Smooth local level parametrization with derivative the tangent inclusion
ProvedImplicitCalculus.smooth_level_parametrizationdifferential-geometryimplicit-function-theorem
Let V be a real Banach space and W a finite-dimensional real normed space. Suppose F from V to W is smooth at y and DF at y is surjective. Write K for its kernel. There exists a map gamma from K into V, smooth at zero, such that
The derivative is the canonical kernel inclusion. This is a local level parametrization; compactness, global regularity, contact forms, and a finite-dimensional source are unnecessary.
Preamble
import Mathlib.Analysis.Calculus.ImplicitContDiff set_option autoImplicit false open Filter open scoped ContDiff Topology
Formal statement
theorem ImplicitCalculus.smooth_level_parametrization
{V W : Type*} [NormedAddCommGroup V] [NormedSpace ℝ V] [CompleteSpace V]
[NormedAddCommGroup W] [NormedSpace ℝ W] [FiniteDimensional ℝ W]
(F : V → W) (y : V) (hF : ContDiffAt ℝ ∞ F y)
(hD : Function.Surjective (fderiv ℝ F y)) :
∃ γ : (fderiv ℝ F y).ker → V,
ContDiffAt ℝ ∞ γ 0 ∧ γ 0 = y ∧
HasFDerivAt γ (fderiv ℝ F y).ker.subtypeL 0 ∧
∀ᶠ z in 𝓝 (0 : (fderiv ℝ F y).ker), F (γ z) = F y := by sorrySource
Smooth implicit-function theorem. Mathlib 0df444a360eaa60ab8c11dca51a86af692955474, https://github.com/leanprover-community/mathlib4/blob/0df444a360eaa60ab8c11dca51a86af692955474/Mathlib/Analysis/Calculus/ImplicitContDiff.lean, ImplicitFunctionData.contDiffAt_implicitFunction; Implicit.lean, to_implicitFunctionOfComplemented and map_implicitFunctionOfComplemented_eq.