Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

§7.3.1, p. 88 — the future state distribution dominates the start distribution: dπ,μ≥μ/Hd_{\pi,\mu}\ge\mu/Hdπ,μ​≥μ/H

Proved
SampleComplexityRL.CPI.future_dist_ge_start

by mikedeng1 · Oct 7, 2026 · Mathlib 0df444a (Lean v4.33.1)

dynamic-programmingp2o-batch-pfp2ap2o-gran-per-chapterp2o-plan-bookp2o-v1reinforcement-learning

Let SSS and AAA be finite nonempty sets, P(⋅∣s,a)P(\cdot\mid s,a)P(⋅∣s,a) a transition kernel on SSS, 0≤γ<10\le\gamma<10≤γ<1, π\piπ a stationary policy and μ\muμ a distribution on SSS. Let H=1/(1−γ)H=1/(1-\gamma)H=1/(1−γ) be the horizon time and dπ,μ(s)=(1−γ)∑t≥0γtPr⁡(st=s∣π,s0∼μ)d_{\pi,\mu}(s)=(1-\gamma)\sum_{t\ge0}\gamma^t\Pr(s_t=s\mid\pi,s_0\sim\mu)dπ,μ​(s)=(1−γ)∑t≥0​γtPr(st​=s∣π,s0​∼μ) the discounted future state distribution. Then for every state sss,

dπ,μ(s) ≥ μ(s)H=(1−γ) μ(s).d_{\pi,\mu}(s)\ \ge\ \frac{\mu(s)}{H}=(1-\gamma)\,\mu(s).dπ,μ​(s) ≥ Hμ(s)​=(1−γ)μ(s).

The future distribution keeps a μ/H\mu/Hμ/H contribution from the starting distribution. This converts a guarantee on future advantages, which weight states by dπ,μd_{\pi,\mu}dπ,μ​, into information about states weighted by μ\muμ.

Formalization Note The point of the statement is pointwise domination by the time-000 term of the series; no reward function enters.

Preamble
import Mathlib
import Definitions.Def_FoundationsML_ReinforcementLearning_IsTransitionKernel
import Definitions.Def_FoundationsML_ReinforcementLearning_IsPolicy
import Definitions.Def_ApproxOptRL_Shared_Model
import Definitions.Def_ApproxOptRL_CPI_PolicyAdvantage
import Definitions.Def_SampleComplexityRL_CPI_ExactCPI
open ApproxOptRL.Shared ApproxOptRL.CPI FoundationsML.ReinforcementLearning
Formal statement
namespace SampleComplexityRL.CPI

/-- Kakade 2003, §7.3.1, p. 88: the future state distribution dominates the start distribution,
`d_{π,μ} ≥ μ/H` with `H = 1/(1 − γ)`, i.e. `d_{π,μ}(s) ≥ (1 − γ) μ(s)` for every state `s`. -/
theorem future_dist_ge_start {S A : Type} [Fintype S] [DecidableEq S] [Fintype A]
    [Nonempty S] [Nonempty A]
    (P : S → A → S → ℝ) (hP : IsTransitionKernel P)
    (γ : ℝ) (hγ0 : 0 ≤ γ) (hγ1 : γ < 1)
    (π : S → A → ℝ) (hπ : IsPolicy π)
    (μ : S → ℝ) (hμ : IsStateDist μ) :
    ∀ s : S, μ s / horizon γ ≤ futureStateDist P γ π μ s := by sorry

end SampleComplexityRL.CPI
Source
Kakade, On the Sample Complexity of Reinforcement Learning, PhD thesis, University College London, 2003, p. 88, Section 7.3.1 (the fact d_{π,μ} ≥ μ/H)
Human review
  • Endorsed by Shuze Chen · Oct 7, 2026

    Confirmed by the moderator at approval.

  • Endorsed by mikedeng1 · Oct 7, 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