Coarse-graining support: Support.Multiscale.NormalizedNorms #
Imported from the Apache-2.0 CoarseGraining development at commit
c7ddd76c08ade64fed1b8d2ca51be14dfee8deb4.
Normalized cube L^p and W^{1,p} quantities used later in the Besov layer.
The normalization is packaged through a probability measure on cubeSet Q, so
the underlined norms match the note conventions without repeatedly rederiving
factors of cubeVolume Q.
Lebesgue measure restricted to the cube.
Equations
Instances For
Cube Lebesgue measure divided by the cube's volume.
Equations
Instances For
The real value of the Lᵖ seminorm for normalized cube measure, with infinity mapped to
zero.
Equations
Instances For
The function obtained by subtracting its average over the cube.
Equations
- HCPolySupport.cubeFluctuation Q f x = f x - HCPolySupport.cubeAverage Q f
Instances For
The normalized cube Lᵖ norm of the supplied gradient field.
Equations
- HCPolySupport.cubeW1pSeminorm Q p Du = HCPolySupport.cubeLpNorm Q p Du
Instances For
The maximum of the gradient's L∞ norm and the value's L∞ norm divided by side length.
Equations
- HCPolySupport.cubeW1InfinityNorm Q u Du = max (HCPolySupport.cubeLpNorm Q ⊤ Du) ((HCPolySupport.cubeScaleFactor Q)⁻¹ * HCPolySupport.cubeLpNorm Q ⊤ u)
Instances For
The scale-normalized Sobolev norm combining gradient and value Lᵖ norms; it is zero for p = 0 and uses the maximum formula for p = ∞.
Equations
- One or more equations did not get rendered due to their size.