Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Finite-horizon Gittins retirement values

Definition
GittinsFiniteRetirementValue

by Harry_Xu · Jul 31, 2026 · Mathlib 0df444a (Lean v4.33.1)

banditsdynamic-programmingoptimal-stopping

The finite-horizon Bellman approximants for the discounted retirement game: horizon zero has value zero, while horizon n+1n+1n+1 takes the maximum of immediate retirement and one charged reward plus the discounted expected horizon-nnn continuation value.

Definition code
import Mathlib.Probability.Kernel.MeasurableIntegral
import Definitions.Def_GittinsRetirementValue

open MeasureTheory ProbabilityTheory

namespace BanditAlgorithm

/-- Finite-horizon Bellman approximants for the discounted retirement game.
At horizon zero the value is zero; at horizon `n+1`, the player either retires
for zero or plays once and receives the continuation value at horizon `n`. -/
noncomputable def gittinsFiniteRetirementValue
    {S : Type*} [MeasurableSpace S]
    (P : Kernel S S) (r : S → ℝ) (α γ : ℝ) : ℕ → S → ℝ
  | 0, _ => 0
  | n + 1, x =>
      max 0 (r x - γ +
        α * ∫ y, gittinsFiniteRetirementValue P r α γ n y ∂P x)

end BanditAlgorithm
Source
Lattimore and Szepesvári, Bandit Algorithms (CUP 2020), Theorem 35.3 on printed pp.442--443 and Lemma 35.7(a) on printed p.449.

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