Backward space-time mollification #
The regularity bootstrap of thm:vorticity-regularity of the Escauriaza–Seregin–Šverák manuscript
works on cylinders whose top is the
terminal time of the hypotheses. A space-time mollification evaluated 5ε earlier in time only
uses values of the field strictly below the top, so smooth approximations are available up to
the top face. This file records the derivative formulas for kernels, the pairing identities with
the reflected kernel, and the resulting pointwise form of constant-coefficient weak identities.
A directional derivative of a smooth compactly supported kernel is again smooth.
A directional derivative of a compactly supported kernel is compactly supported.
A scalar convolution is the integral of the reflected kernel against the function.
The directional derivative of a convolution with a smooth compactly supported kernel is the convolution with the directional derivative of the kernel.
The time shift (0, 5ε) of the backward mollifier.
Instances For
Backward mollification on W: the zero extension of f outside W, convolved with the
space-time kernel of radius ε and evaluated at the earlier time t - 5ε.
Equations
- CKN.vorticityBackMollify W f ε hε z = CKN.spaceTimeMollify (W.indicator f) ε hε (z - CKN.vorticityBackShift ε)
Instances For
The reflected kernel centred at the shifted point: pairing a field with it evaluates the
backward mollification at z.
Equations
- CKN.vorticityBackTest ε hε z y = CKN.spaceTimeMollifier ε hε (z - CKN.vorticityBackShift ε - y)
Instances For
Backward mollification of an integrable zero extension is smooth.
The first derivative of a backward mollification is the backward convolution with the derivative of the kernel.
The second derivative of a backward mollification is the backward convolution with the second derivative of the kernel.
The backward test is smooth.
The backward test is supported in the closed ball of radius ε about the shifted point.
The backward test has compact support.
Pairing the zero extension with the backward test evaluates the backward mollification.
Pairing the zero extension with a first derivative of the backward test gives minus the corresponding derivative of the backward mollification.
Pairing the zero extension with a second derivative of the backward test gives the corresponding second derivative of the backward mollification.
Translation is continuous in L².
Backward mollifications converge in L² to the zero extension.
Backward mollification preserves a pointwise bound of the zero extension.