Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

(2.18), proof of Theorem 2.4 — f(x* − x*ₘeₘ − σx*ₘeᵢ) ≤ f(x*) − σx*ₘ∇ᵢf(x*) + L₂(f)(x*ₘ)² at a CW-minimum with ‖x*‖₀ = s

Proved
SparseNLO.CWOpt.eq_2_18

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

optimality-conditionsp2o-batch-pfp2bp2o-gran-per-chapterp2o-plan-paperp2o-v1sparse-optimization

Let f:Rn→Rf:\mathbb R^n\to\mathbb Rf:Rn→R be continuously differentiable and bounded below with Lipschitz gradient (Assumption 2), let 0<s<n0<s<n0<s<n, and let L2L_2L2​ satisfy the local Lipschitz condition (2.15). Let x∗x^*x∗ be a coordinate-wise minimum of (P) with ∥x∗∥0=s\|x^*\|_0=s∥x∗∥0​=s. Let i∈I0(x∗)i\in I_0(x^*)i∈I0​(x∗), let mmm be an index with ∣xm∗∣=Ms(x∗)|x^*_m|=M_s(x^*)∣xm∗​∣=Ms​(x∗), and set

σ=sgn⁡(xm∗∇if(x∗))∈{−1,0,1}.\sigma=\operatorname{sgn}\big(x^*_m\nabla_i f(x^*)\big)\in\{-1,0,1\}.σ=sgn(xm∗​∇i​f(x∗))∈{−1,0,1}.

Then

f(x∗−xm∗em−σxm∗ei)≤f(x∗)−σxm∗∇if(x∗)+L2 (xm∗)2.f\big(x^*-x^*_me_m-\sigma x^*_me_i\big)\le f(x^*)-\sigma x^*_m\nabla_i f(x^*)+L_2\,(x^*_m)^2 .f(x∗−xm∗​em​−σxm∗​ei​)≤f(x∗)−σxm∗​∇i​f(x∗)+L2​(xm∗​)2.

This is the estimate that, combined with the CW-minimality inequality f(x∗)≤f(x∗−xm∗em−σxm∗ei)f(x^*)\le f(x^*-x^*_me_m-\sigma x^*_me_i)f(x∗)≤f(x∗−xm∗​em​−σxm∗​ei​) of (2.17), yields the bound ∣∇if(x∗)∣≤L2Ms(x∗)|\nabla_i f(x^*)|\le L_2M_s(x^*)∣∇i​f(x∗)∣≤L2​Ms​(x∗) on the zero coordinates in Theorem 2.4.

Formalization Note The statement is the first and the last member of the chain (2.18). The printed middle equality uses ∥xm∗em+σxm∗ei∥2=2(xm∗)2\|x^*_me_m+\sigma x^*_me_i\|^2=2(x^*_m)^2∥xm∗​em​+σxm∗​ei​∥2=2(xm∗​)2, which needs σ2=1\sigma^2=1σ2=1; when ∇if(x∗)=0\nabla_i f(x^*)=0∇i​f(x∗)=0 (σ=0\sigma=0σ=0) the norm is (xm∗)2(x^*_m)^2(xm∗​)2 and the inequality still holds because L2≥0L_2\ge0L2​≥0. The sign is SignType.sign cast to R\mathbb RR, with sgn⁡0=0\operatorname{sgn}0=0sgn0=0. The facts m∈I1(x∗)m\in I_1(x^*)m∈I1​(x∗) and ∇mf(x∗)=0\nabla_m f(x^*)=0∇m​f(x∗)=0 used on the page are consequences of the hypotheses, not assumptions.

Preamble
import Mathlib
import Definitions.Def_SparseNLO_CWOpt_Setting
Formal statement
namespace SparseNLO.CWOpt

/-- (2.18), proof of Theorem 2.4 (p. 12): let `x*` be a CW-minimum with `‖x*‖₀ = s`, let
`i ∈ I₀(x*)`, let `m` be an index with `|x*_m| = M_s(x*)`, and `σ = sgn(x*_m ∇_i f(x*))`. Then
`f(x* − x*_m e_m − σ x*_m e_i) ≤ f(x*) − σ x*_m ∇_i f(x*) + L2 (x*_m)²`. -/
theorem eq_2_18 {n s : ℕ} (f : EuclideanSpace ℝ (Fin n) → ℝ) (hf : ContDiff ℝ 1 f)
    (hbdd : ∃ γ : ℝ, ∀ x, γ ≤ f x) (hs : 0 < s) (hsn : s < n)
    (Lf : NNReal) (hLf : LipschitzWith Lf (gradient f))
    (L2 : ℝ) (hL2 : LocalLipschitz f L2)
    (xstar : EuclideanSpace ℝ (Fin n)) (hcw : IsCWMin f s xstar) (hxs : l0 xstar = s)
    (i : Fin n) (hi : i ∈ I0 xstar) (m : Fin n) (hm : |xstar m| = Ms s xstar) :
    f (xstar - EuclideanSpace.single m (xstar m)
        - EuclideanSpace.single i ((SignType.sign (xstar m * gradient f xstar i) : ℝ) * xstar m))
      ≤ f xstar - (SignType.sign (xstar m * gradient f xstar i) : ℝ) * xstar m * gradient f xstar i
        + L2 * xstar m ^ 2 := by sorry

end SparseNLO.CWOpt
Source
Beck, Eldar, Sparsity Constrained Nonlinear Optimization: Optimality Conditions and Algorithms, SIAM J. Optim. (2013), doi:10.1137/120869778 (source: arXiv:1203.4580v1), p. 12, §2.4, proof of Theorem 2.4, (2.18)
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