Harris ergodic chains are strongly mixing: (Jones Thm 2(i))
ProvedMarkovChainCLT.strongly_mixing_of_harrisLet be a Markov chain with transition kernel , Harris ergodic with invariant probability , and consider its stationary version (initial distribution ). Then the chain is strongly mixing:
where is the strong mixing coefficient of the coordinate process.
This is the bridge that makes the entire classical theory of CLTs for mixing sequences applicable to Markov chain Monte Carlo.
Formalization Note "Harris ergodic" is encoded by its total-variation characterization: is invariant for and for every starting point (equivalent to the classical aperiodic, -irreducible, positive Harris recurrent definition; the "every " quantifier is exactly the Harris property). Convergence in distribution is weak convergence of laws, and is read as the point mass at , which absorbs the source's "" caveat. The -algebra of the state space is additionally assumed countably generated, the standard general-state-space setting of Meyn and Tweedie.
import Definitions.Def_MarkovErgodicity import Definitions.Def_MarkovChainPathMeasure import Definitions.Def_MixingCoefficients open MeasureTheory ProbabilityTheory Filter open scoped ENNReal NNReal Topology ProbabilityTheory /-- **Theorem 2, part 1**: the stationary version of a Harris ergodic chain is strongly mixing: `α(n) → 0`. -/
theorem MarkovChainCLT.strongly_mixing_of_harris {X : Type*} [MeasurableSpace X]
[MeasurableSpace.CountablyGenerated X]
(P : Kernel X X) [IsMarkovKernel P] (π : Measure X) [IsProbabilityMeasure π]
(hP : HarrisErgodic P π) :
Tendsto (fun n => alphaMixingCoef (chainMeasure P π) (fun i ω => ω i) n)
atTop (𝓝 0) := by sorry