Proposition 9.7 — for lower semianalytic
OpenBertsekasShreve.BorelInfinite.Tbar_expect_eq_expect_Tborel-modeldynamic-programming
Let (SM) be an infinite horizon stochastic optimal control model on nonempty Borel spaces (Definitions 8.1, 9.1), with optimal cost over all universally measurable policies, and let (DM) be the associated deterministic model on (Definition 9.4) with optimal cost and operator . Assume one of (P) on , (N) on , (D) and on .
Let be lower semianalytic and satisfy , , or for some , according as (P), (N), or (D) is in force, and define for . Then
where .
This identity transports the DP operator of (DM) to that of (SM); its proof uses the -optimal selection theorem (Proposition 7.50) through Lemma 8.2.
Formalization Note All integrals are extended integrals with .
Preamble
import Mathlib import Definitions.Def_BertsekasShreve_BorelInfinite_Analytic import Definitions.Def_BertsekasShreve_BorelInfinite_ExtArith import Definitions.Def_BertsekasShreve_BorelInfinite_Model import Definitions.Def_BertsekasShreve_BorelInfinite_Policy import Definitions.Def_BertsekasShreve_BorelInfinite_DetModel open MeasureTheory ProbabilityTheory Filter Topology
Formal statement
namespace BertsekasShreve.BorelInfinite
/-- **Proposition 9.7** (p. 224). (P)(N)(D) Let `J : S → R*` be lower semianalytic and satisfy
`J ≥ 0`, `J ≤ 0`, or `−c ≤ J ≤ c` (`c < ∞`), according as (P), (N), or (D) is in force, and let
`J̄(p) = ∫ J(x) p(dx)`. Then `T̄(J̄)(p) = ∫ T(J)(x) p(dx)` for every `p ∈ P(S)`. -/
theorem Tbar_expect_eq_expect_T {S C W : Type*}
[TopologicalSpace S] [BertsekasShreve.AnalyticSelection.IsBorelSpace S] [MeasurableSpace S] [BorelSpace S] [Nonempty S]
[TopologicalSpace C] [BertsekasShreve.AnalyticSelection.IsBorelSpace C] [MeasurableSpace C] [BorelSpace C] [Nonempty C]
[TopologicalSpace W] [BertsekasShreve.AnalyticSelection.IsBorelSpace W] [MeasurableSpace W] [BorelSpace W] [Nonempty W]
(M : Model S C W)
(J : S → EReal) (hJ : IsLowerSemianalyticFun J)
(hcase : (M.CaseP ∧ ∀ x, 0 ≤ J x) ∨ (M.CaseN ∧ ∀ x, J x ≤ 0) ∨
(M.CaseD ∧ ∃ c : ℝ, ∀ x, -(c : EReal) ≤ J x ∧ J x ≤ (c : EReal)))
(p : Measure S) [IsProbabilityMeasure p] :
M.Tbar (fun p' => DupacovaWets.Consistency.expect p' J) p =
DupacovaWets.Consistency.expect p (M.T J) := by sorry
end BertsekasShreve.BorelInfinite
Source
Bertsekas & Shreve, Stochastic Optimal Control: The Discrete-Time Case, Athena Scientific 1996, p. 224–225, Proposition 9.7 of Chapter 9