Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Theorem 4.10.1 — general Helly theorem for infinite families of closed convex sets

Proved
ModernConvex.SLemma.general_helly

by Shuze Chen · Oct 20, 2026 · Mathlib 0df444a (Lean v4.33.1)

convex-geometryhelly-theorem

Let {Aα}α∈I\{A_\alpha\}_{\alpha\in I}{Aα​}α∈I​ be a family (possibly infinite) of closed convex subsets of Rn\mathbb R^nRn such that

  1. every n+1n+1n+1 sets from the family have a point in common, and
  2. there is a finite subfamily whose intersection is bounded.

Then all sets of the family have a point in common:

⋂α∈IAα≠∅.\bigcap_{\alpha\in I} A_\alpha \neq \emptyset.α∈I⋂​Aα​=∅.

The finite Helly theorem needs neither closedness nor boundedness. For infinite families both are needed: the half-lines [k,∞)[k, \infty)[k,∞), k∈Nk \in \mathbb Nk∈N, have the finite intersection property but empty intersection. The book uses this theorem to give a straightforward proof of the S-lemma (Exercises 4.45–4.46).

Formalization Note Rn\mathbb R^nRn is Fin n → ℝ. Its sup-norm bornology has the same bounded sets as the Euclidean one. "Every n+1n+1n+1 sets" is every map Fin (n + 1) → I, so repeated indices are allowed and a family with fewer than n+1n+1n+1 members is covered. The finite subfamily is a Finset I. The index type I is arbitrary and may be empty. This is not the platform's FamousTheorems.helly_theorem, which treats finite families.

Preamble
import Mathlib
Formal statement
namespace ModernConvex.SLemma

/-- Ben-Tal–Nemirovski, Theorem 4.10.1 (General Helley theorem), p. 305: let `{A_α}_{α ∈ I}` be a
family of closed convex sets in `ℝⁿ` such that (1) every `n + 1` sets from the family have a
point in common, and (2) some finite subfamily has a bounded intersection. Then all sets of the
family have a point in common. "Every `n + 1` sets" is every choice of `n + 1` indices, repetitions
allowed (so a family with fewer than `n + 1` members is covered). -/
theorem general_helly {n : ℕ} {I : Type} (A : I → Set (Fin n → ℝ))
    (hclosed : ∀ α, IsClosed (A α)) (hconvex : ∀ α, Convex ℝ (A α))
    (hinter : ∀ f : Fin (n + 1) → I, (⋂ k, A (f k)).Nonempty)
    (hbdd : ∃ s : Finset I, Bornology.IsBounded (⋂ α ∈ s, A α)) :
    (⋂ α, A α).Nonempty := by sorry

end ModernConvex.SLemma
Source
Ben-Tal & Nemirovski, Lectures on Modern Convex Optimization, SIAM 2001, p. 305, Theorem 4.10.1 (Exercise 4.44)
Read-back

What the Lean code literally says, in plain math · claude-opus-5-5

Let nnn be a natural number and III an arbitrary index type, possibly empty or infinite. Let (Aα)α∈I(A_\alpha)_{\alpha \in I}(Aα​)α∈I​ be a family of subsets of Rn\mathbb{R}^nRn. Topology and boundedness are those of Rn\mathbb{R}^nRn with the product (sup) metric, which have the same closed and bounded sets as the Euclidean ones.

Assumptions:

  1. Every AαA_\alphaAα​ is closed.
  2. Every AαA_\alphaAα​ is convex.
  3. For every choice of n+1n+1n+1 indices α0,…,αn∈I\alpha_0, \dots, \alpha_n \in Iα0​,…,αn​∈I, repetitions allowed, the intersection Aα0∩⋯∩AαnA_{\alpha_0} \cap \dots \cap A_{\alpha_n}Aα0​​∩⋯∩Aαn​​ is nonempty.
  4. There is a finite set s⊆Is \subseteq Is⊆I, possibly empty, such that ⋂α∈sAα\bigcap_{\alpha \in s} A_\alpha⋂α∈s​Aα​ is bounded. Over the empty sss this intersection is all of Rn\mathbb{R}^nRn.

Conclusion: ⋂α∈IAα\bigcap_{\alpha \in I} A_\alpha⋂α∈I​Aα​ is nonempty.

Degenerate cases.

  • III empty: assumption 3 is vacuous, because there are no maps from the nonempty index set {0,…,n}\{0,\dots,n\}{0,…,n} into an empty III. Assumption 4 forces s=∅s = \emptysets=∅, so it says Rn\mathbb{R}^nRn is bounded, which holds only for n=0n = 0n=0. The conclusion concerns the empty intersection, which is Rn\mathbb{R}^nRn and so is always nonempty.
  • n=0n = 0n=0: R0\mathbb{R}^0R0 is a single point. Assumption 4 always holds with s=∅s = \emptysets=∅. Assumption 3 with one index says each AαA_\alphaAα​ is nonempty, hence equal to that point.
  • A family with fewer than n+1n + 1n+1 members is covered by assumption 3 through repeated indices.
Human review
  • Endorsed by Community (Bot) · Oct 20, 2026

    Confirmed by the moderator at approval.

  • Endorsed by Shuze Chen · Oct 20, 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