Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Proof of Theorem 6 — the multipliers pkp^kpk of (9) are a Bregman proximal point run on ∂(−d)\partial(-d)∂(−d)

Proved
BregmanPPA.EqMult.theorem6_ppa_run

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

bregman-functionmethod-of-multipliersp2o-batch-pfp2bp2o-gran-per-chapterp2o-plan-paperp2o-v1proximal-point

Consider problem (7) under its standing assumptions: X⊆RnX \subseteq \mathbb R^nX⊆Rn closed and convex, f:Rn→(−∞,+∞]f : \mathbb R^n \to (-\infty,+\infty]f:Rn→(−∞,+∞] convex and lower semicontinuous, inf⁡{f(x)∣x∈X}<∞\inf\{f(x)\mid x \in X\} < \inftyinf{f(x)∣x∈X}<∞, and the dual functional d(p)=inf⁡x∈X{f(x)+⟨p,Ax−b⟩}d(p) = \inf_{x\in X}\{f(x) + \langle p, Ax - b\rangle\}d(p)=infx∈X​{f(x)+⟨p,Ax−b⟩} not identically −∞-\infty−∞. Let hhh be a Bregman function with zone Rm\mathbb R^mRm and im⁡(∇h)=Rm\operatorname{im}(\nabla h) = \mathbb R^mim(∇h)=Rm, let ck>0c_k > 0ck​>0, and let {(xk,pk)}\{(x^k,p^k)\}{(xk,pk)} conform to recursion (9),

xk+1∈arg min⁡x∈X{f(x)+1ckh∗(∇h(pk)+ck(Ax−b))},pk+1=∇h∗(∇h(pk)+ck(Axk+1−b)).x^{k+1} \in \operatorname*{arg\,min}_{x\in X}\Bigl\{ f(x) + \tfrac{1}{c_k} h^*\bigl(\nabla h(p^k) + c_k(Ax - b)\bigr)\Bigr\},\qquad p^{k+1} = \nabla h^*\bigl(\nabla h(p^k) + c_k(Ax^{k+1} - b)\bigr).xk+1∈x∈Xargmin​{f(x)+ck​1​h∗(∇h(pk)+ck​(Ax−b))},pk+1=∇h∗(∇h(pk)+ck​(Axk+1−b)).

Then the multiplier sequence conforms to the Bregman proximal point recursion for ∂(−d)\partial(-d)∂(−d):

pk+1=(∇h+ck ∂(−d))−1(∇h(pk))for all k,p^{k+1} = \bigl(\nabla h + c_k\, \partial(-d)\bigr)^{-1}\bigl(\nabla h(p^k)\bigr)\qquad\text{for all } k,pk+1=(∇h+ck​∂(−d))−1(∇h(pk))for all k,

that is, 1ck(∇h(pk)−∇h(pk+1))∈∂(−d)(pk+1)\frac{1}{c_k}\bigl(\nabla h(p^k) - \nabla h(p^{k+1})\bigr) \in \partial(-d)(p^{k+1})ck​1​(∇h(pk)−∇h(pk+1))∈∂(−d)(pk+1) for every kkk.

This identity reduces the convergence of the nonquadratic method of multipliers to Theorem 1 applied to the maximal monotone operator ∂(−d)\partial(-d)∂(−d).

Formalization Note The recursion is encoded in form (4), as in the definition of runs; the zone is all of Rm\mathbb R^mRm, so the condition pk∈Sp^k \in Spk∈S is automatic. ∂(−d)\partial(-d)∂(−d) is the subdifferential of the EReal function p↦−d(p)p \mapsto -d(p)p↦−d(p). im⁡∇h=Rm\operatorname{im}\nabla h = \mathbb R^mim∇h=Rm is surjectivity of ∇h\nabla h∇h. The paper states this claim in the proof of Theorem 6, under the standing assumptions of §4.1, which are all hypotheses here.

Preamble
import Mathlib
import Definitions.Def_InertialFB_IFB_ConvexAnalysis
import Definitions.Def_fenchelConjugate
import Definitions.Def_BregmanPPA_Convergence_Model
import Definitions.Def_BregmanPPA_EqMult_EqConstrainedProgram

open InertialFB.IFB ConvexOptimization Filter Topology
Formal statement
namespace BregmanPPA.EqMult
theorem theorem6_ppa_run {n m : ℕ}
    (A : EuclideanSpace ℝ (Fin n) →L[ℝ] EuclideanSpace ℝ (Fin m))
    (b : EuclideanSpace ℝ (Fin m)) (X : Set (EuclideanSpace ℝ (Fin n)))
    (f : EuclideanSpace ℝ (Fin n) → EReal)
    (hP : IsEqProgram X f) (hd : ∃ p, dualEq A b X f p ≠ ⊥)
    (h : EuclideanSpace ℝ (Fin m) → ℝ) (hh : BregmanPPA.Convergence.IsBregmanFunction Set.univ h)
    (him : Function.Surjective (gradient h))
    (c : ℕ → ℝ) (hc : ∀ k, 0 < c k)
    (x : ℕ → EuclideanSpace ℝ (Fin n)) (p : ℕ → EuclideanSpace ℝ (Fin m))
    (hrun : IsEqMultRun A b X f h c x p) :
    BregmanPPA.Convergence.IsBregmanPPARun Set.univ h (BregmanPPA.Convergence.subdiffOp (negDual A b X f)) c p := by sorry
end BregmanPPA.EqMult
Source
Eckstein, Nonlinear proximal point algorithms using Bregman functions, with applications to convex programming, Math. Oper. Res. 18(1) (1993), p. 213, proof of Theorem 6 (from "We show that the sequence {p^k} produced by (9) conforms to the Bregman nonlinear proximal point formula" to "Thus, p^{k+1} = (∇h + c_k∂(−d))^{-1}(∇h(p^k)), as claimed.")
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