Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Theorem 4.4 — lim inf⁡t→+∞∥x(t)−x∗∥=0\liminf_{t\to+\infty}\|x(t)-x^*\|=0liminft→+∞​∥x(t)−x∗∥=0, and x(t)→x∗x(t)\to x^*x(t)→x∗ if the trajectory ends inside or outside B(0,∥x∗∥)B(0,\|x^*\|)B(0,∥x∗∥)

Proved
TikhonovHDD.Strong.theorem_4_4

by mikedeng1 · Oct 8, 2026 · Mathlib 0df444a (Lean v4.33.1)

convex-optimizationdynamical-systemsp2o-batch-pfp2bp2o-gran-per-chapterp2o-plan-paperp2o-v1tikhonov-regularization

Let H\mathcal HH be a real Hilbert space, t0>0t_0>0t0​>0, α≥3\alpha\ge 3α≥3, β≥0\beta\ge 0β≥0 and u0,v0∈Hu_0,v_0\in\mathcal Hu0​,v0​∈H. Let g:H→Rg:\mathcal H\to\mathbb Rg:H→R be convex and twice Fréchet differentiable, with gradient Lipschitz continuous on bounded sets and argmin⁡g≠∅\operatorname{argmin} g\neq\emptysetargming=∅, and let ϵ:[t0,+∞)→[0,+∞)\epsilon:[t_0,+\infty)\to[0,+\infty)ϵ:[t0​,+∞)→[0,+∞) be nonincreasing, of class C1C^1C1, with lim⁡t→+∞ϵ(t)=0\lim_{t\to+\infty}\epsilon(t)=0limt→+∞​ϵ(t)=0. Let xxx be a global C2C^2C2-solution of

x¨(t)+αtx˙(t)+β∇2g(x(t))x˙(t)+∇g(x(t))+ϵ(t)x(t)=0,t≥t0,x(t0)=u0, x˙(t0)=v0.\ddot x(t)+\frac{\alpha}{t}\dot x(t)+\beta\nabla^2 g(x(t))\dot x(t)+\nabla g(x(t))+\epsilon(t)x(t)=0,\quad t\ge t_0,\qquad x(t_0)=u_0,\ \dot x(t_0)=v_0 .x¨(t)+tα​x˙(t)+β∇2g(x(t))x˙(t)+∇g(x(t))+ϵ(t)x(t)=0,t≥t0​,x(t0​)=u0​, x˙(t0​)=v0​.

Assume that

∫t0+∞ϵ(t)t dt<+∞andlim⁡t→+∞βϵ(t)tα3+1∫t0tϵ2(s)sα3+1 ds=0,\int_{t_0}^{+\infty}\frac{\epsilon(t)}{t}\,dt<+\infty\qquad\text{and}\qquad\lim_{t\to+\infty}\frac{\beta}{\epsilon(t)t^{\frac{\alpha}{3}+1}}\int_{t_0}^{t}\epsilon^2(s)s^{\frac{\alpha}{3}+1}\,ds=0,∫t0​+∞​tϵ(t)​dt<+∞andt→+∞lim​ϵ(t)t3α​+1β​∫t0​t​ϵ2(s)s3α​+1ds=0,

that there exist a>1a>1a>1 and t1≥t0t_1\ge t_0t1​≥t0​ such that ϵ˙(t)≤−aβ2ϵ2(t)\dot\epsilon(t)\le-\frac{a\beta}{2}\epsilon^2(t)ϵ˙(t)≤−2aβ​ϵ2(t) for every t≥t1t\ge t_1t≥t1​, and in addition that

  1. in case α=3\alpha=3α=3: lim⁡t→+∞t2ϵ(t)=+∞\lim_{t\to+\infty}t^2\epsilon(t)=+\inftylimt→+∞​t2ϵ(t)=+∞;
  2. in case α>3\alpha>3α>3: there exists c>0c>0c>0 such that t2ϵ(t)≥23α(13α−1+βc2)t^2\epsilon(t)\ge\frac23\alpha\left(\frac13\alpha-1+\beta c^2\right)t2ϵ(t)≥32​α(31​α−1+βc2) for ttt large enough.

If x∗x^*x∗ is the element of minimum norm of the nonempty convex closed set argmin⁡g\operatorname{argmin} gargming, then

lim inf⁡t→+∞∥x(t)−x∗∥=0.\liminf_{t\to+\infty}\|x(t)-x^*\|=0 .t→+∞liminf​∥x(t)−x∗∥=0.

In addition,

lim⁡t→+∞∥x(t)−x∗∥=0\lim_{t\to+\infty}\|x(t)-x^*\|=0t→+∞lim​∥x(t)−x∗∥=0

if there exists T≥t0T\ge t_0T≥t0​ such that the trajectory {x(t):t≥T}\{x(t):t\ge T\}{x(t):t≥T} stays either in the open ball B(0,∥x∗∥)B(0,\|x^*\|)B(0,∥x∗∥), or in its complement.

