Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Theorem 2.16 — for α>3\alpha>3α>3 and argmin⁡Φ≠∅\operatorname{argmin}\Phi\neq\emptysetargminΦ=∅, every solution of x¨+αtx˙+∇Φ(x)=0\ddot x+\frac{\alpha}{t}\dot x+\nabla\Phi(x)=0x¨+tα​x˙+∇Φ(x)=0 converges weakly to a minimizer

Proved
InertialAVD.Traj.theorem_2_16

by mikedeng1 · Oct 7, 2026 · Mathlib 0df444a (Lean v4.33.1)

convex-optimizationinertial-dynamicsp2o-batch-pfp2bp2o-gran-per-chapterp2o-plan-paperp2o-v1weak-convergence

Let H\mathcal HH be a real Hilbert space and let Φ:H→R\Phi:\mathcal H\to\mathbb RΦ:H→R be a continuously differentiable convex function such that argmin⁡Φ≠∅\operatorname{argmin}\Phi\neq\emptysetargminΦ=∅. Let α>3\alpha>3α>3, t0>0t_0>0t0​>0, and let x:[t0,+∞[→Hx:[t_0,+\infty[\to\mathcal Hx:[t0​,+∞[→H be a solution of

x¨(t)+αtx˙(t)+∇Φ(x(t))=0.\ddot x(t)+\frac{\alpha}{t}\dot x(t)+\nabla\Phi(x(t))=0.x¨(t)+tα​x˙(t)+∇Φ(x(t))=0.

Then x(t)x(t)x(t) converges weakly, as t→+∞t\to+\inftyt→+∞, to a point in argmin⁡Φ\operatorname{argmin}\PhiargminΦ: there is xˉ\bar xxˉ with Φ(xˉ)≤Φ(y)\Phi(\bar x)\le\Phi(y)Φ(xˉ)≤Φ(y) for all y∈Hy\in\mathcal Hy∈H and

⟨x(t),y⟩⟶⟨xˉ,y⟩(t→+∞)for every y∈H.\langle x(t),y\rangle\longrightarrow\langle\bar x,y\rangle\quad(t\to+\infty)\quad\text{for every }y\in\mathcal H.⟨x(t),y⟩⟶⟨xˉ,y⟩(t→+∞)for every y∈H.

This is the convergence of the trajectories of the continuous-time model of Nesterov's accelerated gradient method, for damping parameters α>3\alpha>3α>3; the case α=3\alpha=3α=3 is open.

Formalization Note The conclusion is weak convergence, tested against every y∈Hy\in\mathcal Hy∈H; in an infinite-dimensional H\mathcal HH it does not imply norm convergence. Solutions are pairs (x,x˙)(x,\dot x)(x,x˙) satisfying (1) on [t0,+∞[[t_0,+\infty[[t0​,+∞[ as a first-order system, with t0>0t_0>0t0​>0.

Preamble
import Mathlib
import Definitions.Def_InertialAVD_Traj_Setting
Formal statement
namespace InertialAVD.Traj

theorem theorem_2_16 {H : Type*} [NormedAddCommGroup H] [InnerProductSpace ℝ H] [CompleteSpace H]
    (Φ : H → ℝ) (hΦconv : ConvexOn ℝ Set.univ Φ) (hΦC1 : ContDiff ℝ 1 Φ)
    (hargmin : ∃ z : H, ∀ y : H, Φ z ≤ Φ y)
    (α t₀ : ℝ) (hα : 3 < α) (x v : ℝ → H) (hsol : IsSolution Φ α t₀ x v) :
    ∃ xbar : H, (∀ y : H, Φ xbar ≤ Φ y) ∧ WeakTendstoAtTop x xbar := by sorry

end InertialAVD.Traj
Source
Attouch, Chbani, Peypouquet, Redont, Fast convergence of inertial dynamics and algorithms with asymptotic vanishing damping, Optimization Online preprint 5179 (Oct. 2015), p. 7, Theorem 2.16
Human review
  • Endorsed by Shuze Chen · Oct 8, 2026

    Confirmed by the moderator at approval.

  • Endorsed by mikedeng1 · Oct 8, 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