First-order weak Sobolev spaces #
This file constructs the first-order, real-valued Sobolev space W^{1,p}(Ω) on an open subset
of a finite-dimensional real inner product space E. An element is an Lᵖ value-gradient jet
(u, ∇u) satisfying the distributional integration-by-parts pairing against test functions, and
this closed-subspace definition is identified with the weak Fréchet derivative predicate
TauCeti.HasWeakFDerivOn. The definitions do not assume finite dimension, but on an
infinite-dimensional E every compactly supported continuous function vanishes
(HasCompactSupport.eq_zero_or_finiteDimensional), so every test function is zero and every jet
is a member.
The quotient issue is handled at the definition boundary. Both components of a jet are Lp
classes for μ.restrict Ω, and the weak relation is the one of the generic closed
weak-derivative graph TauCeti.weakDerivStepSubmodule over Lᵖ(Ω) with the identity base:
TauCeti.w1pSubmodule is its preimage under the continuous linear map that keeps the value and
reads the gradient as the field of functionals ⟪·, ∇u x⟫.
Consequently the admissible jets form an intersection of kernels of continuous linear
functionals. This makes TauCeti.W1p a closed subspace of the ambient Bochner Lᵖ space, and
hence complete when E is complete.
The theorem TauCeti.mem_w1pSubmodule_iff_hasWeakFDerivOn identifies this closed subspace
definition with the weak-derivative predicate, so the construction does not replace the
distributional condition by a merely formal closedness assumption.
The pointwise jet uses the Euclidean product norm on ℝ × E. Thus at p = 2 the inherited norm
is the usual Hilbert norm
(‖u‖²₂ + ‖∇u‖²₂)¹⁄²,
which is the space needed for energy methods in PDE. No boundedness or
boundary regularity of Ω is used.
Main declarations #
TauCeti.Sobolev1Jet: the value-gradient fibreℝ × Ewith its Euclidean product norm.TauCeti.w1pSubmodule: the closed subspace of jets annihilating every weak-derivative test.TauCeti.W1p: the corresponding normed space (complete whenEis complete).TauCeti.mem_w1pSubmodule_iff_hasWeakFDerivOn: membership is exactly weak differentiability of the value component with the recorded gradient.TauCeti.W1p.valueLandTauCeti.W1p.gradientL: the two components as continuous linear projections from the Sobolev space, withTauCeti.W1p.value_coeandTauCeti.W1p.gradient_coeidentifying them with the components of the ambient jet.TauCeti.W1p.locallyIntegrableOn_gradient: the weak gradient is locally integrable onΩ.TauCeti.W1p.gradient_ae_eq_zero_of_value_ae_eq_zero: the weak gradient vanishes wherever the value vanishes on an open subset.TauCeti.W1p.ofExponentLE: on a domain of finite measure,W^{1,q}(Ω) ⊆ W^{1,p}(Ω)forp ≤ q.
References #
The graph-space construction and completeness argument follow L. C. Evans, Partial Differential
Equations, Chapter 5, §5.2. The continuous annihilator presentation is the quotient-respecting
version of the standard proof that weak differentiation is a closed operator on Lᵖ.
Bochner Lᵖ Sobolev jets #
The fibre of a first-order scalar Sobolev jet: a value and its gradient, with the Euclidean product norm.
Equations
- TauCeti.Sobolev1Jet E = WithLp 2 (ℝ × E)
Instances For
The ambient Bochner Lᵖ space of value-gradient jets on Ω.
Equations
- TauCeti.Sobolev1JetLp mu Omega p = MeasureTheory.Lp (TauCeti.Sobolev1Jet E) p (mu.restrict ↑Omega)
Instances For
The continuous linear projection from an Lᵖ Sobolev jet to its value component.
Equations
- TauCeti.Sobolev1JetLp.valueL = ContinuousLinearMap.compLpL p (mu.restrict ↑Omega) (WithLp.fstL 2 ℝ ℝ E)
Instances For
The value component of an Lᵖ Sobolev jet.
Equations
Instances For
Applying the bundled value projection gives the value component of a Sobolev jet.
The bundled value projection is postcomposition with the first projection of the fibre.
The continuous linear projection from an Lᵖ Sobolev jet to its gradient component.
Equations
- TauCeti.Sobolev1JetLp.gradientL = ContinuousLinearMap.compLpL p (mu.restrict ↑Omega) (WithLp.sndL 2 ℝ ℝ E)
Instances For
The gradient component of an Lᵖ Sobolev jet.
Instances For
Applying the bundled gradient projection gives the gradient component of a Sobolev jet.
The bundled gradient projection is postcomposition with the second projection of the fibre.
The value component of a Sobolev jet is, almost everywhere on Ω, the first coordinate of the
jet.
The gradient component of a Sobolev jet is, almost everywhere on Ω, the second coordinate of
the jet.
Two Sobolev jets are equal when their value and gradient components are equal.
The candidate weak Fréchet derivative recorded by the gradient component of a Sobolev jet.
