Theorem 6.39 -- mconvex_minimizer_cut_scaling
ProvedDiscreteConvex.MConvexFunctionsC.mconvex_minimizer_cut_scalingconvex-optimizationdiscrete-convex-analysis
Theorem 6.39 (M-minimizer cut with scaling; p.158). Let be M-convex with , a positive integer, . (1) For , , and minimizing over shifts at , some minimizer of has . (2) The symmetric statement for the lower bound at .
(Murota, Discrete Convex Analysis, SIAM 2003, DOI 10.1137/1.9780898718508, p.158, Theorem 6.39.)
Preamble
import Mathlib import Definitions.Def_DiscreteConvex_MConvexFunctionsC_CharVec import Definitions.Def_DiscreteConvex_MConvexFunctionsC_DomZ import Definitions.Def_DiscreteConvex_MConvexFunctionsC_MExchangeAxiom import Definitions.Def_DiscreteConvex_MConvexFunctionsC_ArgMinOn
Formal statement
namespace DiscreteConvex.MConvexFunctionsC
open scoped Pointwise
open Classical
variable {V : Type*} [Fintype V] [DecidableEq V]
/-- Theorem 6.39 (p.158), M-minimizer cut with scaling. -/
theorem mconvex_minimizer_cut_scaling (f : (V → ℤ) → WithTop ℝ) (hf : MExchangeAxiom f)
(hne : (ArgMinOn f).Nonempty) (alpha : ℤ) (halpha : 0 < alpha) :
(∀ x ∈ DomZ f, ∀ v u : V,
(∀ s : V, f (fun w => x w + alpha * (CharVec v w - CharVec u w)) ≤
f (fun w => x w + alpha * (CharVec v w - CharVec s w))) →
∃ xstar ∈ ArgMinOn f,
xstar u ≤ x u - alpha * (1 - CharVec v u) + ((Fintype.card V : ℤ) - 1) * (alpha - 1)) ∧
(∀ x ∈ DomZ f, ∀ u v : V,
(∀ t : V, f (fun w => x w + alpha * (CharVec v w - CharVec u w)) ≤
f (fun w => x w + alpha * (CharVec t w - CharVec u w))) →
∃ xstar ∈ ArgMinOn f,
xstar v ≥ x v + alpha * (1 - CharVec u v) - ((Fintype.card V : ℤ) - 1) * (alpha - 1)) := by sorry
end DiscreteConvex.MConvexFunctionsC
Source
Murota, Discrete Convex Analysis, SIAM 2003, DOI 10.1137/1.9780898718508, p.158, Theorem 6.39
Human review
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.