§2.2, p. 4 — if M(t) < ∞ near t = 0 and a > µ = E X, then I(a) > 0
ProvedKleywegtSAA.ExpRate.rateFn_poslarge-deviationsp2o-batch-pfp2ap2o-gran-per-chapterp2o-plan-paperp2o-v1rate-function
Let be an integrable real random variable on a probability space with mean , and suppose its moment generating function is finite for all in a neighbourhood of . Then for every ,
Combined with (2.4), this gives an exponentially small probability for every level above the mean.
Formalization Note "Finite in a neighbourhood of " is that is an interior point of the set of for which is integrable. Integrability of follows from it and is kept as a separate hypothesis for readability. is an extended real and may be .
Preamble
import Mathlib import Definitions.Def_KleywegtSAA_ExpRate_Setting open MeasureTheory ProbabilityTheory Filter Topology
Formal statement
namespace KleywegtSAA.ExpRate
/-- §2.2, p. 4: if the moment generating function of `Y` is finite in a neighbourhood of `0` and
`a > μ = E Y`, then `I(a) > 0`. -/
theorem rateFn_pos {Ω : Type*} [MeasurableSpace Ω] (P : Measure Ω) [IsProbabilityMeasure P]
(Y : Ω → ℝ) (hY : Integrable Y P) (hA : (0 : ℝ) ∈ interior (integrableExpSet Y P))
(a : ℝ) (ha : ∫ ω, Y ω ∂P < a) :
0 < rateFn P Y a := by sorry
end KleywegtSAA.ExpRate
Source
Kleywegt & Shapiro, The sample average approximation method for stochastic discrete optimization, preprint (two-author version, sha256 56657748…), p. 4, first paragraph ("Moreover, suppose ... It follows that in that case I(a) > 0")
Human review
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.