(PSI) policy costs are universally measurable
OpenBertsekasShreve.ImperfectInfo.costHat_univMeasurableLet an (ISI) model with finite horizon be given, together with a statistic sufficient for control (Definition 10.6) and the associated perfect state information model (PSI) of Definition 10.7: states , control space , constraint sets , one-stage costs and transition kernels . For a (PSI) policy (universally measurable kernels with ), denotes its (0-originating) cost at (eq. (31)), and is the optimal cost (eq. (32)).
If or holds for (PSI), then for every policy the cost function
is universally measurable on .
This is the remark following Definition 8.3 (" is universally measurable in ", a consequence of Proposition 7.46), applied to the model (PSI). It is needed to integrate against in Propositions 10.2 and 10.3 and to evaluate the set in the definition of -optimality.
Formalization Note is σ.costHat μ (the stage-wise sum (31) with the convention ); universal measurability is UnivMeasurable.
import Mathlib import Definitions.Def_BertsekasShreve_ImperfectInfo_SufficientStatistic open MeasureTheory ProbabilityTheory
namespace BertsekasShreve.ImperfectInfo
theorem costHat_univMeasurable {S C Z : Type}
[TopologicalSpace S] [MeasurableSpace S] [BorelSpace S]
[TopologicalSpace C] [MeasurableSpace C] [BorelSpace C]
[TopologicalSpace Z] [MeasurableSpace Z] [BorelSpace Z]
(M : ISIModel S C Z) {Y : ℕ → Type} [∀ k, TopologicalSpace (Y k)] [∀ k, MeasurableSpace (Y k)]
(σ : SuffStat M Y)
(hF : σ.FhatPlus ∨ σ.FhatMinus)
(μ : (n : ℕ) → Obs Y C n → ProbabilityMeasure C) (hμ : σ.IsPolicy μ) :
UnivMeasurable (σ.costHat μ) := by sorry
end BertsekasShreve.ImperfectInfo