Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Theorem 4.1, p. 56 — strong second-order sufficiency and independent binding gradients make the KKT system (4.3) strongly regular

Proved
RobinsonSR.NLP.theorem_4_1

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

generalized-equationkkt-conditionslinear-independencenonlinear-programmingp2o-batch-pfp2ap2o-gran-per-chapterp2o-plan-paperp2o-v1second-order-sufficiencystrong-regularity

Consider the nonlinear program

minimize θ(x)subject to g(x)≤0, h(x)=0,(4.1)\text{minimize } \theta(x)\quad\text{subject to } g(x)\le0,\ h(x)=0,\qquad(4.1)minimize θ(x)subject to g(x)≤0, h(x)=0,(4.1)

where θ,g,h\theta,g,hθ,g,h are defined on an open set Ω⊆Rn\Omega\subseteq\mathbb R^nΩ⊆Rn with values in R\mathbb RR, Rp\mathbb R^pRp, Rq\mathbb R^qRq, are Fréchet differentiable on Ω\OmegaΩ, and are twice differentiable at a point x0∈Ωx_0\in\Omegax0​∈Ω. Let u0∈Rpu_0\in\mathbb R^pu0​∈Rp and v0∈Rqv_0\in\mathbb R^qv0​∈Rq be such that (x0,u0,v0)(x_0,u_0,v_0)(x0​,u0​,v0​) solves the KKT generalized equation

0∈[L′(x,u,v)−g(x)−h(x)]+∂ψRn×R+p×Rq[xuv],(4.3)0\in\begin{bmatrix}\mathcal L'(x,u,v)\\-g(x)\\-h(x)\end{bmatrix}+\partial\psi_{\mathbb R^n\times\mathbb R^p_+\times\mathbb R^q}\begin{bmatrix}x\\u\\v\end{bmatrix},\qquad(4.3)0∈​L′(x,u,v)−g(x)−h(x)​​+∂ψRn×R+p​×Rq​​xuv​​,(4.3)

with L(x,u,v)=θ(x)+⟨u,g(x)⟩+⟨v,h(x)⟩\mathcal L(x,u,v)=\theta(x)+\langle u,g(x)\rangle+\langle v,h(x)\rangleL(x,u,v)=θ(x)+⟨u,g(x)⟩+⟨v,h(x)⟩. Write L′′=L′′(x0,u0,v0)\mathcal L''=\mathcal L''(x_0,u_0,v_0)L′′=L′′(x0​,u0​,v0​) for the Hessian in xxx. Assume

  1. (a′) strong second-order sufficiency: ⟨y,L′′y⟩>0\langle y,\mathcal L''y\rangle>0⟨y,L′′y⟩>0 for every nonzero yyy with ⟨∇gi(x0),y⟩=0\langle\nabla g_i(x_0),y\rangle=0⟨∇gi​(x0​),y⟩=0 whenever u0,i>0u_{0,i}>0u0,i​>0 and ⟨∇hj(x0),y⟩=0\langle\nabla h_j(x_0),y\rangle=0⟨∇hj​(x0​),y⟩=0 for all jjj;
  2. (b) linear independence: the gradients ∇gi(x0)\nabla g_i(x_0)∇gi​(x0​) of the binding inequality constraints (gi(x0)=0g_i(x_0)=0gi​(x0​)=0) together with all ∇hj(x0)\nabla h_j(x_0)∇hj​(x0​) are linearly independent.

Then (4.3) is strongly regular at (x0,u0,v0)(x_0,u_0,v_0)(x0​,u0​,v0​): for some λ\lambdaλ, the linearisation of (4.3) at (x0,u0,v0)(x_0,u_0,v_0)(x0​,u0​,v0​) has, for all perturbations in a neighbourhood of 000, a unique solution near (x0,u0,v0)(x_0,u_0,v_0)(x0​,u0​,v0​), depending λ\lambdaλ-Lipschitz-continuously on the perturbation.

