Plünnecke–Ruzsa with the Ruzsa triangle inequality: small sumset implies small difference set
ProvedFinset.ruzsa_sumset_to_differenceadditive-combinatoricsbalog-szemeredi-gowersruzsa-calculussumsets
Let be an additive commutative group and let , . Let be nonempty finite sets with
Then the difference set satisfies
The proof is ordinary Ruzsa calculus, in two moves, and owes nothing to the Balog-Szemeredi-Gowers literature. First the Plunnecke-Ruzsa inequality (Mathlib's Finset.pluennecke_ruzsa_inequality_nsmul_add, in the form ) upgrades the hypothesis to
Then the Ruzsa triangle inequality (Finset.ruzsa_triangle_inequality_sub_add_add) gives
and the relative-density hypothesis finishes by division.
In the Balog-Szemeredi-Gowers project this is the last step, converting the graph step's output — a bound on the honest sum — into the stated bound on the difference set , at the cost of cubing and dividing by the relative density .
Preamble
import Mathlib open scoped Pointwise
Formal statement
theorem Finset.ruzsa_sumset_to_difference {G : Type*} [AddCommGroup G] [DecidableEq G] :
∀ K c : ℝ, 0 < K → 0 < c → ∀ A B : Finset G, A.Nonempty → B.Nonempty →
((A + B).card : ℝ) ≤ K * A.card →
c * (A.card : ℝ) ≤ (B.card : ℝ) →
((A - B).card : ℝ) ≤ K ^ 3 / c * A.card := by sorry
Source
Plunnecke-Ruzsa inequality composed with the Ruzsa triangle inequality; proved here via Mathlib's Finset.pluennecke_ruzsa_inequality_nsmul_add. NOT stated in any of the Balog-Szemeredi, Gowers, Fox-Sudakov or Tao-Vu BSG arguments. See Petridis, New proofs of Plunnecke-type estimates for product sets in groups, Combinatorica 32 (2012) 721-733, arXiv:1101.3507. Formalized in https://github.com/mysticflounder/lean-formalizations/blob/dd46c17a2a034d7bfa0df02e7f77834d35592864/lean/LeanFormalizations/Combinatorics/Additive/BalogSzemerediGowers.lean#L593-L672
Human review
Confirmed by the mission captain (proposal self-audit).