Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Theorem 5 (EDPwC) — the flow relaxation has integrality gap ≥ β₁/(4c) against integral routings of congestion c − 1

Proved
EDPHardness.IntegralityGap.theorem_5_integrality_gap

by mikedeng1 · Oct 6, 2026 · Mathlib 0df444a (Lean v4.33.1)

edge-disjoint-pathsintegrality-gapmulticommodity-flowp2o-batch-pfp2bp2o-gran-per-chapterp2o-plan-paperp2o-v1

For a ccc-uniform hypergraph HHH on nnn vertices with m=⌊β2n⌋m=\lfloor\beta_2 n\rfloorm=⌊β2​n⌋ hyperedges let G(H)G(H)G(H) be the edge-disjoint-paths instance of the mission: vertices s(v),t(v)s(v),t(v)s(v),t(v) for every vertex vvv and ℓi,ri\ell_i,r_iℓi​,ri​ for every hyperedge, special edges (ℓi,ri)(\ell_i,r_i)(ℓi​,ri​), regular edges along the hyperedges containing each vvv in increasing order, and pairs (s(v),t(v))(s(v),t(v))(s(v),t(v)). Here

β1=14(log⁡n150(log⁡log⁡n)2)1/c,β2=6(2β1)c−1ln⁡β1,\beta_1=\frac14\Big(\frac{\log n}{150(\log\log n)^2}\Big)^{1/c},\qquad\beta_2=6(2\beta_1)^{c-1}\ln\beta_1,β1​=41​(150(loglogn)2logn​)1/c,β2​=6(2β1​)c−1lnβ1​,

with log⁡\loglog to base 222 and ln⁡\lnln natural.

There are absolute constants α>0\alpha>0α>0 and N0N_0N0​ such that for every n≥N0n\ge N_0n≥N0​ and every integer ccc with

2≤c≤α log⁡log⁡nlog⁡log⁡log⁡n2\le c\le\alpha\,\frac{\log\log n}{\log\log\log n}2≤c≤αlogloglognloglogn​

there is a hypergraph HHH as above for which

  1. the multicommodity flow relaxation of G(H)G(H)G(H) has a feasible fractional solution (capacity 111 on every edge) of value n/cn/cn/c, and
  2. every integral routing in G(H)G(H)G(H) in which every edge carries at most c−1c-1c−1 paths routes at most 4n/β14n/\beta_14n/β1​ pairs.

Hence the integrality gap of the relaxation on G(H)G(H)G(H), against integral solutions of congestion c−1c-1c−1, is at least

n/c4n/β1=β14c=Ω(1c(log⁡n(log⁡log⁡n)2)1/c).\frac{n/c}{4n/\beta_1}=\frac{\beta_1}{4c}=\Omega\Big(\frac1c\Big(\frac{\log n}{(\log\log n)^2}\Big)^{1/c}\Big).4n/β1​n/c​=4cβ1​​=Ω(c1​((loglogn)2logn​)1/c).

Since G(H)G(H)G(H) has 2n+2m=O(nlog⁡n)2n+2m=O(n\log n)2n+2m=O(nlogn) vertices, this is the EDPwC part of Theorem 5 of the paper, there written for congestion c−1c-1c−1 in terms of the number VVV of vertices.

Formalization Note The theorem is stated in the explicit form Section 2 proves, in the hyperedge size ccc (the integral congestion is c−1c-1c−1); the O(⋅)O(\cdot)O(⋅) bound on ccc and "sufficiently large nnn" are the absolute constants α\alphaα and N0N_0N0​, chosen before nnn and ccc. The instance is required to be G(H)G(H)G(H) for a hypergraph with exactly ⌊β2n⌋\lfloor\beta_2n\rfloor⌊β2​n⌋ hyperedges of size ccc on nnn vertices. The ANFwC part of Theorem 5 and its superconstant remark are not formalized.

Preamble
import Mathlib
import Definitions.Def_EDPHardness_IntegralityGap_FlowRelaxation
import Definitions.Def_EDPHardness_IntegralityGap_GapInstance
Formal statement
namespace EDPHardness.IntegralityGap

theorem theorem_5_integrality_gap :
    ∃ α : ℝ, 0 < α ∧ ∃ N₀ : ℕ, ∀ n : ℕ, N₀ ≤ n → ∀ c : ℕ, 2 ≤ c →
      (c : ℝ) ≤ α * Real.logb 2 (Real.logb 2 n) / Real.logb 2 (Real.logb 2 (Real.logb 2 n)) →
      ∃ H : Hyp n (numEdges n c) c,
        (∃ F : FracSol (gapGraph H) (src n (numEdges n c)) (snk n (numEdges n c)),
          F.value = (n : ℝ) / c) ∧
        ∀ R : IntRouting (gapGraph H) (src n (numEdges n c)) (snk n (numEdges n c)) (c - 1),
          (R.routed.card : ℝ) ≤ 4 * (n : ℝ) / beta1 n c := by sorry

end EDPHardness.IntegralityGap
Source
Andrews, Chuzhoy, Guruswami, Khanna, Talwar, Zhang, Inapproximability of Edge-Disjoint Paths and Low Congestion Routing on Undirected Graphs, Combinatorica 30 (2010), p. 489, Theorem 5 (EDPwC part), in the explicit form of Section 2, pp. 491 and 495–496
Human review
  • Endorsed by Shuze Chen · Oct 6, 2026

    Confirmed by the moderator at approval.

  • Endorsed by mikedeng1 · Oct 6, 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