Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

(4.1), p. 26 — for quadratic f, x⁺ = (P + ρAᵀA)⁻¹(ρAᵀv − q)

Proved
BoydADMM.Prox.quadratic_xUpdate

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

admmlinear-algebrap2o-batch-pfp2bp2o-gran-per-chapterp2o-plan-bookp2o-v1quadratic-programming

Let f(x)=12xTPx+qTx+rf(x)=\tfrac12x^TPx+q^Tx+rf(x)=21​xTPx+qTx+r with P∈S+nP\in\mathbf S^n_+P∈S+n​ (symmetric positive semidefinite), q∈Rnq\in\mathbb R^nq∈Rn, r∈Rr\in\mathbb Rr∈R. Let A∈Rp×nA\in\mathbb R^{p\times n}A∈Rp×n, ρ>0\rho>0ρ>0, and assume that P+ρATAP+\rho A^TAP+ρATA is invertible. Then:

  1. the coefficient matrix P+ρATAP+\rho A^TAP+ρATA is positive definite;
  2. for every v∈Rpv\in\mathbb R^pv∈Rp, the xxx-update x+=argmin⁡x(f(x)+ρ2∥Ax−v∥22)x^+=\operatorname*{argmin}_x\bigl(f(x)+\tfrac{\rho}{2}\|Ax-v\|_2^2\bigr)x+=argminx​(f(x)+2ρ​∥Ax−v∥22​) over all of Rn\mathbb R^nRn is unique and is the affine function of vvv
x+=(P+ρATA)−1(ρATv−q).(4.1)x^+=(P+\rho A^TA)^{-1}(\rho A^Tv-q).\tag{4.1}x+=(P+ρATA)−1(ρATv−q).(4.1)

So computing an xxx-update with a quadratic objective term amounts to solving one linear system with a positive definite coefficient matrix, which is what makes factorization caching (§4.2.3) and the matrix inversion lemma (§4.2.4) useful.

Formalization Note P∈S+nP\in\mathbf S^n_+P∈S+n​ is P.PosSemidef (which includes symmetry). "Assuming P+ρATAP+\rho A^TAP+ρATA is invertible" is the explicit hypothesis IsUnit (P + ρ • (Aᵀ * A)). The claim "x+x^+x+ is given by (4.1)" is stated as an equivalence: xxx is a minimizer if and only if xxx equals the right-hand side of (4.1), which gives both existence and uniqueness. The positive definiteness, stated by the book in the next sentence, is the first conjunct.

Preamble
import Mathlib
import Definitions.Def_BoydADMM_Prox_Basic

open Matrix
Formal statement
namespace BoydADMM.Prox

theorem quadratic_xUpdate {n p : ℕ} (P : Matrix (Fin n) (Fin n) ℝ) (hP : P.PosSemidef)
    (q : EuclideanSpace ℝ (Fin n)) (r : ℝ) (A : Matrix (Fin p) (Fin n) ℝ) (ρ : ℝ) (hρ : 0 < ρ)
    (hinv : IsUnit (P + ρ • (Aᵀ * A))) :
    (P + ρ • (Aᵀ * A)).PosDef ∧
      ∀ (v : EuclideanSpace ℝ (Fin p)) (x : EuclideanSpace ℝ (Fin n)),
        IsXUpdate Set.univ (quadObj P q r) ρ A v x ↔
          x = Matrix.toEuclideanLin (P + ρ • (Aᵀ * A))⁻¹
            (ρ • Matrix.toEuclideanLin Aᵀ v - q) := by sorry

end BoydADMM.Prox
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. 26, §4.2, (4.1)
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