Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Lemma 3.1, p. 4 — the energy inequality (3.1) for z ∈ S, n ≥ 2

Proved
ProjReflGrad.Weak.lemma_3_1

by mikedeng1 · Oct 9, 2026 · Mathlib 0df444a (Lean v4.33.1)

monotone-operatorp2o-batch-pfp2ap2o-gran-per-chapterp2o-plan-paperp2o-v1projection-methodvariational-inequality

Let HHH be a real Hilbert space, C⊆HC\subseteq HC⊆H closed and convex, and F:H→HF:H\to HF:H→H monotone and Lipschitz continuous with constant L>0L>0L>0. Let λ>0\lambda>0λ>0, let (xn),(yn)(x_n),(y_n)(xn​),(yn​) be generated by Algorithm 3.1 (x0=y0∈Hx_0=y_0\in Hx0​=y0​∈H, xn+1=PC(xn−λF(yn))x_{n+1}=P_C(x_n-\lambda F(y_n))xn+1​=PC​(xn​−λF(yn​)), yn+1=2xn+1−xny_{n+1}=2x_{n+1}-x_nyn+1​=2xn+1​−xn​), and let z∈Sz\in Sz∈S be a solution of the variational inequality (1.1). Then for every n≥2n\ge2n≥2

∥xn+1−z∥2≤∥xn−z∥2−(1−λL(1+2))∥xn−xn−1∥2+λL∥xn−yn−1∥2−(1−2λL)∥xn+1−yn∥2−2λ⟨F(z),yn−z⟩.\|x_{n+1}-z\|^2\le\|x_n-z\|^2-\big(1-\lambda L(1+\sqrt2)\big)\|x_n-x_{n-1}\|^2+\lambda L\|x_n-y_{n-1}\|^2-\big(1-\sqrt2\lambda L\big)\|x_{n+1}-y_n\|^2-2\lambda\langle F(z),y_n-z\rangle .∥xn+1​−z∥2≤∥xn​−z∥2−(1−λL(1+2​))∥xn​−xn−1​∥2+λL∥xn​−yn−1​∥2−(1−2​λL)∥xn+1​−yn​∥2−2λ⟨F(z),yn​−z⟩.

This is the central estimate of the convergence proof.

Formalization Note. The paper states (3.1) without a range of nnn; its proof uses xn−1∈Cx_{n-1}\in Cxn−1​∈C, which needs n≥2n\ge2n≥2 because x0x_0x0​ may lie outside CCC, and the proof of Theorem 3.2 uses (3.1) only for n≥2n\ge2n≥2. The step bound λ<(2−1)/L\lambda<(\sqrt2-1)/Lλ<(2​−1)/L is not needed for this inequality and is not assumed; nonemptiness of SSS is given by zzz.

Preamble
import Mathlib
import Definitions.Def_ProjReflGrad_Weak_Setting
Formal statement
namespace ProjReflGrad.Weak

/-- Lemma 3.1, inequality (3.1) (Malitsky 2015, p. 4), for `n ≥ 2`: along a run of Algorithm 3.1
and for `z ∈ S`,
`‖x_{n+1} - z‖² ≤ ‖x_n - z‖² - (1 - λL(1 + √2))‖x_n - x_{n-1}‖² + λL‖x_n - y_{n-1}‖²
  - (1 - √2λL)‖x_{n+1} - y_n‖² - 2λ⟨F(z), y_n - z⟩`. -/
theorem lemma_3_1 {H : Type*} [NormedAddCommGroup H] [InnerProductSpace ℝ H] [CompleteSpace H]
    (C : Set H) (hCcl : IsClosed C) (hCcv : Convex ℝ C) (F : H → H) (L lam : ℝ)
    (hC2 : IsMonotoneMap F) (hL : 0 < L) (hC3 : IsLipschitzMap F L) (hlam0 : 0 < lam)
    (x y : ℕ → H) (hrun : IsPRGRun C F lam x y) (z : H) (hz : z ∈ solSet C F) :
    ∀ n : ℕ, 2 ≤ n →
      ‖x (n + 1) - z‖ ^ 2 ≤ ‖x n - z‖ ^ 2
        - (1 - lam * L * (1 + Real.sqrt 2)) * ‖x n - x (n - 1)‖ ^ 2
        + lam * L * ‖x n - y (n - 1)‖ ^ 2
        - (1 - Real.sqrt 2 * lam * L) * ‖x (n + 1) - y n‖ ^ 2
        - 2 * lam * inner ℝ (F z) (y n - z) := by sorry

end ProjReflGrad.Weak
Source
Malitsky, Projected Reflected Gradient Methods for Monotone Variational Inequalities, arXiv:1502.04968v1, p. 4, Lemma 3.1, (3.1)
Human review
  • Endorsed by Shuze Chen · Oct 9, 2026

    Confirmed by the moderator at approval.

  • Endorsed by mikedeng1 · Oct 9, 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