Alphabeta Math
LemmaStatement: AI-generatedProof: AI-generated
How statement and proof provenance work

The first chip identifies the source of the statement or construction; the second identifies the source of its local proof or verification.

  • Literature-sourced: the exact statement appears in a cited source; only wording and notation differ.
  • AI-adapted: a semantically identical restatement of literature-sourced material, modulo indexing, notation, and boundary cases adopted by the library.
  • AI-generated: a genuinely novel statement formulated by AI, with no source for the claim itself.

These labels describe origin, not correctness: citations and verification chips remain separate evidence.

A compactly supported common L²/H² extension from a bounded smooth domain

Statement

Assume AC. Let G⊂Rn be bounded with smooth boundary, and let V be a fixed open neighborhood of G‾. There is one linear map EG:L2(G)→L2(Rn), independent of an endpoint choice, with rGEG=I, bounded also W2,2(G)→W2,2(Rn), such that supp⁡EGu⊂K⋐V for one fixed compact set K. Moreover the weak space W2,2(G) equals the restriction quotient of whole-space H2(Rn) with equivalent norms; at order zero the restriction quotient is isometrically L2(G).

Here n≥1, G is nonempty and open, and smooth boundary means the one-sided graph convention at every finite order, including the finite-union-of-intervals convention when n=1. Fix K∈{R,C}; all maps and weak spaces are over K. The weak spaces and their derivative-sum norms are those of the integer Sobolev definition. Whole-space Hj(Rn;C), j∈{0,2}, is the Bessel-potential completion, identified by the integer Fourier theorem with its unique L2 class; its norm is ∥⟨ξ⟩jF2F∥2 with ⟨ξ⟩=(1+∣ξ∣2)1/2. For real scalars, take its real-valued L2 subspace with this inherited norm. Thus H0(Rn;K)=L2(Rn;K) isometrically.

Precisely, rGF=F∣G is restriction of the L2 class, and the endpoint restriction space and its infimum norm mean Rj(G;K)={rGF:F∈Hj(Rn;K)}⊆L2(G;K),∥u∥Rj=inf⁡F∈Hj(Rn;K)rGF=u a.e.∥F∥Hj,j∈{0,2}. Equality of restrictions is always equality almost everywhere on G. Equivalently, this is the quotient by the kernel of rG. The assertions above are R2(G;K)=W2,2(G;K) with equivalent norms and R0(G;K)=L2(G;K) isometrically.

Facts & Assumptions

Given: The hypotheses and conventions of the Statement.

[F1]

Assume the Axiom of Choice The Axiom of Choice; in particular its countable and dependent instances are available. AC selects a graph chart at each boundary point before the finite subcover is taken; its countable instance is inherited through the Sobolev, weak-chain and Fourier interfaces. The subsequent finite cutoffs and the explicit reflection require no further choice principle.

[F2]

One second-order flat reflection on L² and H² supplies one flat formula bounded at both the zeroth and second weak-derivative endpoints and preserving the smaller-chart support.

[F3]

Bounded Ck domains and boundary charts supplies the one-sided smooth graph charts and their compact-patch derivative bounds. Heine-Borel in ℝⁿ: with the Euclidean metric a subset of ℝⁿ is compact if and only if it is closed and bounded, and the proof by bisection uses no choice principle; the same holds on the real line makes G‾ and ∂G compact. Step 1.1 below derives the finite nested chart cover from these inputs. Test function cutoffs and euclidean localization supplies cutoffs and subordinate partitions. The construction and matched-half-patch exhaustion in proof steps 1.1, 2.1 and 2.2 of Bounded Ck domains admit integer-order Sobolev extension are applicable with k=p=2; we use their localization with the explicit common reflection, without asserting that separate endpoint extension operators agree.

[F4]

Ck boundary flattening preserves local Wk,p gives the weak chain formula for k=2 and change of variables on fixed compact patches, Weak Leibniz rule with a smooth factor gives cutoff derivative bounds through order two, and Compactly supported Sobolev functions extend by zero in every integer order gives zero extension of the compact interior and compact ambient-chart pieces.

[F5]

Integer-order W^k,2 and Hᵏ agree with equivalent norms identifies Wj,2(Rn;C) with the Bessel-potential Hj via the unique regular-distribution L2 representative, for j=0,2. Its norms are equivalent at j=2 and equal at j=0, with the prescribed 2π Fourier normalization. Restricting this identification to real-valued classes gives the stated real convention with the same bounds.

[F6]

Integer-order Sobolev spaces and their norms defines the weak spaces over R and C as almost-everywhere L2 classes with all weak derivatives through the stated integer order in L2, with the derivative-sum norm and W0,2=L2. Real-order Bessel-potential completion Hˢ defines whole-space Hj as the weighted-Schwartz completion; F5 supplies the L2 identification used in the local definition of Rj.

