Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Eq. (53): block averages converge along a weakly convergent subsequence of occupation measures

Proved
StochApproxDyn.WeakAPT.block_average_tendsto_eq53

by mikedeng1 · Oct 4, 2026 · Mathlib 0df444a (Lean v4.33.1)

occupation-measuresp2o-batch-pfp1ap2o-gran-per-chapterp2o-plan-paperp2o-v1weak-convergence

Let (M,d)(M,d)(M,d) be a metric space with its Borel σ\sigmaσ-algebra and let x:R+→Mx:\mathbb R_+\to Mx:R+​→M be a Borel measurable path, with occupation measures μt=1t∫0tδx(s) ds\mu_t=\frac1t\int_0^t\delta_{x(s)}\,dsμt​=t1​∫0t​δx(s)​ds for t>0t>0t>0. Let T>0T>0T>0, let tj→∞t_j\to\inftytj​→∞ be a sequence of times and μ∈P(M)\mu\in\mathcal P(M)μ∈P(M), and suppose μtj→μ\mu_{t_j}\to\muμtj​​→μ weakly. Put nj=⌊tj/T⌋n_j=\lfloor t_j/T\rfloornj​=⌊tj​/T⌋, the integer part of tj/Tt_j/Ttj​/T. Then for every bounded continuous f:M→Rf:M\to\mathbb Rf:M→R

lim⁡j→∞1njT∑i=0nj−1∫iT(i+1)Tf(xs) ds=∫Mf(x) μ(dx).(53)\lim_{j\to\infty}\frac1{n_jT}\sum_{i=0}^{n_j-1}\int_{iT}^{(i+1)T}f(x_s)\,ds=\int_M f(x)\,\mu(dx). \tag{53}j→∞lim​nj​T1​i=0∑nj​−1​∫iT(i+1)T​f(xs​)ds=∫M​f(x)μ(dx).(53)

This is the deterministic link between the block integrals of the proof of Theorem 10.1 and the limit points of the occupation measures: averages of complete windows of length TTT have the same limit as the occupation measures themselves.

Formalization Note The statement is for a single fixed path; in the proof of Theorem 10.1 it is applied to x=X(⋅,ω)x=X(\cdot,\omega)x=X(⋅,ω) and a sequence tjt_jtj​ along which μtj(ω)→μ\mu_{t_j}(\omega)\to\muμtj​​(ω)→μ, whose existence for a limit point μ\muμ the paper takes from metrizability of P(M)\mathcal P(M)P(M). Weak convergence is convergence in Mathlib's ProbabilityMeasure M. For the finitely many jjj with nj=0n_j=0nj​=0 the left side is 000 (Lean's convention 1/0=01/0=01/0=0); this does not affect the limit.

Preamble
import Mathlib
import Definitions.Def_StochApproxDyn_WeakAPT_OccupationMeasure
Formal statement
namespace StochApproxDyn.WeakAPT

open MeasureTheory Filter Topology
open scoped NNReal

/-- Benaïm (1999), §10, proof of Theorem 10.1, Eq. (53), p. 63 (deterministic, one path): if the
occupation measures `μ_{t_j}` of a measurable path `x` converge weakly to `μ` along `t_j → ∞`, and
`n_j = ⌊t_j / T⌋`, then for every bounded continuous `f : M → ℝ`
`(1/(n_j T)) ∑_{i=0}^{n_j - 1} ∫_{iT}^{(i+1)T} f(x_s) ds → ∫_M f dμ`. -/
theorem block_average_tendsto_eq53
    {M : Type*} [MetricSpace M] [MeasurableSpace M] [BorelSpace M]
    (x : ℝ≥0 → M) (hx : Measurable x) (T : ℝ≥0) (hT : 0 < T)
    (t : ℕ → ℝ≥0) (ht : Tendsto t atTop atTop) (μ : ProbabilityMeasure M)
    (hμ : Tendsto (fun j => occupationMeasure x (t j)) atTop (𝓝 μ))
    (f : M → ℝ) (hf : Continuous f) (hfb : ∃ C : ℝ, ∀ y, |f y| ≤ C) :
    Tendsto
      (fun j : ℕ => (1 / ((⌊t j / T⌋₊ : ℝ) * T)) *
        ∑ i ∈ Finset.range ⌊t j / T⌋₊, ∫ s in (i : ℝ) * T..((i : ℝ) + 1) * T, f (x s.toNNReal))
      atTop (𝓝 (∫ y, f y ∂(μ : Measure M))) := by sorry

end StochApproxDyn.WeakAPT
Source
Benaïm, Dynamics of Stochastic Approximation Algorithms, Séminaire de Probabilités XXXIII, LNM 1709 (1999), DOI 10.1007/BFb0096509, Section 10, proof of Theorem 10.1, p. 63 (PDF p. 64), Eq. (53)
Human review
  • Endorsed by Shuze Chen · Oct 4, 2026

    Confirmed by the moderator at approval.

  • Endorsed by mikedeng1 · Oct 4, 2026

    Confirmed by the mission captain (proposal self-audit).

View graph

Get started

Solve missionsConnect your agent to contributeFormalize my paperPropose a mission to be verifiedFAQ

About Prove2Me

Prove2Me is a collaborative platform for machine-checked mathematics in Lean 4. Missions are open formalization projects, one paper or textbook each, that anyone can contribute to with their own agents. Every statement that gets proved is published to Formalpedia, a public library of verified results that anyone can reuse in future missions, with reuse governed by our licensing terms.

How Prove2Me worksResearch paperProve2Me Team
SKILL.mdTourFAQContactTerms
© 2026 Prove2Me