Theorem 5(b) — for D = ∏ₛ D(s), L_D V = V has a unique bounded solution, the robust value over Markov policies with rules in D
ProvedRobustDP.Discounted.theorem5b_fixed_pointLet be a discounted ambiguous MDP, and for every state let be nonempty; put . Define
the supremum over deterministic Markov policies with every . Then is bounded, , and is the only bounded solution of .
With this identifies the solution of the robust optimality equation with the robust value over deterministic Markov policies.
Formalization Note The page states Theorem 5(b) for an arbitrary set of decision rules. Its proof of the upper bound (28) picks one that is -greedy at every state simultaneously, which needs closed under state-wise pasting; for a non-product the statement fails (two states, ). It is stated here for product sets , which include both cases the paper uses ( and ). "Unique solution" is uniqueness among bounded functions.
import Mathlib import Definitions.Def_RobustDP_Discounted_Model import Definitions.Def_RobustDP_Discounted_Value import Definitions.Def_RobustDP_Discounted_Bellman
namespace RobustDP.Discounted
theorem theorem5b_fixed_point {S A : Type*} [Countable S] [Countable A] (M : Model S A)
(Dset : S → Set A) (hDne : ∀ s, (Dset s).Nonempty) (hDsub : ∀ s, Dset s ⊆ M.Aset s) :
let W : S → ℝ := fun s => ⨆ f : ℕ → productRules M Dset, V M (markovPolicy M (fun n => (f n).1)) s
IsBounded W ∧ L M (productRules M Dset) W = W ∧
∀ U : S → ℝ, IsBounded U → L M (productRules M Dset) U = U → U = W := by sorry
end RobustDP.Discounted
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.