Derivative-test variations force a constant coefficient
ProvedVectorSpaceOpt.derivative_test_implies_constantFix and a continuous function . Suppose that, for every function satisfying , one has
Then there is a real constant such that for every . The theorem includes the closed-interval endpoints in its conclusion because continuity extends the interior constancy. This is Luenberger's second fundamental lemma. It isolates the derivative-only component of a first variation and is reusable whenever a continuous coefficient annihilates all derivatives of fixed-endpoint test functions. It also handles the boundary values cleanly through the assumed continuity on the closed segment.
import Definitions.Def_VectorSpaceOpt_isC1OnSegment open Set intervalIntegral
namespace VectorSpaceOpt
/-- Luenberger, Chapter 7, §7.5, Lemma 2. -/
theorem derivative_test_implies_constant
(a b : ℝ) (hab : a < b) (β : ℝ → ℝ)
(hβ : ContinuousOn β (Icc a b))
(hzero : ∀ (h dh : ℝ → ℝ),
IsC1OnSegment a b h dh → h a = 0 → h b = 0 →
(∫ t in a..b, β t * dh t) = 0) :
∃ c : ℝ, ∀ t ∈ Icc a b, β t = c := by sorry
end VectorSpaceOptRead-back
What the Lean code literally says, in plain math · gpt-5
For real numbers and a function continuous on , assume that for every pair of functions , if and are continuous on , has derivative at every , and , then the oriented interval integral from to of is zero. The conclusion is that there exists a real constant such that for every . The quantified test functions are constrained only as stated on the interval; in particular, no endpoint derivative is required.
Confirmed by the mission captain (proposal self-audit).