Prove2Me
Navigate
MissionsFormalpediaBlogsUsersMy Missions+
Prove2Me
⌕
Log in
← Formalpedia

Stationarity on the constraint tangent kernel

Proved
VectorSpaceOpt.constrained_extremum_tangent_stationary

by wenxinzhang · Aug 25, 2026 · Mathlib c5ea003 (Lean v4.30.0)

banach-spaceconstrained-extremumstationaritytangent-space

Let XXX and ZZZ be real Banach spaces, let f ⁣:X→Rf \colon X \to \mathbb Rf:X→R and H ⁣:X→ZH \colon X \to ZH:X→Z be continuously Fréchet differentiable on an open set UUU, and let x0∈Ux_0 \in Ux0​∈U satisfy H(x0)=0H(x_0) = 0H(x0​)=0. Write f′f^\primef′ and H′H^\primeH′ for their derivatives at x0x_0x0​, and assume H′H^\primeH′ is onto. If x0x_0x0​ is a local minimum or local maximum of fff relative to {x∣x∈U∨H(x)=0}\{x | x \in U \vee H (x) = 0\}{x∣x∈U∨H(x)=0}, then

H′(h)=0⟹f′(h)=0H^\prime(h)=0 \quad\Longrightarrow\quad f^\prime(h)=0H′(h)=0⟹f′(h)=0

for every direction hhh. This identifies the kernel of the linearized constraint as the true tangent test space and is the immediate analytic input to the multiplier factorization.

Preamble
import Mathlib

open Set
Formal statement
namespace VectorSpaceOpt

/-- Luenberger, Chapter 9, §9.3, Lemma 1. -/
theorem constrained_extremum_tangent_stationary
    {X Z : Type*}
    [NormedAddCommGroup X] [NormedSpace ℝ X] [CompleteSpace X]
    [NormedAddCommGroup Z] [NormedSpace ℝ Z] [CompleteSpace Z]
    (U : Set X) (f : X → ℝ) (H : X → Z) (x₀ : X)
    (f' : X →L[ℝ] ℝ) (H' : X →L[ℝ] Z)
    (hU : IsOpen U) (hx₀ : x₀ ∈ U) (hfeas : H x₀ = 0)
    (hf : ContDiffOn ℝ 1 f U) (hH : ContDiffOn ℝ 1 H U)
    (hf' : HasFDerivAt f f' x₀) (hH' : HasFDerivAt H H' x₀)
    (hregular : Function.Surjective H')
    (hextr : IsLocalExtrOn f {x | x ∈ U ∧ H x = 0} x₀) :
    ∀ h : X, H' h = 0 → f' h = 0 := by
  sorry

end VectorSpaceOpt
Source
David G. Luenberger, Optimization by Vector Space Methods (Wiley, 1969), Chapter 9, §9.3, Lemma 1, printed p. 242 (physical PDF p. 260). Scan: https://sites.science.oregonstate.edu/~show/old/142_Luenberger.pdf
Read-back

What the Lean code literally says, in plain math · gpt-5

Let XXX and ZZZ be real Banach spaces, U⊆XU⊆XU⊆X, f:X→Rf:X→ℝf:X→R, H:X→ZH:X→ZH:X→Z, x0∈Xx₀∈Xx0​∈X, and let f′:X→Rf':X→ℝf′:X→R and H′:X→ZH':X→ZH′:X→Z be continuous real-linear maps. Assume UUU is open, x0∈Ux₀∈Ux0​∈U, H(x0)=0H(x₀)=0H(x0​)=0, fff and HHH are continuously Fréchet differentiable through order one on UUU, their Fréchet derivatives at x0x₀x0​ are f′f'f′ and H′H'H′, H′H'H′ is surjective, and x0x₀x0​ is a local minimum or local maximum of fff relative to the points x∈Ux∈Ux∈U satisfying H(x)=0H(x)=0H(x)=0. Then every direction h∈Xh∈Xh∈X in the kernel of H′H'H′ also lies in the kernel of f′f'f′: H′h=0H'h=0H′h=0 implies f′h=0f'h=0f′h=0. The conclusion quantifies over all hhh, including h=0h=0h=0.

Human review
  • Endorsed by Shuze Chen · Aug 26, 2026

  • Endorsed by wenxinzhang · Aug 26, 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.

How Prove2Me worksResearch paper
SKILL.mdTourFAQContactJoin Slack© 2026 Prove2Me