ⟨a,x⟩=b\langle a,x\rangle=b⟨a,x⟩=b for all x∈Vx\in Vx∈V.

Here Λ\LambdaΛ may span a proper subspace VVV of WWW, as happens for Λ=Zn∩W\Lambda=\mathbb Z^n\cap WΛ=Zn∩W when WWW is an irrational subspace (for example W={x1+x2+2x3=0}⊂R3W=\{x_1+x_2+\sqrt2x_3=0\}\subset\mathbb R^3W={x1​+x2​+2​x3​=0}⊂R3, whose integer points span V={x1+x2=0, x3=0}V=\{x_1+x_2=0,\ x_3=0\}V={x1​+x2​=0, x3​=0}). The theorem reduces the classification in WWW to Lovász's classification in VVV (Theorem 10) and adds the two genuinely new cases (ii) and (iii). It is the geometric input for the characterization of minimal valid inequalities of the multi-row relaxation Rf(W)R_f(W)Rf​(W).

Formalization Note The page does not state dim⁡W≥1\dim W\ge1dimW≥1; it is added because for W=V={0}W=V=\{0\}W=V={0} the only maximal Λ\LambdaΛ-free set is ∅\emptyset∅, which satisfies none of the three cases (the theorem as stated is true when V={0}≠WV=\{0\}\neq WV={0}=W). "One-to-one correspondence, with F∩VF\cap VF∩V the facet corresponding to FFF" is stated as: F∩VF\cap VF∩V is a facet of S∩VS\cap VS∩V for every facet FFF of SSS, the map is injective, and every facet of S∩VS\cap VS∩V is of this form. Dimensions are integer valued with dim⁡∅=−1\dim\emptyset=-1dim∅=−1; hyperplane conditions are written dim⁡L+1=dim⁡W\dim L+1=\dim WdimL+1=dimW in natural numbers.

Preamble
import Mathlib
import Definitions.Def_MaxLatticeFree_Geometry_IsLatticeOf
import Definitions.Def_MaxLatticeFree_Geometry_IsLambdaSubspace
import Definitions.Def_MaxLatticeFree_Geometry_intW
import Definitions.Def_MaxLatticeFree_Geometry_IsLambdaFree
import Definitions.Def_MaxLatticeFree_Geometry_affDim
import Definitions.Def_MaxLatticeFree_Geometry_Polyhedron

open Pointwise
Formal statement
namespace MaxLatticeFree.Geometry

theorem maxLambdaFree_in_W_iff {n : ℕ} (Λ : AddSubgroup (EuclideanSpace ℝ (Fin n)))
    (V W : Submodule ℝ (EuclideanSpace ℝ (Fin n))) (hΛ : IsLatticeOf Λ V) (hVW : V ≤ W)
    (hW : 0 < Module.finrank ℝ W) (S : Set (EuclideanSpace ℝ (Fin n))) :
    IsMaxLambdaFree Λ W S ↔
      (IsPolyhedronIn W S ∧ affDim S = (Module.finrank ℝ W : ℤ) ∧
        IsMaxLambdaFree Λ V (S ∩ (V : Set (EuclideanSpace ℝ (Fin n)))) ∧
        (∀ F : Set (EuclideanSpace ℝ (Fin n)), IsFacet S F → IsFacet (S ∩ (V : Set (EuclideanSpace ℝ (Fin n)))) (F ∩ (V : Set (EuclideanSpace ℝ (Fin n))))) ∧
        (∀ F G : Set (EuclideanSpace ℝ (Fin n)), IsFacet S F → IsFacet S G →
          F ∩ (V : Set (EuclideanSpace ℝ (Fin n))) = G ∩ (V : Set (EuclideanSpace ℝ (Fin n))) → F = G) ∧
        (∀ G : Set (EuclideanSpace ℝ (Fin n)), IsFacet (S ∩ (V : Set (EuclideanSpace ℝ (Fin n)))) G →
          ∃ F : Set (EuclideanSpace ℝ (Fin n)), IsFacet S F ∧ F ∩ (V : Set (EuclideanSpace ℝ (Fin n))) = G)) ∨
      (∃ v ∈ W, ∃ L : Submodule ℝ (EuclideanSpace ℝ (Fin n)), L ≤ W ∧
        Module.finrank ℝ L + 1 = Module.finrank ℝ W ∧
        S = v +ᵥ (L : Set (EuclideanSpace ℝ (Fin n))) ∧
        Module.finrank ℝ (L ⊓ V : Submodule ℝ (EuclideanSpace ℝ (Fin n))) + 1 = Module.finrank ℝ V ∧
        ¬ IsLambdaSubspace Λ V (L ⊓ V)) ∨
      (∃ (a : EuclideanSpace ℝ (Fin n)) (b : ℝ), (∃ w ∈ W, inner ℝ a w ≠ 0) ∧
        S = (W : Set (EuclideanSpace ℝ (Fin n))) ∩ {x | inner ℝ a x ≤ b} ∧
        (V : Set (EuclideanSpace ℝ (Fin n))) ⊆ {x | inner ℝ a x = b}) := by sorry

end MaxLatticeFree.Geometry
Source
Basu, Conforti, Cornuéjols, Zambelli, Maximal lattice-free convex sets in linear subspaces, arXiv:1701.06543v1, p. 8, Theorem 9 (proof pp. 9–10)
Human review
  • Endorsed by Shuze Chen · Sep 27, 2026

    Confirmed by the moderator at approval.

  • Endorsed by mikedeng1 · Sep 27, 2026

    Confirmed by the mission captain (proposal self-audit).

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