Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

(11), proof of Theorem 2.1, p. 14 — for n > e^{γ/(ε−ε₀)}, P{α ≥ n/((2+ε)/γ·log n)} ≥ P{α ≥ n/((2+ε₀)/γ·log n + 1)} ≥ P{d/log n ≤ (2+ε₀)/γ}

Proved
RecMarkovGraph.IndepLB.eq_11

by mikedeng1 · Oct 24, 2026 · Mathlib 0df444a (Lean v4.33.1)

independence-numberp2o-batch-pfp2ap2o-gran-per-chapterp2o-plan-paperp2o-v1random-graph

Let p∈(0,1]p\in(0,1]p∈(0,1], r∈(0,1)r\in(0,1)r∈(0,1), γ=1−r\gamma=1-rγ=1−r, and fix ε>ε0>0\varepsilon>\varepsilon_0>0ε>ε0​>0. For every n>eγ/(ε−ε0)n>e^{\gamma/(\varepsilon-\varepsilon_0)}n>eγ/(ε−ε0​),

P{α(Gn,pr)≥n2+εγlog⁡n} ≥ P{α(Gn,pr)≥n2+ε0γlog⁡n+1} ≥ P{d(Gn,pr)log⁡n≤2+ε0γ}.\mathbb P\Bigl\{\alpha(G^r_{n,p})\ge\frac{n}{\frac{2+\varepsilon}{\gamma}\log n}\Bigr\}\ \ge\ \mathbb P\Bigl\{\alpha(G^r_{n,p})\ge\frac{n}{\frac{2+\varepsilon_0}{\gamma}\log n+1}\Bigr\}\ \ge\ \mathbb P\Bigl\{\frac{d(G^r_{n,p})}{\log n}\le\frac{2+\varepsilon_0}{\gamma}\Bigr\}.P{α(Gn,pr​)≥γ2+ε​lognn​} ≥ P{α(Gn,pr​)≥γ2+ε0​​logn+1n​} ≥ P{lognd(Gn,pr​)​≤γ2+ε0​​}.

Together with Theorem 2.2, which sends the right-hand side to 111, this display proves the goal.

Differences from the source:

  • reading: Definition 1, p. 3: edge (v_j, v_i) depends on edge (v_{j−1}, v_i) and "is independent of all other edges"; P(present | previous absent) = P(previous present), P(present | previous present) = r·P(previous present) → the law rmrgWeight: the rows (X^i_1, …, X^i_{i−1}), i = 2, …, n, are mutually independent, each row is a Markov chain with X^i_1 ~ Bernoulli(p) and the transition matrix (2), p. 8, whose entries use the marginal of the previous edge, i.e. the iterate recSeq (1 − r) p (j − 1) of f_γ; reason: read literally, independence of all other edges contradicts the dependence on the next edge of the same row; the generation process (pp. 2–3), the transition matrix (2) and the proof of Lemma 3.2 ("intra-iteration dependency but inter-iteration independency", p. 9) fix this reading; Lemma 3.1 certifies that the marginals of the weight obey the recursion Definition 1 presupposes
  • encoding: vertices v_1, …, v_n; edge variable X^i_j for edge (v_j, v_i), j < i; probabilities under G^r_{n,p} → vertex set Fin n with v_k = k − 1, X^i_j = G.Adj (j − 1) (i − 1); probabilities are rmrgProb n p r E, the weighted sum of rmrgWeight over all SimpleGraph (Fin n); reason: the same finite probability space (Notation, p. 4); every 1-based index range is shifted by one, no instance changes
  • encoding: "if n is sufficiently large", made explicit on the page as n > e^{γ/(ε−ε_0)} → hypothesis Real.exp ((1 − r)/(ε − ε₀)) < n; reason: the page's own threshold; not weakened to an eventual statement
  • encoding: γ := 1 − r (display (1), p. 4); log is the natural log → (1 − r) written out; Real.log of the cast (n : ℝ); reason: notation only
Preamble
import Mathlib
import Definitions.Def_RecMarkovGraph_IndepLB_Model
Formal statement
namespace RecMarkovGraph.IndepLB

theorem eq_11 (p r : ℝ) (hp : 0 < p) (hp1 : p ≤ 1) (hr : 0 < r) (hr1 : r < 1)
    (ε ε₀ : ℝ) (hε₀ : 0 < ε₀) (hε₀ε : ε₀ < ε) (n : ℕ)
    (hn : Real.exp ((1 - r) / (ε - ε₀)) < (n : ℝ)) :
    rmrgProb n p r (fun G =>
        (n : ℝ) / ((2 + ε₀) / (1 - r) * Real.log n + 1) ≤ (G.indepNum : ℝ)) ≤
      rmrgProb n p r (fun G =>
        (n : ℝ) / ((2 + ε) / (1 - r) * Real.log n) ≤ (G.indepNum : ℝ)) ∧
    rmrgProb n p r (fun G => avgDeg G / Real.log n ≤ (2 + ε₀) / (1 - r)) ≤
      rmrgProb n p r (fun G =>
        (n : ℝ) / ((2 + ε₀) / (1 - r) * Real.log n + 1) ≤ (G.indepNum : ℝ)) := by sorry

end RecMarkovGraph.IndepLB
Source
Gupte, Zhu, Large independent sets in recursive Markov random graphs, arXiv:2207.04514v4, p. 14, display (11) in the proof of Theorem 2.1
Human review
  • Endorsed by Shuze Chen · Oct 24, 2026

    Confirmed by the moderator at approval.

  • Endorsed by mikedeng1 · Oct 24, 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