B.2, p. 27 β int π© = π© β ββ² β β for the separating functional
ProvedBootRobust.Optimality.N_subset_disappdisappointmentdistributionally-robust-optimizationp2o-batch-pfp2ap2o-gran-per-chapterp2o-plan-paperp2o-v1
Let , , , a loss and be given, and let be open in . Suppose
Then (interior relative to ) and
the disappointment set of the nominal formulation with cost estimator .
This places the separating neighbourhood inside the disappointment set, so a lower bound on the bootstrap probability of is one for .
Formalization Note The supremum is taken in the extended reals.
Preamble
import Mathlib import Definitions.Def_BootRobust_Optimality_Setting
Formal statement
namespace BootRobust.Optimality
/-- B.2, p. 27: if `πΌ_D[G] β€ a < πΌ_{D'}[G]` for every `D` in the `R`-ball of radius `r` around `D_tr`
and every `D'` in a relatively open set `N`, then `int N = N` (interior relative to `πβ`) and `N` lies
in the disappointment set `β` of the nominal formulation with cost estimator `D β¦ πΌ_D[G]`. -/
theorem N_subset_disapp {ΞΉ : Type*} [Fintype ΞΉ] [DecidableEq ΞΉ]
(R : (ΞΉ β β) β (ΞΉ β β) β EReal) (Dtr : ΞΉ β β) (r : β)
(N : Set (ΞΉ β β)) (hN : relOpen N) (G : ΞΉ β β) (a : β)
(hball : β D β stdSimplex β ΞΉ, R D Dtr β€ (r : EReal) β β i, D i * G i β€ a)
(hNsep : β D' β N, a < β i, D' i * G i) :
relInt N = N β§ N β linDisapp R Dtr r G := by sorry
end BootRobust.Optimality
Source
Bertsimas and Van Parys, Bootstrap robust prescriptive analytics, arXiv:1711.09974v2, B.2 (proof of Proposition 1), p. 27, "Hence, int N = N β Rβ² β¦"
Human review
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.