Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Theorem 7 (the ergodic theorem) — a cumulative process satisfies w_t/t → κ₁/μ₁ with probability one

Proved
SmithRegenerative.Ergodic.theorem_7

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

cumulative-processergodic-theoremp2o-batch-pfp2ap2o-gran-per-chapterp2o-plan-paperp2o-v1regenerative-processrenewal-reward

Let t1,t2,…t_1, t_2, \dotst1​,t2​,… be a renewal process (independent, identically distributed, non-negative, P{t1=0}<1P\{t_1 = 0\} < 1P{t1​=0}<1) with delay t0=0t_0 = 0t0​=0 and regeneration epochs Tk=t1+⋯+tkT_k = t_1 + \dots + t_kTk​=t1​+⋯+tk​. Let wtw_twt​ be a cumulative process relative to it:

  1. (C1) the cycle increments yn=wTn−wTn−1y_n = w_{T_n} - w_{T_{n-1}}yn​=wTn​​−wTn−1​​, n≥1n \ge 1n≥1, are independent and identically distributed;
  2. (C2) with probability one, www is of bounded variation on every finite interval;
  3. the cycle variations y~n\tilde y_ny~​n​ (the total variation of www over [Tn−1,Tn][T_{n-1}, T_n][Tn−1​,Tn​]) are identically distributed.

Suppose μ1=Et1<∞\mu_1 = E t_1 < \inftyμ1​=Et1​<∞, κ~1=Ey~1<∞\tilde\kappa_1 = E \tilde y_1 < \inftyκ~1​=Ey~​1​<∞, and w0=0w_0 = 0w0​=0. Then, with κ1=Ey1\kappa_1 = E y_1κ1​=Ey1​,

lim⁡t→∞wtt=κ1μ1with probability one.\lim_{t \to \infty} \frac{w_t}{t} = \frac{\kappa_1}{\mu_1} \quad \text{with probability one.}t→∞lim​twt​​=μ1​κ1​​with probability one.

This is the ergodic theorem for regenerative processes: the long-run average rate of a quantity accumulated over a regenerative system equals the expected accumulation per cycle divided by the expected cycle length (the renewal–reward theorem in later terminology).

Formalization Note (C1) is read literally; no independence between cycle lengths and increments is assumed. μ1<∞\mu_1 < \inftyμ1​<∞ and κ~1<∞\tilde\kappa_1 < \inftyκ~1​<∞ are integrability hypotheses; μ1>0\mu_1 > 0μ1​>0 is a consequence of P{t1=0}<1P\{t_1 = 0\} < 1P{t1​=0}<1 and is not assumed separately; κ1\kappa_1κ1​ is finite because ∣y1∣≤y~1|y_1| \le \tilde y_1∣y1​∣≤y~​1​. The conditions t0=0t_0 = 0t0​=0 and w0=0w_0 = 0w0​=0 hold at every sample point. The limit is along real t→∞t \to \inftyt→∞.

Preamble
import Mathlib
import Definitions.Def_SmithRegenerative_Ergodic_Renewal
import Definitions.Def_SmithRegenerative_Ergodic_CumulativeProcess

open MeasureTheory ProbabilityTheory Filter Topology
Formal statement
namespace SmithRegenerative.Ergodic

/-- **Theorem 7 (the ergodic theorem)** (Smith, *Regenerative stochastic processes*, Proc. R.
Soc. Lond. A 232(1188):6–31 (1955), §5·3, p. 27): "If w_t is a cumulative process (satisfying
(C1) and (C2)); and if μ₁ < ∞ and κ̃₁ < ∞; and if t₀ = 0 and we define w₀ = 0, then
lim_{t=∞} w_t/t = κ₁/μ₁ with probability one."

Formalization Note: `t` is the general renewal process `t₀, t₁, …` (`IsRenewalProcess`: `t₁, t₂,
…` i.i.d., non-negative, `P{t₁ = 0} < 1`) with `t₀ = 0` at every sample point; `w` is a real
process with `w₀ = 0` at every sample point. `IsCumulativeProcess` is (C1) read literally (the
cycle increments `y_n = w_{T_n} − w_{T_{n−1}}` i.i.d.), (C2) (almost every path of bounded
variation on every `[0, s]`), and the cycle variations `ỹ_n` identically distributed (the
paper's notation `κ̃_r` presupposes it); no joint independence of cycle lengths and increments
is assumed. `μ₁ < ∞` is integrability of `t₁`, `κ̃₁ < ∞` integrability of `ỹ₁`;
`μ₁ = ∫ t₁ dP` is strictly positive under the hypotheses and is not assumed positive separately;
`κ₁ = ∫ y₁ dP` (`y₁` is integrable since `|y₁| ≤ ỹ₁` almost surely). The limit is along real
`t → ∞`. -/
theorem theorem_7 {Ω : Type*} [MeasurableSpace Ω] {P : Measure Ω} [IsProbabilityMeasure P]
    (t : ℕ → Ω → ℝ) (w : ℝ → Ω → ℝ) (hren : IsRenewalProcess P t)
    (ht0 : ∀ ω, t 0 ω = 0) (hw0 : ∀ ω, w 0 ω = 0) (hμ : Integrable (t 1) P)
    (hcum : IsCumulativeProcess P t w) (hκ : Integrable (cycleVariation t w 1) P) :
    ∀ᵐ ω ∂P, Tendsto (fun s : ℝ => w s ω / s) atTop
      (𝓝 ((∫ ω, cycleIncrement t w 1 ω ∂P) / ∫ ω, t 1 ω ∂P)) := by sorry

end SmithRegenerative.Ergodic
Source
Smith, Regenerative stochastic processes, Proc. R. Soc. Lond. A 232(1188):6–31 (1955), DOI 10.1098/rspa.1955.0198, p. 27, §5·3, Theorem 7
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