This is the strong-convergence result for which the Tikhonov term ϵ(t)x(t)\epsilon(t)x(t)ϵ(t)x(t) is added to the Hessian-damped inertial dynamics: without it, trajectories are only known to converge weakly to some minimizer, while here they approach the distinguished minimizer of minimum norm.

Formalization Note Stated for every global C2C^2C2-solution of (5) (existence and uniqueness is Theorem 2.1 of the paper). lim inf⁡t→+∞∥x(t)−x∗∥=0\liminf_{t\to+\infty}\|x(t)-x^*\|=0liminft→+∞​∥x(t)−x∗∥=0 is written as: for every δ>0\delta>0δ>0, ∥x(t)−x∗∥<δ\|x(t)-x^*\|<\delta∥x(t)−x∗∥<δ for arbitrarily large ttt, which avoids the junk value of Lean's Filter.liminf. The alternative "in the ball, or in its complement" uses one TTT for both branches: either ∥x(t)∥<∥x∗∥\|x(t)\|<\|x^*\|∥x(t)∥<∥x∗∥ for all t≥Tt\ge Tt≥T, or ∥x(t)∥≥∥x∗∥\|x(t)\|\ge\|x^*\|∥x(t)∥≥∥x∗∥ for all t≥Tt\ge Tt≥T. tα/3+1t^{\alpha/3+1}tα/3+1 is the real power. Positivity of ϵ\epsilonϵ is not assumed; it follows from the hypotheses for α=3\alpha=3α=3 and α>3\alpha>3α>3.

Preamble
import Mathlib
import Definitions.Def_TikhonovHDD_Strong_Setting

open Set Filter Topology
Formal statement
namespace TikhonovHDD.Strong

/-- Theorem 4.4 (p. 21): `liminf_{t→+∞} ‖x(t) − x⋆‖ = 0` (for every `δ > 0`,
`‖x(t) − x⋆‖ < δ` for arbitrarily large `t`), and `‖x(t) − x⋆‖ → 0` if there is `T ≥ t₀`
such that `{x(t) : t ≥ T}` stays in the open ball `B(0, ‖x⋆‖)` or stays in its complement.
Standing hypotheses (those of Theorem 4.4, arXiv:1911.12845v2, p. 21): the General
assumption, `t₀ > 0`, `α ≥ 3`, `β ≥ 0`, `x` a global `C²`-solution of (5),
`∫_{t₀}^{+∞} ε(t)/t dt < +∞`, `β/(ε(t) t^{α/3+1}) ∫_{t₀}^t ε²(s) s^{α/3+1} ds → 0`,
`ε̇(t) ≤ −(aβ/2) ε²(t)` for `t ≥ t₁` (some `a > 1`, `t₁ ≥ t₀`), for `α = 3`: `t² ε(t) → +∞`,
for `α > 3`: `t² ε(t) ≥ (2/3)α((1/3)α − 1 + βc²)` eventually (some `c > 0`), and `x⋆` the
minimum-norm element of `argmin g`. -/
theorem theorem_4_4
    {H : Type*} [NormedAddCommGroup H] [InnerProductSpace ℝ H] [CompleteSpace H]
    (g : H → ℝ) (α β t₀ : ℝ) (ε ε' : ℝ → ℝ) (u₀ v₀ : H) (x xd xdd : ℝ → H) (xstar : H)
    (ht₀ : 0 < t₀) (hα : 3 ≤ α) (hβ : 0 ≤ β)
    (hGA : GeneralAssumption g t₀ ε ε')
    (hx : IsSolution g α β t₀ ε u₀ v₀ x xd xdd)
    (hε_int : MeasureTheory.IntegrableOn (fun t => ε t / t) (Ici t₀))
    (hε_lim : Tendsto
      (fun t => β / (ε t * t ^ (α / 3 + 1)) * ∫ s in t₀..t, ε s ^ 2 * s ^ (α / 3 + 1))
      atTop (𝓝 0))
    (hε_deriv : ∃ a > (1 : ℝ), ∃ t₁ ≥ t₀, ∀ t ≥ t₁, ε' t ≤ -(a * β / 2) * ε t ^ 2)
    (hα3 : α = 3 → Tendsto (fun t => t ^ 2 * ε t) atTop atTop)
    (hα3lt : 3 < α → ∃ c > (0 : ℝ),
      ∀ᶠ t in atTop, 2 / 3 * α * (1 / 3 * α - 1 + β * c ^ 2) ≤ t ^ 2 * ε t)
    (hxstar : IsMinNormMinimizer g xstar) :
    (∀ δ > (0 : ℝ), ∃ᶠ t in atTop, ‖x t - xstar‖ < δ) ∧
      ((∃ T ≥ t₀, (∀ t ≥ T, ‖x t‖ < ‖xstar‖) ∨ (∀ t ≥ T, ‖xstar‖ ≤ ‖x t‖)) →
        Tendsto (fun t => ‖x t - xstar‖) atTop (𝓝 0)) := by sorry

end TikhonovHDD.Strong
Source
Boţ, Csetnek, László, Tikhonov regularization of a second order dynamical system with Hessian driven damping, arXiv:1911.12845v2, p. 21, Theorem 4.4
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