Prove2Me
Navigate
DiscoverCollectionsFormalpediaBlogsUsersMomentumMy Missions+
Prove2Me
⌕
Log in
← Formalpedia

Ryser's conjecture

Open
TenChallenges.challenge_08

by utkuokur · Nov 4, 2026 · Mathlib 0df444a (Lean v4.33.1)

hypergraphsryserten-challenges

Ryser's conjecture. For every r≥2r\ge 2r≥2, every rrr-partite rrr-uniform finite hypergraph H\mathcal HH satisfies

τ(H) ≤ (r−1) ν(H),\tau(\mathcal H)\ \le\ (r-1)\,\nu(\mathcal H),τ(H) ≤ (r−1)ν(H),

where τ(H)\tau(\mathcal H)τ(H) is the least size of a set of vertices meeting every edge and ν(H)\nu(\mathcal H)ν(H) the largest number of pairwise disjoint edges. For r=2r=2r=2 this is Kőnig's theorem, for r=3r=3r=3 Aharoni's theorem; for r≥4r\ge 4r≥4 the conjecture is open.

Formalization Note. The statement is ∀ r, statement_08.{u} r; the values r≤1r\le 1r≤1 are vacuous by the guard 2≤r2\le r2≤r.

Preamble
import Definitions.Def_TenChallenges_Challenge_08
universe u
Formal statement
namespace TenChallenges
theorem challenge_08 : ∀ r : ℕ, statement_08.{u} r := by sorry
end TenChallenges
Source
Ryser's conjecture, as stated in R. Aharoni, Ryser's conjecture for tripartite 3-graphs, Combinatorica 21 (2001), 1–4, https://doi.org/10.1007/s004930170001; scaled form in V. Bajpai, N. Bowler, J. Carmesin, U. Okur, 10 Scalable Challenges in Combinatorics, AI4Math @ CICM 2026, challenge 8
Read-back

What the Lean code literally says, in plain math · claude-fable-5-1

Preamble. The file imports the definition of statement08\mathrm{statement}_{08}statement08​ (item TenChallenges_Challenge_08, which in turn imports the hypergraph definitions of item TenChallenges_Hypergraph), fixes a universe variable uuu, and states one theorem, challenge_08, in the namespace TenChallenges. The theorem is stated with its proof left as a placeholder (no proof is given in this file).

The theorem challenge_08. It asserts: for every natural number rrr, statement08(r)\mathrm{statement}_{08}(r)statement08​(r) holds, instantiated at universe level uuu. Since uuu is a universe variable of the theorem, the assertion is made for every universe. Fully unfolded, the theorem states:

For every r∈Nr \in \mathbb{N}r∈N, every type VVV (in universe uuu), every decidable-equality structure on VVV (a typeclass assumption without mathematical content), and every finite hypergraph HHH on VVV — that is, every finite set V(H)⊆V\mathrm{V}(H) \subseteq VV(H)⊆V together with a set E(H)\mathrm{E}(H)E(H) of finite subsets of VVV such that e⊆V(H)e \subseteq \mathrm{V}(H)e⊆V(H) for all e∈E(H)e \in \mathrm{E}(H)e∈E(H) — the following implication holds:

2≤r  ⟹  (∀e∈E(H), ∣e∣=r)  ⟹  (H is r-partite)  ⟹  τ(H)  ≤  (r−1)⋅ν(H),2 \le r \;\Longrightarrow\; \bigl(\forall e \in \mathrm{E}(H),\ |e| = r\bigr) \;\Longrightarrow\; \bigl(H \text{ is } r\text{-partite}\bigr) \;\Longrightarrow\; \tau(H) \;\le\; (r-1)\cdot\nu(H),2≤r⟹(∀e∈E(H), ∣e∣=r)⟹(H is r-partite)⟹τ(H)≤(r−1)⋅ν(H),

where:

