The implicit function theorem with a moving level #
Let f : E₁ × E₂ → F be strictly differentiable at u with invertible partial derivative
∂₂f(u) : E₂ →L[𝕜] F. Mathlib's HasStrictFDerivAt.implicitFunctionOfProdDomain solves
f (z, v) = f u for v as a function of z, at the fixed level f u. Here the level w varies
as well: the joint implicit function ψ (z, w) solves f (z, ψ (z, w)) = w for (z, w) near
(u.1, f u). For each z it is a local inverse of v ↦ f (z, v), and it depends jointly on the
parameter z and the value w.
The construction applies Mathlib's ImplicitFunctionData for f and Prod.fst
(HasStrictFDerivAt.implicitFunctionDataOfProdDomain), that is, the local inverse of
(z, v) ↦ (f (z, v), z), and reads off the second coordinate. So ψ inherits the joint
regularity of that local inverse: it is strictly differentiable, C^n when f is (over ℝ or
ℂ), and analytic when f is. Over ℂ in finite dimensions this gives the holomorphic implicit
map of a holomorphic equation, see TauCeti.Analysis.Complex.ImplicitFunction.
Main results #
HasStrictFDerivAt.jointImplicitFunctionOfProdDomain: the joint implicit functionψ.HasStrictFDerivAt.jointImplicitFunctionOfProdDomain_apply_self:ψ (u.1, f u) = u.2.HasStrictFDerivAt.eventually_apply_jointImplicitFunctionOfProdDomain: near(u.1, f u),f (z, ψ (z, w)) = w.HasStrictFDerivAt.eventually_jointImplicitFunctionOfProdDomain_apply: nearu,ψ (z, f (z, v)) = v.HasStrictFDerivAt.eventually_apply_eq_iff_jointImplicitFunctionOfProdDomain: near(u, f u),f (z, v) = w ↔ ψ (z, w) = v.HasStrictFDerivAt.eventuallyEq_jointImplicitFunctionOfProdDomainandHasStrictFDerivAt.eventuallyEq_jointImplicitFunctionOfProdDomain_of_continuousAt: local uniqueness ofψ, from a left-inverse identity, or from a right-inverse identity together with continuity and the value at the base point.HasStrictFDerivAt.hasStrictFDerivAt_jointImplicitFunctionOfProdDomain: the derivative ofψat(u.1, f u)is(z, w) ↦ ∂₂f(u)⁻¹ (w - ∂₁f(u) z).HasStrictFDerivAt.contDiffAt_jointImplicitFunctionOfProdDomainandHasStrictFDerivAt.analyticAt_jointImplicitFunctionOfProdDomain:ψis as regular asf.
References #
- R. C. Gunning, H. Rossi, Analytic Functions of Several Complex Variables, Prentice-Hall (1965), Chapter I (the holomorphic inverse and implicit function theorems).
The joint implicit function ψ : E₁ × F → E₂ of f : E₁ × E₂ → F at u, defined near
(u.1, f u) by f (z, ψ (z, w)) = w. It is the second coordinate of the local inverse of
(z, v) ↦ (f (z, v), z) given by HasStrictFDerivAt.implicitFunctionDataOfProdDomain.
Instances For
At the fixed level f u, the joint implicit function is Mathlib's implicit function.
Near (u.1, f u), the joint implicit function solves f (z, ψ (z, w)) = w.
Near u, the joint implicit function inverts v ↦ f (z, v): ψ (z, f (z, v)) = v.
The joint implicit function takes the value u.2 at (u.1, f u).
Local uniqueness of solutions. For (z, v) near u and w near f u, the equation
f (z, v) = w holds exactly when v = ψ (z, w).
Local uniqueness of the joint implicit function, left-inverse form: a function φ with
φ (z, f (z, v)) = v for (z, v) near u agrees with ψ near (u.1, f u).
Local uniqueness of the joint implicit function, right-inverse form: a function φ,
continuous at (u.1, f u) with value u.2 there and with f (z, φ (z, w)) = w near
(u.1, f u), agrees with ψ near (u.1, f u).
The joint implicit function is strictly differentiable at (u.1, f u), with derivative
(z, w) ↦ ∂₂f(u)⁻¹ (w - ∂₁f(u) z).
The joint implicit function of an analytic equation is analytic.
The joint implicit function of a C^n equation is C^n.