Theorem 1: finite limit almost surely and in mean, with E(ξ) = γ
ProvedKingmanSubadditive.Ergodic.theorem_1ergodic-theorylimit-theoremp2o-batch-pfp2ap2o-gran-per-chapterp2o-plan-paperp2o-v1subadditive-ergodic-theorem
Let be a subadditive process on a probability space , satisfying S₁, full joint stationarity S₂, and S₃. There is an integrable real random variable for which
and
This is Kingman's finite subadditive ergodic theorem. It gives the long-run growth rate even when the shift is not ergodic and the limit is therefore random.
Formalization Note Time is discrete and nonnegative. The limit is real valued; the distinct case allowing a limit of is Theorem 2, outside this mission.
Preamble
import Mathlib import Definitions.Def_KingmanSubadditive_Ergodic_Process
Formal statement
namespace KingmanSubadditive.Ergodic
open MeasureTheory Filter Topology
/-- Kingman, *Subadditive ergodic theory*, Ann. Probab. 1(6):883–899 (1973),
p. 885, Theorem 1, (1.2.1)–(1.2.2). `IsSubadditiveProcess` contains the
§1.1 standing conditions S₁, full joint-law stationarity S₂, and S₃.
The limit is real and integrable; convergence in mean is the `L¹`
conjunct, distinct from almost-sure convergence. -/
theorem theorem_1 {Ω : Type*} [MeasurableSpace Ω] (P : Measure Ω)
[IsProbabilityMeasure P] (x : ℕ → ℕ → Ω → ℝ)
(hx : IsSubadditiveProcess P x) :
∃ ξ : Ω → ℝ, Integrable ξ P ∧
(∀ᵐ ω ∂P, Tendsto (fun t : ℕ => x 0 t ω / (t : ℝ)) atTop (𝓝 (ξ ω))) ∧
Tendsto (fun t : ℕ => ∫ ω, |x 0 t ω / (t : ℝ) - ξ ω| ∂P) atTop (𝓝 0) ∧
∫ ω, ξ ω ∂P = gamma P x := by sorry
end KingmanSubadditive.Ergodic
Source
Kingman, Subadditive ergodic theory, Ann. Probab. 1(6):883–899 (1973), DOI 10.1214/aop/1176996798, p. 885, Theorem 1, (1.2.1)–(1.2.2)
Human review
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.