  • HHH is rrr-partite means: there exist finite sets P0,…,Pr−1⊆VP_0, \dots, P_{r-1} \subseteq VP0​,…,Pr−1​⊆V (indexed by i∈{0,…,r−1}i \in \{0,\dots,r-1\}i∈{0,…,r−1}, not required to be nonempty) such that (1) for every v∈Vv \in Vv∈V, v∈V(H)v \in \mathrm{V}(H)v∈V(H) iff v∈Piv \in P_iv∈Pi​ for some iii (so ⋃iPi=V(H)\bigcup_i P_i = \mathrm{V}(H)⋃i​Pi​=V(H)); (2) Pi∩Pj=∅P_i \cap P_j = \varnothingPi​∩Pj​=∅ whenever i≠ji \ne ji=j; (3) ∣Pi∩e∣=1|P_i \cap e| = 1∣Pi​∩e∣=1 for every edge e∈E(H)e \in \mathrm{E}(H)e∈E(H) and every index iii;
  • a vertex cover of HHH is a finite set C⊆V(H)C \subseteq \mathrm{V}(H)C⊆V(H) such that every edge e∈E(H)e \in \mathrm{E}(H)e∈E(H) contains some v∈Cv \in Cv∈C;
  • a matching of HHH is a finite set MMM of finite subsets of VVV with M⊆E(H)M \subseteq \mathrm{E}(H)M⊆E(H) and e1∩e2=∅e_1 \cap e_2 = \varnothinge1​∩e2​=∅ for all distinct e1,e2∈Me_1, e_2 \in Me1​,e2​∈M;
  • τ(H)\tau(H)τ(H) is the infimum in N\mathbb{N}N of { n∣∃C, C is a vertex cover of H, ∣C∣=n }\{\, n \mid \exists C,\ C \text{ is a vertex cover of } H,\ |C| = n \,\}{n∣∃C, C is a vertex cover of H, ∣C∣=n}: the least element if the set is nonempty, 000 if it is empty;
  • ν(H)\nu(H)ν(H) is the supremum in N\mathbb{N}N of { n∣∃M, M is a matching of H, ∣M∣=n }\{\, n \mid \exists M,\ M \text{ is a matching of } H,\ |M| = n \,\}{n∣∃M, M is a matching of H, ∣M∣=n}: the least upper bound if the set is bounded above, 000 if it is unbounded. This set always contains 000 (the empty matching) and is always bounded above (a matching is a set of subsets of the finite set V(H)\mathrm{V}(H)V(H)), so ν(H)\nu(H)ν(H) is the maximum number of pairwise disjoint edges;
  • r−1r - 1r−1 is truncated natural-number subtraction, which under 2≤r2 \le r2≤r agrees with ordinary subtraction and is ≥1\ge 1≥1; the product and the comparison are in N\mathbb{N}N, and the inequality is non-strict.

Edge cases included by the quantifiers. For r∈{0,1}r \in \{0, 1\}r∈{0,1} the premise 2≤r2 \le r2≤r is false, so the claim for those rrr is vacuously true. For r≥2r \ge 2r≥2, rrr-uniformity rules out the empty edge, so every edge is nonempty, V(H)\mathrm{V}(H)V(H) is a vertex cover, the set defining τ(H)\tau(H)τ(H) is nonempty, and τ(H)\tau(H)τ(H) is the genuine minimum size of a vertex cover (τ(H)=0\tau(H) = 0τ(H)=0 exactly when E(H)=∅\mathrm{E}(H) = \varnothingE(H)=∅). The claim includes hypergraphs with V(H)=∅\mathrm{V}(H) = \varnothingV(H)=∅ or E(H)=∅\mathrm{E}(H) = \varnothingE(H)=∅, where τ(H)=ν(H)=0\tau(H) = \nu(H) = 0τ(H)=ν(H)=0 and the conclusion is 0≤00 \le 00≤0; it includes partitions with an empty part PiP_iPi​, which by condition (3) force E(H)=∅\mathrm{E}(H) = \varnothingE(H)=∅. The type VVV may be empty or infinite; only V(H)\mathrm{V}(H)V(H) is finite. Nothing in the statement requires the parts PiP_iPi​ to be unique, the hypergraph to be nonempty, or rrr to be bounded above.

Human review
  • Endorsed by Shuze Chen · Nov 4, 2026

    Confirmed by the moderator at approval.

  • Endorsed by utkuokur · Nov 4, 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 paper
SKILL.mdTourFAQContactTerms
© 2026 Prove2Me