Corollary 7.5.10 — (CAV*) implies (CAV), hence (BOR)
ProvedSennottDP.BOR.cav_of_irreducible_standardaverage-costmarkov-chainp2o-batch-b23ap2o-gran-per-chapterp2o-plan-bookp2o-v1
Assume the (CAV*) assumptions: (CAV1) there is a standard policy — standard for a state — whose positive recurrent class is all of , ; (CAV2) for every the set is finite. Then the (CAV) assumptions hold for and , and hence (BOR) is valid for , and some .
Formalization Note "Standard" is read as " standard for some state "; the conclusion is stated for that same .
Preamble
import Mathlib import Definitions.Def_SennottDP_BOR_Assumptions open scoped ENNReal NNReal open Filter Topology
Formal statement
namespace SennottDP.BOR
/-- Sennott (1999), Corollary 7.5.10, p. 149. Assume (CAV*): (CAV*1) there is a standard policy
`d` (here: `z` standard for the state `z`) with `R_d = S`; (CAV*2) for every `U > 0` the set
`D_U = {i | C(i,a) ≤ U for some a}` is finite. Then the (CAV) assumptions hold, and hence (BOR)
is valid (for `z`, `d` and some `ε > 0`). -/
theorem cav_of_irreducible_standard {S Act : Type} [Countable S] (M : SennottDP.Discounted.MDC S Act) (z : S)
(d : RandStationaryPolicy M) (hd : IsZStandard d.toPolicy z)
(hR : recClass M d z = Set.univ)
(hD : ∀ U : ℝ, 0 < U → (lowCostSet M (ENNReal.ofReal U)).Finite) :
CAVAssumptions M z d ∧ ∃ ε : ℝ, BORAssumptions M z d ε := by sorry
end SennottDP.BOR
Source
Sennott, Stochastic Dynamic Programming and the Control of Queueing Systems (Wiley, 1999), p. 149, Corollary 7.5.10
Human review
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.