Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Corollary 3.1, p. 6 — approximate value iteration with a nonexpansive M_F converges in max norm at the rate γ

Proved
StableFA.Discounted.corollary_3_1

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

approximate-value-iterationcontractionconvergence-ratefunction-approximationp2o-batch-pfp2ap2o-gran-per-chapterp2o-plan-paperp2o-v1

Let MMM, γ\gammaγ, TMT_MTM​ and MFM_FMF​ be as in Theorem 3.1: a finite MDP on nnn states with stochastic transition rows, discount 0≤γ<10\le\gamma<10≤γ<1, its parallel value backup operator TMT_MTM​, and a max-norm nonexpansion MFM_FMF​ on Rn\mathbb R^nRn. The approximate value iteration algorithm based on FFF produces Vk+1=MF(TM(Vk))V_{k+1}=M_F(T_M(V_k))Vk+1​=MF​(TM​(Vk​)). There is a value function V∞V_\inftyV∞​ such that, from every initial guess V0V_0V0​, the iterates converge to V∞V_\inftyV∞​ at the geometric rate γ\gammaγ:

∥(MF∘TM)k(V0)−V∞∥≤γk ∥V0−V∞∥for all k≥0,\|(M_F\circ T_M)^k(V_0)-V_\infty\|\le\gamma^k\,\|V_0-V_\infty\|\qquad\text{for all }k\ge0,∥(MF​∘TM​)k(V0​)−V∞​∥≤γk∥V0​−V∞​∥for all k≥0,

and in particular (MF∘TM)k(V0)→V∞(M_F\circ T_M)^k(V_0)\to V_\infty(MF​∘TM​)k(V0​)→V∞​ as k→∞k\to\inftyk→∞.

The corollary turns the contraction of Theorem 3.1 into a quantitative convergence statement for the algorithm actually run in practice.

Formalization Note "Converges at the rate γ\gammaγ" is read as the geometric rate of the Contraction Mapping theorem (Theorem 2.1, p. 2): the error after kkk steps is at most γk\gamma^kγk times the initial error, measured from the common limit V∞V_\inftyV∞​ (called Vinf in Lean). ∥⋅∥\|\cdot\|∥⋅∥ on Fin n → ℝ is the max norm; γ0=1\gamma^0=1γ0=1, so the case k=0k=0k=0 is trivial.

Preamble
import Mathlib
import Definitions.Def_BertsekasSSPModel
import Definitions.Def_StableFA_Discounted_Setting

open Filter Topology
Formal statement
namespace StableFA.Discounted

theorem corollary_3_1 {n : ℕ} {C : Type} [Fintype C] (M : BertsekasSSPModel n C)
    (γ : ℝ) (hγ0 : 0 ≤ γ) (hγ1 : γ < 1)
    (hp1 : ∀ i, ∀ u ∈ M.U i, ∑ j, M.p i u j = 1)
    (MF : (Fin n → ℝ) → (Fin n → ℝ)) (hMF : ∀ Y Z : Fin n → ℝ, ‖MF Y - MF Z‖ ≤ ‖Y - Z‖) :
    ∃ Vinf : Fin n → ℝ,
      (∀ V : Fin n → ℝ,
        Tendsto (fun k => (MF ∘ BertsekasDiscountedBellmanOp M γ)^[k] V) atTop (𝓝 Vinf)) ∧
      ∀ (V : Fin n → ℝ) (k : ℕ),
        ‖(MF ∘ BertsekasDiscountedBellmanOp M γ)^[k] V - Vinf‖ ≤ γ ^ k * ‖V - Vinf‖ := by sorry

end StableFA.Discounted
Source
Gordon, Stable Function Approximation in Dynamic Programming, Tech. Rep. CMU-CS-95-103, Carnegie Mellon University (1995), p. 6, Corollary 3.1 (rate as in Theorem 2.1, p. 2)
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