The theorem does not assume strict complementary slackness: weakly active constraints (gi(x0)=0=u0,ig_i(x_0)=0=u_{0,i}gi​(x0​)=0=u0,i​) are allowed, and (a′) is imposed only on the directions annihilated by the gradients of constraints with positive multipliers.

Formalization Note "Differentiable on Ω\OmegaΩ" is the standing assumption of (4.1); "twice differentiable at x0x_0x0​" is encoded as differentiability at x0x_0x0​ of each gradient ∇θ\nabla\theta∇θ, ∇gi\nabla g_i∇gi​, ∇hj\nabla h_j∇hj​. Symmetry of L′′\mathcal L''L′′ is not assumed; it follows from these hypotheses. Triples (x,u,v)(x,u,v)(x,u,v) carry the Euclidean norm of Rn+p+q\mathbb R^{n+p+q}Rn+p+q.

Preamble
import Mathlib
import Definitions.Def_RobinsonSR_NLP_Setting
open scoped RealInnerProductSpace Matrix
Formal statement
namespace RobinsonSR.NLP

/-- Theorem 4.1, p. 56: let `θ, g, h` be functions from an open set `Ω ⊆ ℝⁿ` to `ℝ, ℝᵖ, ℝ^q`,
differentiable on `Ω` (the standing assumption of (4.1)) and twice differentiable at `x₀ ∈ Ω`.
Suppose `(x₀, u₀, v₀)` solves (4.3). If the strong second-order sufficient condition (a′) holds
at `(x₀, u₀, v₀)` together with linear independence (b) of the gradients of the binding
constraints, then (4.3) is strongly regular there. -/
theorem theorem_4_1 {n p q : ℕ} (Ω : Set (EuclideanSpace ℝ (Fin n)))
    (hΩ : IsOpen Ω) (x0 : EuclideanSpace ℝ (Fin n)) (hx0 : x0 ∈ Ω)
    (θ : EuclideanSpace ℝ (Fin n) → ℝ)
    (g : EuclideanSpace ℝ (Fin n) → EuclideanSpace ℝ (Fin p))
    (h : EuclideanSpace ℝ (Fin n) → EuclideanSpace ℝ (Fin q))
    (hθ : DifferentiableOn ℝ θ Ω)
    (hg : ∀ i, DifferentiableOn ℝ (fun x => g x i) Ω)
    (hh : ∀ j, DifferentiableOn ℝ (fun x => h x j) Ω)
    (hθ2 : DifferentiableAt ℝ (gradient θ) x0)
    (hg2 : ∀ i, DifferentiableAt ℝ (gradient (fun x => g x i)) x0)
    (hh2 : ∀ j, DifferentiableAt ℝ (gradient (fun x => h x j)) x0)
    (u0 : EuclideanSpace ℝ (Fin p)) (v0 : EuclideanSpace ℝ (Fin q))
    (hsol : -(kktMap θ g h (mk x0 u0 v0)) ∈ RobinsonSR.Reduction.normalCone (kktCone n p q) (mk x0 u0 v0))
    (hSSOSC : ∀ y : EuclideanSpace ℝ (Fin n), y ≠ 0 →
      (∀ i, 0 < u0 i → ⟪gradient (fun x => g x i) x0, y⟫ = 0) →
      (∀ j, ⟪gradient (fun x => h x j) x0, y⟫ = 0) →
      0 < ⟪y, lagHess θ g h x0 u0 v0 y⟫)
    (hLI : LinearIndependent ℝ
      (Sum.elim (fun i : {i : Fin p // g x0 i = 0} => gradient (fun x => g x i) x0)
        (fun j : Fin q => gradient (fun x => h x j) x0))) :
    StronglyRegularAt (kktMap θ g h) (kktCone n p q) (mk x0 u0 v0) := by sorry

end RobinsonSR.NLP
Source
Robinson, Strongly regular generalized equations, Math. Oper. Res. 5 (1980), p. 56, Theorem 4.1; conditions (a′) and (b) p. 55; (4.1)–(4.3) pp. 53–54
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