Equations
- TauCeti.Sobolev1JetLp.candidateWeakFDeriv J x = (innerSL ℝ) (↑↑(TauCeti.Sobolev1JetLp.gradient J) x)
Instances For
The weak Sobolev space W^{1,p}(Ω) #
The first-order weak Sobolev subspace: the preimage of the closed weak-derivative graph
TauCeti.weakDerivStepSubmodule over Lᵖ(Ω), with the identity base, under the map reading the
gradient of a jet as a field of functionals. Its members are the Lᵖ value-gradient jets
satisfying the weak integration-by-parts identity against every test function
(TauCeti.mem_w1pSubmodule_iff).
Equations
- One or more equations did not get rendered due to their size.
Instances For
Membership in w1pSubmodule is the family of weak integration-by-parts identities.
The first-order, real-valued weak Sobolev space W^{1,p}(Ω), represented by its value and
weak gradient.
Equations
- TauCeti.W1p mu Omega p = ↑(TauCeti.w1pSubmodule mu Omega p)
Instances For
W^{1,p}(Ω) is complete in its value-gradient graph norm.
The continuous linear projection from W1p to its Lᵖ value component.
Equations
- TauCeti.W1p.valueL = TauCeti.Sobolev1JetLp.valueL ∘SL (↑(TauCeti.w1pSubmodule mu Omega p)).subtypeL
Instances For
The Lᵖ value component of a Sobolev function.
Equations
Instances For
The continuous linear projection from W1p to its Lᵖ weak-gradient component.
Equations
- TauCeti.W1p.gradientL = TauCeti.Sobolev1JetLp.gradientL ∘SL (↑(TauCeti.w1pSubmodule mu Omega p)).subtypeL
Instances For
The Lᵖ weak-gradient component of a Sobolev function.
Equations
Instances For
W1p.valueL is the ambient jet projection precomposed with the inclusion, so the Sobolev
value component is the value component of the underlying ambient jet.
The Sobolev value component agrees almost everywhere with the first component of its ambient value-gradient jet.
The value of a finite sum of Sobolev functions is almost everywhere the sum of their values.
As for TauCeti.W1p.value_coe: the Sobolev gradient component is the gradient component of
the underlying ambient jet.
The Sobolev gradient component agrees almost everywhere with the second component of its ambient value-gradient jet.
Two Sobolev functions are equal when their value and weak-gradient components are equal.
The norm of a Sobolev function controls the norm of its value component.
The norm of a Sobolev function controls the norm of its weak gradient.
At exponent two, the norm on W1p is the Hilbert graph norm.
The squared pointwise norm of a Sobolev gradient is integrable.
The squared pointwise value of a real Sobolev function is integrable.
The integral of the squared Sobolev gradient is its squared L² norm.
The integral of the squared Sobolev value is its squared L² norm.
The L² pairing of the value components of two Sobolev functions, as an integral over Ω.
Convergence in the first-order Sobolev norm is equivalent to convergence of both the
value and the weak gradient in Lᵖ.
Identification with weak Fréchet derivatives #
A jet belongs to w1pSubmodule exactly when its value component has the recorded gradient as
its weak Fréchet derivative.
Construct a Sobolev function from its Lᵖ value and weak-gradient components.
Equations
- TauCeti.W1p.mk u g h = ⟨TauCeti.assembleSobolev1JetLp✝ u g, ⋯⟩
Instances For
The value component of W1p.mk u g h is u.
The gradient component of W1p.mk u g h is g.
The value-gradient pair represented by an element of W1p satisfies the weak derivative
identity.
The weak gradient of a Sobolev function is locally integrable on the domain, as its value component is.
A Sobolev function vanishing on an open subset has vanishing weak gradient there. The
weak gradient is determined almost everywhere by the function on every open set
(TauCeti.HasWeakFDerivOn.ae_eq), and the zero function has zero weak gradient.
Two Sobolev functions are equal when their Lᵖ value components are equal. Uniqueness of
weak derivatives determines the gradient component.
On a domain of finite measure, W^{1,q}(Ω) ⊆ W^{1,p}(Ω) for p ≤ q: the value and the
weak gradient of u ∈ W^{1,q}(Ω) are also in Lᵖ(Ω), and are still related by the weak
derivative identity.
Instances For
The underlying ambient Lᵖ jet of W1p.ofExponentLE is the canonical subtype inclusion of the
underlying L^q jet via Lp.antitone.
Lowering the exponent does not change the value of a Sobolev function.
Lowering the exponent does not change the weak gradient of a Sobolev function.
Coercion of W1p.ofExponentLE to the ambient jet space preserves zero.
Lowering the exponent sends zero to zero.
Coercion of W1p.ofExponentLE to the ambient jet space preserves addition.
Lowering the exponent preserves addition.
Coercion of W1p.ofExponentLE to the ambient jet space preserves real scalar multiplication.
Lowering the exponent preserves real scalar multiplication.
Coercion of W1p.ofExponentLE to the ambient jet space at equal exponents is the identity.
Lowering the exponent from p to itself is the identity.
Coercions of composed ambient exponent inclusions compose transitively.
Exponent inclusions compose transitively.