Lemma B.4 — Two-point inequality for a convex function
ProvedFourBlockIP350.Main.Supplement.two_point_majorizationconvex-analysisp2o-batch-stoc27sp2o-gran-per-chapterp2o-plan-paperp2o-v1
Let be convex. Suppose satisfy
Then
The paper cites this as Theorem 1 of Kadelburg, Đukić, Lukić and Matić (2005). It is applied coordinatewise, to the component functions of the objective and to , in the proof of Lemma B.5.
Differences from the source: none.
Preamble
import Mathlib
Formal statement
namespace FourBlockIP350.Main.Supplement
/-- Lemma B.4 (Theorem 1 of [KÐLM05]), p. 24. -/
theorem two_point_majorization (ϕ : ℝ → ℝ) (hϕ : ConvexOn ℝ Set.univ ϕ)
(a b c d : ℝ) (hsum : a + b = c + d)
(hc : min a b ≤ c ∧ c ≤ max a b) (hd : min a b ≤ d ∧ d ≤ max a b) :
ϕ c + ϕ d ≤ ϕ a + ϕ b := by sorry
end FourBlockIP350.Main.Supplement
Source
Klaus Jansen, Felix Ohnesorge, Corinna Wambsganz, A Fixed-Parameter Algorithm for 4-Block Integer Programming, STOC 2027 submission #350 (October 2026), p. 24, Lemma B.4 (Theorem 1 of [KÐLM05])