Theorem 2.5 — for sets at distance
ProvedAlonMilman.Diameter.theorem_2_5Let be a connected finite simple graph on vertices, with maximum degree and algebraic connectivity . Let and let be an integer such that every vertex of is at graph distance at least from every vertex of . With and ,
This is the one-step isoperimetric inequality: a set of relative size leaves only a small proportion of vertices at distance at least from it. Iterating it gives the concentration inequality of Theorem 2.6.
Formalization Note As in Lemma 2.1, is any integer lower bound for the pairwise distances, which is equivalent to the paper's exact distance since the bound weakens as decreases. and may be empty: then (bound ) or , and the inequality holds as stated, so no nonemptiness hypothesis is needed. is Mathlib's SimpleGraph.maxDegree, which is at least for a connected graph with .
import Mathlib import Definitions.Def_AlonMilman_Diameter_lambda1
namespace AlonMilman.Diameter
theorem theorem_2_5 {V : Type*} [Fintype V] [DecidableEq V]
(G : SimpleGraph V) [DecidableRel G.Adj] (hG : G.Connected) (hn : 2 ≤ Fintype.card V)
(A B : Finset V) (ρ : ℕ) (hρ : 1 < ρ)
(hdist : ∀ u ∈ A, ∀ v ∈ B, ρ ≤ G.dist u v) :
(B.card : ℝ) / Fintype.card V ≤
(1 - (A.card : ℝ) / Fintype.card V) /
(1 + (lambda1 G / (G.maxDegree : ℝ)) * ((A.card : ℝ) / Fintype.card V) * (ρ : ℝ) ^ 2) := by sorry
end AlonMilman.Diameter
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.