Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Theorem 3.2(ii), p. 484 — a dual sufficient condition: h*(y) − g*(y) ≥ h*(y*) − g*(y*) for some y ∈ ∂h(x) makes x* a local minimizer

Proved
TaoAnDCA.LocalOpt.theorem_3_2_ii

by mikedeng1 · Oct 6, 2026 · Mathlib 0df444a (Lean v4.33.1)

dc-programmingdualitylocal-optimalityp2o-batch-pfp2ap2o-gran-per-chapterp2o-plan-paperp2o-v1sufficient-condition

Let g,h∈Γ0(Rn)g, h \in \Gamma_0(\mathbb R^n)g,h∈Γ0​(Rn) satisfy the standing inclusions dom⁡g⊂dom⁡h\operatorname{dom} g \subset \operatorname{dom} hdomg⊂domh and dom⁡h∗⊂dom⁡g∗\operatorname{dom} h^* \subset \operatorname{dom} g^*domh∗⊂domg∗, with differences taken under the convention +∞−(+∞)=+∞+\infty - (+\infty) = +\infty+∞−(+∞)=+∞. Let x∗x^*x∗ be a critical point of g−hg - hg−h and let

y∗∈∂g(x∗)∩∂h(x∗).y^* \in \partial g(x^*) \cap \partial h(x^*).y∗∈∂g(x∗)∩∂h(x∗).

Let UUU be a neighbourhood of x∗x^*x∗ such that U∩dom⁡g⊂dom⁡∂hU \cap \operatorname{dom} g \subset \operatorname{dom} \partial hU∩domg⊂dom∂h. Suppose that for every x∈U∩dom⁡gx \in U \cap \operatorname{dom} gx∈U∩domg there is y∈∂h(x)y \in \partial h(x)y∈∂h(x) with

h∗(y)−g∗(y)≥h∗(y∗)−g∗(y∗).h^*(y) - g^*(y) \ge h^*(y^*) - g^*(y^*).h∗(y)−g∗(y)≥h∗(y∗)−g∗(y∗).

Then x∗x^*x∗ is a local minimizer of g−hg - hg−h. More precisely,

g(x)−h(x)≥g(x∗)−h(x∗)∀x∈U∩dom⁡g.g(x) - h(x) \ge g(x^*) - h(x^*) \qquad \forall x \in U \cap \operatorname{dom} g.g(x)−h(x)≥g(x∗)−h(x∗)∀x∈U∩domg.

The authors call this condition new: it certifies local optimality in the primal program through values of the dual objective h∗−g∗h^* - g^*h∗−g∗. Applied to the dual pair (h∗,g∗)(h^*, g^*)(h∗,g∗) it is the core of Corollary 3.5.

Formalization Note. "x∗x^*x∗ is a critical point and y∗∈∂g(x∗)∩∂h(x∗)y^* \in \partial g(x^*) \cap \partial h(x^*)y∗∈∂g(x∗)∩∂h(x∗)" is the single hypothesis on y∗y^*y∗, which already implies criticality. dom⁡∂h\operatorname{dom}\partial hdom∂h is the set of points where ∂h\partial h∂h is nonempty. Both dual values are the convention-respecting difference dcSub (conj h) (conj g), and the conclusion contains both the local-minimizer claim and its "more precisely" form.

Preamble
import Mathlib
import Definitions.Def_TaoAnDCA_LocalOpt_Setting

open Filter Topology
Formal statement
namespace TaoAnDCA.LocalOpt

/-- Theorem 3.2 (ii), p. 484: let `x*` be a critical point of `g − h` and
`y* ∈ ∂g(x*) ∩ ∂h(x*)`, and let `U` be a neighbourhood of `x*` with `U ∩ dom g ⊂ dom ∂h`.
If for every `x ∈ U ∩ dom g` there is `y ∈ ∂h(x)` with `h*(y) − g*(y) ≥ h*(y*) − g*(y*)`,
then `x*` is a local minimizer of `g − h`; more precisely
`g(x) − h(x) ≥ g(x*) − h(x*)` for all `x ∈ U ∩ dom g`. -/
theorem theorem_3_2_ii {n : ℕ} (g h : EuclideanSpace ℝ (Fin n) → EReal) (hgh : TaoAnDCA.GlobalOpt.DCStanding g h)
    (xs ys : EuclideanSpace ℝ (Fin n)) (hys : ys ∈ TaoAnDCA.GlobalOpt.subdiff g xs ∩ TaoAnDCA.GlobalOpt.subdiff h xs)
    (U : Set (EuclideanSpace ℝ (Fin n))) (hU : U ∈ 𝓝 xs)
    (hdom : U ∩ TaoAnDCA.GlobalOpt.effDom g ⊆ {x | (TaoAnDCA.GlobalOpt.subdiff h x).Nonempty})
    (hcond : ∀ x ∈ U ∩ TaoAnDCA.GlobalOpt.effDom g, ∃ y ∈ TaoAnDCA.GlobalOpt.subdiff h x,
      TaoAnDCA.GlobalOpt.dcSub (CondatPD.FinDim.conj h) (CondatPD.FinDim.conj g) ys ≤
        TaoAnDCA.GlobalOpt.dcSub (CondatPD.FinDim.conj h) (CondatPD.FinDim.conj g) y) :
    IsDCLocalMin g h xs ∧ ∀ x ∈ U ∩ TaoAnDCA.GlobalOpt.effDom g, TaoAnDCA.GlobalOpt.dcSub g h xs ≤ TaoAnDCA.GlobalOpt.dcSub g h x := by sorry

end TaoAnDCA.LocalOpt
Source
Pham Dinh & Le Thi, A d.c. optimization algorithm for solving the trust-region subproblem, SIAM J. Optim. 8 (1998), p. 484, Theorem 3.2(ii)
Human review
  • Endorsed by Shuze Chen · Oct 6, 2026

    Confirmed by the moderator at approval.

  • Endorsed by mikedeng1 · Oct 6, 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