Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Lemma 1 — convergence of a scalar stochastic approximation recursion

Proved
AsyncSA.Contraction.lemma1

by mikedeng1 · Oct 6, 2026 · Mathlib 0df444a (Lean v4.33.1)

martingalep2o-batch-pfp2bp2o-gran-per-chapterp2o-plan-paperp2o-v1stochastic-approximation

Let F(t)\mathcal F(t)F(t) be an increasing filtration and let α(t)\alpha(t)α(t), w(t)w(t)w(t), and B(t)B(t)B(t) be scalar random variables, with α(t)\alpha(t)α(t) and B(t)B(t)B(t) measurable at time ttt and w(t)w(t)w(t) measurable at time t+1t+1t+1. Suppose w(t)w(t)w(t) has conditional mean zero and conditional second moment at most B(t)B(t)B(t). Almost surely, let 0≤α(t)≤10\le\alpha(t)\le10≤α(t)≤1, ∑tα(t)=∞\sum_t\alpha(t)=\infty∑t​α(t)=∞, ∑tα(t)2≤C\sum_t\alpha(t)^2\le C∑t​α(t)2≤C for one deterministic CCC, and let the path B(t)B(t)B(t) be bounded. If WWW satisfies

W(t+1)=(1−α(t))W(t)+α(t)w(t),W(t+1)=(1-\alpha(t))W(t)+\alpha(t)w(t),W(t+1)=(1−α(t))W(t)+α(t)w(t),

then W(t)→0W(t)\to0W(t)→0 almost surely. The initial scalar W(0)W(0)W(0) may be any finite random value.

This lemma controls the scalar noise recursion used in both convergence and boundedness arguments.

Formalization Note Conditional moments are generalized measurable-set statements, and both step-size series are encoded by partial sums. The paper's notation says w(t−1)w(t-1)w(t−1) is F(t)\mathcal F(t)F(t)-measurable; the equivalent nonnegative-index condition is that w(t)w(t)w(t) is F(t+1)\mathcal F(t+1)F(t+1)-measurable.

Preamble
import Mathlib
import Definitions.Def_AsyncSA_Contraction_Model
Formal statement
namespace AsyncSA.Contraction

open MeasureTheory Filter Topology

/-- Lemma 1, p. 190. The recursion may start from any finite random value. -/
theorem lemma1 {Ω : Type*} [m₀ : MeasurableSpace Ω]
    (P : Measure Ω) [IsProbabilityMeasure P]
    (𝓕 : Filtration ℕ m₀) (α w B W : ℕ → Ω → ℝ) (C : ℝ)
    (hαmeas : ∀ t, Measurable[𝓕 t] (α t))
    (hwmeas : ∀ t, Measurable[𝓕 (t + 1)] (w t))
    (hBmeas : ∀ t, Measurable[𝓕 t] (B t))
    (hmean : ∀ t, CondMeanZero m₀ (𝓕 t) P (w t))
    (hsq : ∀ t, CondSqLe m₀ (𝓕 t) P (w t) (B t))
    (hαmem : ∀ᵐ ω ∂P, ∀ t, α t ω ∈ Set.Icc (0 : ℝ) 1)
    (hdiv : ∀ᵐ ω ∂P,
      Tendsto (fun T => ∑ t ∈ Finset.range T, α t ω) atTop atTop)
    (hsum : ∀ᵐ ω ∂P,
      ∀ T, ∑ t ∈ Finset.range T, α t ω ^ 2 ≤ C)
    (hBbound : ∀ᵐ ω ∂P, ∃ K : ℝ, ∀ t, |B t ω| ≤ K)
    (hW : ∀ t ω, W (t + 1) ω =
      (1 - α t ω) * W t ω + α t ω * w t ω) :
    ∀ᵐ ω ∂P, Tendsto (fun t => W t ω) atTop (𝓝 0) := by sorry

end AsyncSA.Contraction
Source
Tsitsiklis, Asynchronous Stochastic Approximation and Q-Learning, Machine Learning 16 (1994), p. 190, §3, Lemma 1
Human review
  • Endorsed by Shuze Chen · Oct 6, 2026

    Confirmed by the moderator at approval.

  • Endorsed by mikedeng1 · Oct 6, 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