Lemma 8, PDF p. 20 — there is an optimal affine solution with P̂_ij = μ (i ≠ j), P̂_jj = θ, q̂_j = λ
ProvedAffinePolicies.LargeGap.lemma_8affine-policiesp2o-batch-pfp2ap2o-gran-per-chapterp2o-plan-paperp2o-v1robust-optimizationsymmetry
Let , , and consider the instance (19). There exists an optimal affine solution ,
and real numbers such that
- for all and for all ;
- for all .
The lemma reduces the analysis of the best affine policy on this instance to three scalar parameters, which the proof of Theorem 3 then bounds.
Formalization Note The existence of an optimal affine solution (attainment of the minimum) is part of the conclusion, as on the page. The hypothesis excludes the empty instance; the threshold (18) is not needed.
Preamble
import Mathlib import Definitions.Def_AffinePolicies_LargeGap_Setting
Formal statement
namespace AffinePolicies.LargeGap
theorem lemma_8 (δ : ℝ) (m : ℕ) (hm0 : 0 < m) :
∃ (x : Fin m → ℝ) (P : Matrix (Fin m) (Fin m) ℝ) (q : Fin m → ℝ) (μ θ lam : ℝ),
AffinePolicies.Simplex.IsOptimalAff (A19 m) (B19 m δ) (c19 m) (d19 m) (U19 m δ) x P q ∧
(∀ i j, i ≠ j → P i j = μ) ∧ (∀ j, P j j = θ) ∧ (∀ j, q j = lam) := by sorry
end AffinePolicies.LargeGap
Source
Bertsimas & Goyal, On the power and limitations of affine policies in two-stage adaptive optimization, Math. Program. Ser. A, DOI 10.1007/s10107-011-0444-4, Lemma 8, PDF p. 20
Human review
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.