[F7]

Bounded restriction and cutoff localisation in Sobolev spaces gives restriction to every open subset as a contraction on Wk,p for every admitted order and exponent, over either scalar field, and identifies the restricted weak derivatives. We use k=0,2 and p=2.

Proof

Goal: Establish every assertion of the Statement under the stated hypotheses.

Given: The hypotheses and data in the Statement.

1.1F1F2F3given

At each x∈∂G, choose a graph chart in the convention of F3, reversing its normal coordinate so that G maps to t>0. Since its flattened image is open and contains the image of x, choose three nested doubled cylinders centered there, with successively larger positive tangential and normal radii, whose outer closure lies inside that image and whose inverse image lies compactly in V. For n=1 only the normal intervals occur. The inverse images of the innermost cylinders form a boundary cover; compactness of ∂G gives finitely many. The compact set left in G‾ after removing these innermost patches misses ∂G and lies in G, so finitely many balls with closures in G give a relatively compact interior open set covering it. Apply the partition supplier to the resulting finite cover of a neighborhood N of G‾. Choose a compact neighborhood of G‾ in N; local finiteness leaves only finitely many partition pieces meeting it. Group those by cover member and extend their compactly supported sums by zero outside N. The resulting ambient smooth ρ0,…,ρm sum to 1 on G, with supp⁡ρ0⋐G and the other supports compactly in the innermost patches. The intervening and outer cylinders provide the positive margins required by the common reflection. Choose ψi compactly supported in each outer chart and equal to 1 near supp⁡ρi. All these choices precede u, and their finitely many supports lie compactly in V.

2.1F3F4F6F7step 1.1

For i>0 flatten ρiu to the positive half-cylinder and extend it by zero across only the lateral and top artificial edges, as a function fi on H+. At order zero, change of variables bounds this map on L2. At order two the weak chain rule has first derivatives bounded by a constant times the first input derivatives, and second derivatives bounded by a constant times the first and second input derivatives; these constants depend on the bounded first and second derivatives of the chart and its inverse on the larger closed patch. To handle the half-patch reaching t=0, apply the local chain formula on an exhaustion of matched interior subpatches. The same ambient bounds apply on every subpatch; all tests are covered and increasing the integrals gives the norm bound on the whole half-patch. The localized function vanishes near the artificial edges, so tests in H+ prove that its zero extension across those edges has the corresponding weak derivatives extended by zero. No test in H+ crosses the true boundary t=0.

3.1F2F3F4step 2.1

Apply the common reflection to fi, pull the result back to the larger doubled chart, multiply by ψi and extend the product by zero outside the ambient chart. Call this Eiu. The support assertion of the flat lemma and the chosen chart margins keep the reflected support in the larger chart. Chart, reflection and cutoff bounds give ∥Eiu∥2≤Ci∥u∥L2(G) and ∥Eiu∥W2,2≤Ci′∥u∥W2,2(G). The same operations act on L2 representatives and on their weakly differentiable subclass; all coordinate maps preserve null sets. On G, the reflected positive side is unchanged and ψiρi=ρi, so rGEiu=ρiu.

4.1F3F4step 1.1step 3.1algebra

Let E0u be the zero extension of ρ0u and define EGu=E0u+∑i=1mEiu. This is one linear formula on all of L2(G); the preceding estimates give both endpoint bounds for its restriction to W2,2(G). Restriction gives rGEGu=u. Its support is contained in the fixed union of the compact interior and ambient-chart cutoffs, so the same compact K⋐V works for every input.

5.1F5F6F7step 4.1algebra∎

If u=rGF∈R2(G;K), F5 identifies F with its weak W2,2 class and F7 restricts that class and each of its derivatives to G, giving ∥u∥W2,2(G)≤C∥F∥H2. Conversely, if u∈W2,2(G;K), step 4.1 and F5 give the whole-space extension EGu∈H2 with ∥EGu∥H2≤C′∥u∥W2,2(G). In the real case the same construction preserves real values, so this extension also lies in the specified real subspace. Infimizing the first bound over exactly the extensions in the definition of R2 and using EGu for the second gives C−1∥u∥W2,2(G)≤∥u∥R2≤C′∥u∥W2,2(G). This proves equality of the spaces and equivalence of norms; definiteness of the infimum follows also from ∥u∥L2(G)≤∥F∥H2, while homogeneity and the triangle inequality follow by scaling and adding extensions and then taking infima. At order zero F5 and F7 give ∥u∥2≤∥F∥H0 for every extension. Extension by zero of any L2(G) class has exactly the same L2=H0 norm, since its squared integral outside G is zero. Thus every such class belongs to R0 and its infimum norm is exactly ∥u∥2.

Depends on

Used by · 2

Dependency tree · 2 levels

Each result is drawn at its shortest distance from this item. Arrows run from prerequisites to the items that use them; this item has a heavier outline. External source nodes link directly to the cited text. The view shows up to two dependency steps; open a prerequisite to continue.