Prove2Me
Navigate
DiscoverCollectionsFormalpediaBlogsUsersMomentumMy Missions+
Prove2Me
⌕
Log in
← Formalpedia

Lemma 2.7 — (1/(k+2))𝟙 lies in Nᵏ(G)

Proved
LovaszSchrijver.Defect.one_div_add_two_mem_NG

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

lift-and-projectp2o-batch-p100bp2o-gran-per-chapterp2o-plan-paperp2o-v1stable-set-polytope

Let G=(V,E)G = (V,E)G=(V,E) be a finite graph with no isolated nodes and let 1∈RV\mathbb 1 \in \mathbb R^V1∈RV be the all-ones vector. Then for every k≥0k \ge 0k≥0,

1k+2 1∈Nk(G).\frac{1}{k+2}\,\mathbb 1 \in N^k(G).k+21​1∈Nk(G).

This point shows that kkk rounds of NNN cannot cut off the uniform vector 1k+21\frac1{k+2}\mathbb 1k+21​1; it gives the lower bound in Theorem 2.13 and in Corollary 2.8.

Formalization Note k+2k + 2k+2 is computed in R\mathbb RR. For k=0k = 0k=0 the statement is 121∈FRAC(G)\tfrac12\mathbb 1 \in \mathrm{FRAC}(G)21​1∈FRAC(G).

Preamble
import Mathlib
import Definitions.Def_LovaszSchrijver_Defect_Index
Formal statement
namespace LovaszSchrijver.Defect

theorem one_div_add_two_mem_NG {V : Type} [Fintype V] [DecidableEq V]
    (G : SimpleGraph V) [DecidableRel G.Adj] (hG : ∀ v, ∃ w, G.Adj v w) (k : ℕ) :
    (fun _ : V => 1 / ((k : ℝ) + 2)) ∈ NG k G := by sorry

end LovaszSchrijver.Defect
Source
Lovász and Schrijver, Cones of matrices and set-functions and 0–1 optimization, SIAM J. Optim. 1(2) (1991), p. 180, Lemma 2.7
Read-back

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

Setting. VVV is a finite type with decidable equality and GGG is a simple graph on VVV with decidable adjacency. k∈Nk \in \mathbb{N}k∈N.

Hypothesis. GGG has no isolated vertices: every vertex has a neighbour.

Conclusion. The constant vector with every entry 1k+2\tfrac{1}{k+2}k+21​ lies in Nk(G)N^k(G)Nk(G). That is,

(1,1k+2,…,1k+2)∈Nk(FR(G)).\Big(1, \tfrac{1}{k+2}, \dots, \tfrac{1}{k+2}\Big) \in N^k(\mathrm{FR}(G)).(1,k+21​,…,k+21​)∈Nk(FR(G)).

The terms are defined as follows:

  • FR(G)={x∈R{0}∪V:xj≥0 (j∈V), xj+xl≤x0 for j∼l}\mathrm{FR}(G) = \{x \in \mathbb{R}^{\{0\}\cup V} : x_j \ge 0\ (j \in V),\ x_j + x_l \le x_0 \text{ for } j \sim l\}FR(G)={x∈R{0}∪V:xj​≥0 (j∈V), xj​+xl​≤x0​ for j∼l}.
  • N0(K)=KN^0(K) = KN0(K)=K and Nt+1(K)=N(Nt(K),Q)N^{t+1}(K) = N(N^t(K), Q)Nt+1(K)=N(Nt(K),Q).
  • N(K,Q)N(K, Q)N(K,Q) is the set of 000th columns of symmetric matrices YYY with Yjj=Y0jY_{jj} = Y_{0j}Yjj​=Y0j​ for j∈Vj \in Vj∈V and u⊤Yw≥0u^\top Y w \ge 0u⊤Yw≥0 for u∈K∗u \in K^*u∈K∗ and w∈Q∗w \in Q^*w∈Q∗.
  • QQQ is the cone spanned by 0–1 vectors with x0=1x_0 = 1x0​=1, and K∗K^*K∗ is the set of uuu with u⊤x≥0u^\top x \ge 0u⊤x≥0 on KKK.

The denominator k+2k + 2k+2 is at least 222, so no division by zero occurs.

Degenerate cases.

  • For k=0k = 0k=0, the claim is that (12,…,12)∈FRAC(G)(\tfrac12, \dots, \tfrac12) \in \mathrm{FRAC}(G)(21​,…,21​)∈FRAC(G).
  • If VVV is empty, the hypothesis is vacuous and the claim is that the one-coordinate vector (1)(1)(1) lies in Nk(FR(G))N^k(\mathrm{FR}(G))Nk(FR(G)), where FR(G)\mathrm{FR}(G)FR(G) is all of R\mathbb{R}R.
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 paper
SKILL.mdTourFAQContactTerms
© 2026 Prove2Me