Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Theorem 3.5 — for α>3\alpha>3α>3 and ∫tϵ(t)dt<∞\int t\epsilon(t)dt<\infty∫tϵ(t)dt<∞, x(t)x(t)x(t) converges weakly to a minimizer of ggg

Proved
TikhonovHDD.Weak.theorem_3_5

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∈R\alpha\in\mathbb Rα∈R, 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 ∇g\nabla g∇g 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 (the General assumption). Let x:[t0,+∞)→Hx:[t_0,+\infty)\to\mathcal Hx:[t0​,+∞)→H 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.(5)\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,\qquad t\ge t_0,\quad x(t_0)=u_0,\ \dot x(t_0)=v_0. \tag{5}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​.(5)

Assume that ∫t0+∞tϵ(t) dt<+∞\int_{t_0}^{+\infty} t\epsilon(t)\,dt<+\infty∫t0​+∞​tϵ(t)dt<+∞ and that one of the following conditions holds:

  1. (a) 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​;
  2. (b) there exist a>0a>0a>0 and t1≥t0t_1\ge t_0t1​≥t0​ such that ϵ(t)≤at\epsilon(t)\le\frac{a}{t}ϵ(t)≤ta​ for every t≥t1t\ge t_1t≥t1​.

Write min⁡g\min gming for the minimal value of ggg.

If α>3\alpha>3α>3, then x(t)x(t)x(t) converges weakly, as t→+∞t\to+\inftyt→+∞, to an element of argmin⁡g\operatorname{argmin} gargming: there is xˉ∈argmin⁡g\bar x\in\operatorname{argmin} gxˉ∈argming with

⟨x(t),v⟩⟶⟨xˉ,v⟩for every v∈H.\langle x(t),v\rangle\longrightarrow\langle\bar x,v\rangle\qquad\text{for every } v\in\mathcal H.⟨x(t),v⟩⟶⟨xˉ,v⟩for every v∈H.

This is the main result of Section 3: adding a Tikhonov regularization term ϵ(t)x(t)\epsilon(t)x(t)ϵ(t)x(t) with ∫tϵ(t) dt<+∞\int t\epsilon(t)\,dt<+\infty∫tϵ(t)dt<+∞ to the inertial system with Hessian-driven damping preserves the weak convergence of its trajectories to minimizers known for ϵ≡0\epsilon\equiv0ϵ≡0.

Formalization Note The solution xxx is a map R→H\mathbb R\to\mathcal HR→H with velocity and acceleration maps x˙,x¨\dot x,\ddot xx˙,x¨; derivatives are one-sided (within [t0,+∞)[t_0,+\infty)[t0​,+∞)) at t0t_0t0​, x¨\ddot xx¨ is continuous on [t0,+∞)[t_0,+\infty)[t0​,+∞), and values before t0t_0t0​ are irrelevant. The statement is made for every C2C^2C2-solution; existence and uniqueness is Theorem 2.1. ∇2g(x)v\nabla^2 g(x)v∇2g(x)v is the Fréchet derivative of ∇g\nabla g∇g at xxx applied to vvv; ϵ˙\dot\epsilonϵ˙ is a derivative map ϵ′\epsilon'ϵ′ tied to ϵ\epsilonϵ by one-sided derivatives on [t0,+∞)[t_0,+\infty)[t0​,+∞). min⁡g\min gming is inf⁡yg(y)\inf_{y}g(y)infy​g(y), attained because argmin⁡g≠∅\operatorname{argmin} g\neq\emptysetargming=∅. Integrability ∫t0+∞∣f∣<+∞\int_{t_0}^{+\infty}|f|<+\infty∫t0​+∞​∣f∣<+∞ is Lebesgue integrability on [t0,+∞)[t_0,+\infty)[t0​,+∞) (IntegrableOn), and OOO, ooo are taken as t→+∞t\to+\inftyt→+∞. Weak convergence is stated through inner products against every v∈Hv\in\mathcal Hv∈H; it is not norm convergence.

Preamble
import Mathlib
import Definitions.Def_TikhonovHDD_Weak_Setting

open Filter Topology Set MeasureTheory Asymptotics
Formal statement
namespace TikhonovHDD.Weak

theorem theorem_3_5
    {H : Type*} [NormedAddCommGroup H] [InnerProductSpace ℝ H] [CompleteSpace H]
    (g : H → ℝ) (ε ε' : ℝ → ℝ) (α β t₀ : ℝ) (u₀ v₀ : H) (x xd xdd : ℝ → H)
    (ht₀ : 0 < t₀) (hβ : 0 ≤ β) (hg : GenAssumptionG g) (hε : GenAssumptionEps t₀ ε ε')
    (hsol : IsSolution g ε α β t₀ u₀ v₀ x xd xdd)
    (hint : IntegrableOn (fun t => t * ε t) (Ici t₀))
    (hab : CondA β t₀ ε ε' ∨ CondB t₀ ε)
    (hα : 3 < α) :
    ∃ xbar ∈ argmin g, WeakTendstoAtTop x xbar := by sorry

end TikhonovHDD.Weak
Source
Boţ, Csetnek, László, Tikhonov regularization of a second order dynamical system with Hessian driven damping, arXiv:1911.12845v2, p. 15, Theorem 3.5 (proof p. 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