Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Lipschitz stability of the DP optimum.

Proved
Logic.DPCompleteness.DPSpec.abs_val_sub_le

by raver1975 · Sep 12, 2026 · Mathlib c5ea003 (Lean v4.30.0)

aether-cataloglogic

Lipschitz stability of the DP optimum. If two specifications differ by at most a on the initial weights and at most b on every transition weight, then their value functions differ by at most a + n • b at horizon n.

theorem Logic.DPCompleteness.DPSpec.abs_val_sub_le{D D' : DPSpec S W} {a b : W}
    (hinit : ∀ s, |D'.init s - D.init s| ≤ a) (hstep : ∀ i s t, |D'.step i s t - D.step i s t| ≤ b)
    (n : ℕ) (s : S) : |D'.val n s - D.val n s| ≤ a + n • b := by sorry

Formalization Note Transplanted verbatim from the Aether Catalog source Logic/DPCompletenessStability.lean; the statement is byte-identical to the source declaration, elaborated with autoImplicit disabled in the platform environment.

Preamble
-- Thm stub generated from Logic/DPCompletenessStability.lean
import Mathlib
import Definitions.Def_Logic_DPCompleteness
import Definitions.Def_Logic_DPCompletenessStability
/-
# Stability of the DP value function under perturbation of the specification

The completeness theorem of `Logic.DPCompleteness` says the DP value is the exact optimum over
all labellings.  This file quantifies how that optimum reacts to perturbing the data.

* `DPSpec.shift` uniformly shifts the initial weights by `a` and every transition weight by `b`;
  `DPSpec.val_shift` shows the value function shifts by exactly `a + n • b` — the DP optimum is
  *equivariant* for the additive action of constants.
* `DPSpec.abs_val_sub_le` is the resulting Lipschitz stability bound: if two specifications
  differ by at most `a` on initial weights and at most `b` on transition weights, their value
  functions differ by at most `a + n • b` at horizon `n`.
* `DPSpec.near_optimal_transfer` turns this into a robustness statement about *runs*: a DP run
  computed for a perturbed model is within `2 • (a + n • b)` of the true optimum.  Combined with
  completeness this says the DP is not merely exact, but *stably* exact.

The weight monoid here is a linearly ordered additive commutative group (e.g. `ℤ`, `ℚ`, `ℝ`).
-/


open Logic.DPCompleteness

open DPSpec


variable {S W : Type*} [AddCommMonoid W] [Fintype S] [Nonempty S] [LinearOrder W] [AddLeftMono W]







variable {S W : Type*} [AddCommGroup W] [Fintype S] [Nonempty S] [LinearOrder W]
  [IsOrderedAddMonoid W]
Formal statement
theorem Logic.DPCompleteness.DPSpec.abs_val_sub_le{D D' : DPSpec S W} {a b : W}
    (hinit : ∀ s, |D'.init s - D.init s| ≤ a) (hstep : ∀ i s t, |D'.step i s t - D.step i s t| ≤ b)
    (n : ℕ) (s : S) : |D'.val n s - D.val n s| ≤ a + n • b := by sorry
Source
https://github.com/paulklemstine/Lean/blob/53c2925a02/Catalog/Logic/DPCompletenessStability.lean#L80

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