Proposition 2.2, (2.5), p. 5 — f′(x, d) = E_P h′_ω(x, d), with h′_ω(x, d) integrable
ProvedSAARate.Sharp.prop_2_2_eq_2_5convex-analysisdirectional-derivativep2o-batch-pfp2ap2o-gran-per-chapterp2o-plan-paperp2o-v1stochastic-programming
Let be a probability measure on and . Suppose that
- for every the function is convex;
- the expected value function is well defined and finite valued.
Then for all the function is -integrable and
where is the directional derivative of at in the direction and that of .
Exchanging the directional derivative with the expectation is what lets the strong law of large numbers act on directional derivatives in (2.6).
Formalization Note. Assumption 2 is encoded as integrability of for every (which includes its measurability). The integrability of is part of the conclusion: it is the page's "the right hand side of (2.5) is well defined" (proof of Proposition 2.2, p. 5), and without it the Bochner integral of a non-integrable function would be .
Preamble
import Mathlib import Definitions.Def_SAARate_Sharp_Setting
Formal statement
namespace SAARate.Sharp
open MeasureTheory ProbabilityTheory Filter Topology
theorem prop_2_2_eq_2_5 {m : ℕ} {Ω : Type*} [MeasurableSpace Ω] (P : Measure Ω)
[IsProbabilityMeasure P] (h : E m → Ω → ℝ)
(hconv : ∀ ω, ConvexOn ℝ Set.univ (fun x => h x ω))
(hint : ∀ x, Integrable (h x) P) :
∀ x d : E m, Integrable (fun ω => dirDeriv (fun y => h y ω) x d) P ∧
dirDeriv (expectedObj P h) x d = ∫ ω, dirDeriv (fun y => h y ω) x d ∂P := by sorry
end SAARate.Sharp
Source
Shapiro & Homem-de-Mello, On Rate of Convergence of Optimal Solutions of Monte Carlo Approximations of Stochastic Programs, preprint (SPEPS copy, edoc.hu-berlin.de), p. 5, Proposition 2.2, (2.5)
Human review
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.