Abnormal equality multiplier
ProvedVectorSpaceOpt.abnormal_lagrange_multiplierIn the same Banach-space equality-constrained setting, retain feasibility, continuous Fréchet differentiability, and local extremality, but replace surjectivity of by the assumption that its range is closed. Then there exist a scalar and a continuous functional , not both zero, such that
The coefficient is allowed to vanish, so the assertion records the abnormal alternative rather than silently normalizing the objective multiplier. This corollary is not required for the regular root theorem, but it captures the exact boundary case highlighted by Luenberger and provides a reusable precursor to Fritz John and abnormal optimal-control conditions.
import Mathlib open Set
namespace VectorSpaceOpt
/-- Luenberger, Chapter 9, §9.3, Corollary 1. -/
theorem abnormal_lagrange_multiplier
{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)
(hrange : IsClosed (Set.range H'))
(hextr : IsLocalExtrOn f {x | x ∈ U ∧ H x = 0} x₀) :
∃ (r₀ : ℝ) (z₀ : Z →L[ℝ] ℝ), (r₀, z₀) ≠ (0, 0) ∧
r₀ • f' + z₀.comp H' = 0 := by
sorry
end VectorSpaceOptRead-back
What the Lean code literally says, in plain math · gpt-5
Let and be real Banach spaces, , , , , and let and be continuous real-linear maps. Assume is open, , , both and are continuously Fréchet differentiable through order one on , and their Fréchet derivatives at are respectively and . Assume also that the range of is closed and that is a local extremum—either a local minimum or a local maximum—of relative to the set of with . Then there exist a real scalar and a continuous real-linear functional , not simultaneously the zero scalar and zero functional, such that the continuous linear functional on is identically zero. No surjectivity of is assumed.
Confirmed by the mission captain (proposal self-audit).