Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Consensus ADMM: yˉk+1=0\bar y^{k+1}=0yˉ​k+1=0, zk=xˉkz^k=\bar x^kzk=xˉk and the simplified iteration

Proved
BoydADMM.Consensus.consensus_dual_average_zero

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

admmconsensusdual-variablesp2o-batch-pfp2bp2o-gran-per-chapterp2o-plan-bookp2o-v1

Let xik,zk,yikx_i^k, z^k, y_i^kxik​,zk,yik​ be any run of global variable consensus ADMM (§7.1) with N≥1N\ge1N≥1 agents and ρ>0\rho>0ρ>0. Then:

  1. the dual variables have average zero after the first iteration: yˉk+1=0\bar y^{k+1}=0yˉ​k+1=0 for every k≥0k\ge0k≥0;
  2. zk=xˉkz^k=\bar x^kzk=xˉk for every k≥2k\ge2k≥2;
  3. for every k≥2k\ge2k≥2, xik+1x_i^{k+1}xik+1​ minimizes fi(xi)+yikT(xi−xˉk)+(ρ/2)∥xi−xˉk∥22f_i(x_i)+y_i^{kT}(x_i-\bar x^k)+(\rho/2)\|x_i-\bar x^k\|_2^2fi​(xi​)+yikT​(xi​−xˉk)+(ρ/2)∥xi​−xˉk∥22​ over dom⁡fi\operatorname{dom} f_idomfi​;
  4. for every k≥1k\ge1k≥1, yik+1=yik+ρ(xik+1−xˉk+1)y_i^{k+1}=y_i^k+\rho(x_i^{k+1}-\bar x^{k+1})yik+1​=yik​+ρ(xik+1​−xˉk+1).

Items 3–4 are the book's simplified algorithm, in which the global variable zzz no longer appears:

xik+1:=argmin⁡xi(fi(xi)+yikT(xi−xˉk)+(ρ/2)∥xi−xˉk∥22),yik+1:=yik+ρ(xik+1−xˉk+1).x_i^{k+1}:=\operatorname*{argmin}_{x_i}\big(f_i(x_i)+y_i^{kT}(x_i-\bar x^k)+(\rho/2)\|x_i-\bar x^k\|_2^2\big),\qquad y_i^{k+1}:=y_i^k+\rho(x_i^{k+1}-\bar x^{k+1}).xik+1​:=xi​argmin​(fi​(xi​)+yikT​(xi​−xˉk)+(ρ/2)∥xi​−xˉk∥22​),yik+1​:=yik​+ρ(xik+1​−xˉk+1).

Formalization Note The book says "after the first iteration" and then "using zk=xˉkz^k=\bar x^kzk=xˉk" without an index range. Since zk=xˉk+(1/ρ)yˉk−1z^{k}=\bar x^{k}+(1/\rho)\bar y^{k-1}zk=xˉk+(1/ρ)yˉ​k−1, the identity zk=xˉkz^k=\bar x^kzk=xˉk needs yˉk−1=0\bar y^{k-1}=0yˉ​k−1=0, i.e. k≥2k\ge2k≥2 for an arbitrary starting y0y^0y0; we state exactly these ranges. N≥1N\ge1N≥1 and ρ>0\rho>0ρ>0 are made explicit.

Preamble
import Mathlib
import Definitions.Def_BoydADMM_Consensus_Model

open scoped BigOperators InnerProductSpace
Formal statement
namespace BoydADMM.Consensus

/-- §7.1, p. 50: along every run of global variable consensus ADMM with `N ≥ 1`, `ρ > 0`:
(1) `ȳ^{k+1} = 0` for every `k ≥ 0` (the dual variables have average zero after the first
iteration); (2) `z^k = x̄^k` for every `k ≥ 2`; and hence the simplified iteration holds:
(3) for `k ≥ 2`, `x_i^{k+1}` minimizes `f_i(x_i) + y_i^{kT}(x_i − x̄^k) + (ρ/2)‖x_i − x̄^k‖²` over
`dom f_i`, and (4) for `k ≥ 1`, `y_i^{k+1} = y_i^k + ρ(x_i^{k+1} − x̄^{k+1})`. -/
theorem consensus_dual_average_zero {n N : ℕ} (hN : 0 < N) (ρ : ℝ) (hρ : 0 < ρ)
    (Cf : Fin N → Set (EuclideanSpace ℝ (Fin n))) (f : Fin N → EuclideanSpace ℝ (Fin n) → ℝ)
    (x : ℕ → Fin N → EuclideanSpace ℝ (Fin n)) (z : ℕ → EuclideanSpace ℝ (Fin n))
    (y : ℕ → Fin N → EuclideanSpace ℝ (Fin n)) (hrun : IsConsensusADMMRun Cf f ρ x z y) :
    (∀ k, avg (y (k + 1)) = 0) ∧
    (∀ k, 2 ≤ k → z k = avg (x k)) ∧
    (∀ k, 2 ≤ k → ∀ i, ∀ x' ∈ Cf i,
      f i (x (k + 1) i) + ⟪y k i, x (k + 1) i - avg (x k)⟫_ℝ +
          (ρ / 2) * ‖x (k + 1) i - avg (x k)‖ ^ 2 ≤
        f i x' + ⟪y k i, x' - avg (x k)⟫_ℝ + (ρ / 2) * ‖x' - avg (x k)‖ ^ 2) ∧
    (∀ k, 1 ≤ k → ∀ i, y (k + 1) i = y k i + ρ • (x (k + 1) i - avg (x (k + 1)))) := by sorry

end BoydADMM.Consensus
Source
Boyd, Parikh, Chu, Peleato, Eckstein, Distributed Optimization and Statistical Learning via the Alternating Direction Method of Multipliers, Found. Trends Mach. Learn. 3(1) (2011), p. 50, §7.1 (ȳ^{k+1} = 0 and the simplified algorithm)
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