Lemma 7.1 (positive invariance) — for all
ProvedStochApproxDyn.Attractor.attainableSet_isPositivelyInvariantattainable-setp2o-batch-pfp1ap2o-gran-per-chapterp2o-plan-paperp2o-v1stochastic-approximation
Let be a semiflow on a metric space and let be a process on satisfying the standing assumptions of Section 7 (continuous paths, adapted to , condition (24) with a function ). Then the set of attainable points is positively invariant under :
This is the second assertion of Lemma 7.1: if the process can come close to at arbitrarily late times with positive probability, it can also come close to , because condition (24) makes it follow the flow for a time with high conditional probability.
Formalization Note The standing assumptions are those of the definition SatisfiesStandingAssumption; is a probability measure.
Preamble
import Mathlib import Definitions.Def_StochApproxDyn_Attractor_Dynamics import Definitions.Def_StochApproxDyn_Attractor_Process open scoped NNReal ENNReal open MeasureTheory
Formal statement
namespace StochApproxDyn.Attractor
/-- Lemma 7.1, second assertion (Benaïm 1999, p. 31): `Att(X)` is positively invariant under `Φ`. -/
theorem attainableSet_isPositivelyInvariant {Ω M : Type*} {m0 : MeasurableSpace Ω}
[MetricSpace M] [MeasurableSpace M] [BorelSpace M] (Φ : Flow ℝ≥0 M) (P : Measure Ω)
[IsProbabilityMeasure P] (ℱ : Filtration ℝ≥0 m0) (X : ℝ≥0 → Ω → M)
(w : ℝ≥0 → ℝ≥0 → ℝ≥0 → ℝ≥0) (hX : SatisfiesStandingAssumption Φ P ℱ X w) :
IsPositivelyInvariant Φ (attainableSet P X) := by sorry
end StochApproxDyn.Attractor
Source
Benaïm, Dynamics of Stochastic Approximation Algorithms, Séminaire de Probabilités XXXIII, LNM 1709 (1999), p. 31, Section 7.1, Lemma 7.1 (second assertion)
Human review
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.