Localization of the flat Sobolev trace #
For 1 ≤ p < ∞, the trace of a smooth bounded multiplier times a W^{1,p} function is the
boundary restriction of the multiplier times the trace. This holds for both the whole-space
hyperplane trace and the one-sided half-space trace. In particular, equality of Sobolev values
wherever a multiplier is nonzero implies equality of their traces wherever it is nonzero on the
boundary.
These identities allow local boundary traces to be patched with a partition of unity: localizing before taking the trace agrees with localizing the boundary data. Bounds on both the cutoff and its gradient are explicit. Compact support is not required.
References #
- L. C. Evans, Partial Differential Equations, 2nd ed., §5.5, Theorem 1 (trace theorem background).
TauCeti.Analysis.Sobolev.Trace.Extension: flat Sobolev traces.TauCeti.Analysis.Sobolev.W1p.Multiplication: bounded smooth multiplication.- Mathlib's
ContinuousLinearMap.holderL:L∞ × Lᵖ → Lᵖmultiplication.
The whole-space W^{1,p} trace commutes with multiplication by a smooth function
whose value and gradient are bounded. The multiplier on the boundary is its classical
restriction to the hyperplane.
The one-sided W^{1,p} trace commutes with smooth bounded multiplication. No symmetry
of the multiplier across the boundary is assumed.
Half-space Sobolev functions agreeing almost everywhere wherever a smooth bounded multiplier is nonzero have equal traces almost everywhere wherever its boundary restriction is nonzero. This is the local agreement needed when gluing boundary charts.