Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Theorem 4 — every limit point of policy improvement is a rule of C'' optimal over C

Proved
DermanDenumerable.AvgCost.theorem4

by mikedeng1 · Oct 5, 2026 · Mathlib 0df444a (Lean v4.33.1)

average-costdynamic-programmingp2o-batch-pfp2bp2o-gran-per-chapterp2o-plan-paperp2o-v1policy-improvement

Assume conditions (A), (B), (C), (E) and (F) for a Markovian decision process on a denumerable state space III with finitely many decisions per state and bounded costs. Let {gR,vjR}\{g^R, v_j^R\}{gR,vjR​} be the uniformly bounded solutions of (2) given by (E), and let R1,R2,…R_1, R_2, \dotsR1​,R2​,… be any sequence of policy improvement iterations starting from an arbitrary R1∈C′′R_1 \in C''R1​∈C′′: the decisions of Rn+1R_{n+1}Rn+1​ minimize

wik+∑j∈Iqij(k)vjRnw_{ik} + \sum_{j\in I} q_{ij}(k) v_j^{R_n}wik​+j∈I∑​qij​(k)vjRn​​

at every state iii. Then:

  1. the sequence has a limit point R∗∈C′′R^* \in C''R∗∈C′′, i.e. a subsequence RnνR_{n_\nu}Rnν​​ whose decision at each state equals that of R∗R^*R∗ for all large ν\nuν;
  2. every limit point R∗R^*R∗ is optimal over CCC: QR∗(i)≤QR(i)Q_{R^*}(i) \le Q_R(i)QR∗​(i)≤QR​(i) for every rule R∈CR \in CR∈C and every i∈Ii \in Ii∈I;
  3. for every limit point R∗R^*R∗ and every iii, gRn→QR∗(i)g^{R_n} \to Q_{R^*}(i)gRn​→QR∗​(i) as n→∞n \to \inftyn→∞.

This is the paper's main theorem: on a denumerable state space, the policy improvement procedure "converges to a rule R∗∈CR^* \in CR∗∈C which is optimal over CCC".

Formalization Note. The paper's "converges" is read as items 1–2: its proof chooses a subsequence converging to R∗R^*R∗ and never shows that the whole sequence converges, which can fail when the minimization has ties. Item 3 records the convergence of the values gRng^{R_n}gRn​ that the proof establishes (gRn↓g∗=QR∗g^{R_n} \downarrow g^* = Q_{R^*}gRn​↓g∗=QR∗​). Condition (D) is implied by (E) and is not a separate hypothesis. The sequence is indexed from 000: R 0 is the paper's R1R_1R1​.

Preamble
import Mathlib
import Definitions.Def_DermanDenumerable_AvgCost_Model
open scoped Topology
open Filter SennottDP.AvgFinite
Formal statement
namespace DermanDenumerable.AvgCost

/-- Derman (1966), Theorem 4, p. 1552: under (A), (B), (C), (E) (hence (D)) and (F), let `R_1, R_2, ⋯`
be any sequence of policy improvement iterations (relative to the family `{g^R, v^R}` of (E)) from an
arbitrary initial rule of `C''`. Then (i) the sequence has a limit point `R*` in `C''`; (ii) every
limit point `R*` is optimal over the class `C` of all rules; and (iii) `g^{R_n}` converges to the
optimal average cost `Q_{R*}(i)`. -/
theorem theorem4 {S Act : Type} [Countable S] (M : MDC S Act)
    (w : S → Act → ℝ) (hB : CostBounded M w) (hC : CondC M) (gR : StationaryPolicy M → ℝ)
    (vR : StationaryPolicy M → S → ℝ) (hE : IsUniformEvaluation M w gR vR) (hF : CondF M)
    (R : ℕ → StationaryPolicy M) (hR : IsPISequence M w vR R) :
    (∃ e : StationaryPolicy M, IsLimitPoint R e) ∧
      ∀ e : StationaryPolicy M, IsLimitPoint R e →
        OptimalOverC w e.toPolicy ∧
          ∀ i : S, Tendsto (fun n : ℕ => gR (R n)) atTop (𝓝 (avgCostR e.toPolicy w i)) := by sorry

end DermanDenumerable.AvgCost
Source
Derman, Denumerable State Markovian Decision Processes—Average Cost Criterion, Ann. Math. Statist. 37(6) (1966), p. 1552, Theorem 4 (procedure and (F) on p. 1551)
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