Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Discounted Gittins retirement value

Definition
GittinsRetirementValue

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

banditsgittins-indexoptimal-stopping

For a single Markov reward process, define vγ(x)v_\gamma(x)vγ​(x) as the supremum expected discounted net reward when the player may either retire immediately for value zero or play for at least one round and stop at an adapted stopping time. This is the retirement-game value in Eq. (35.7), written with zero-indexed trajectories.

Definition code
import Definitions.Def_GittinsIndex

open MeasureTheory ProbabilityTheory

namespace BanditAlgorithm

/-- The value of the discounted single-arm retirement game at charge `γ`.
The player may retire immediately for value zero, or play for at least one
round and then stop at an adapted stopping time. This is Lattimore--
Szepesvári Eq. (35.7), in the same zero-indexed convention as
`discountedStoppedSum`. -/
noncomputable def gittinsRetirementValue
    {S : Type*} [MeasurableSpace S]
    (P : Kernel S S) [IsMarkovKernel P]
    (r : S → ℝ) (α γ : ℝ) (x : S) : ℝ :=
  sSup (insert 0 {v : ℝ |
    ∃ τ : (ℕ → S) → ℕ∞,
      IsTrajStoppingTime τ ∧ (∀ ω, 1 ≤ τ ω) ∧
      v = ∫ ω, discountedStoppedSum α (fun y ↦ r y - γ) τ ω
        ∂markovChainMeasure P x})

end BanditAlgorithm
Source
Lattimore and Szepesvári, Bandit Algorithms (CUP 2020), discounted retirement game, Eq. (35.7), printed p.448.

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