The scale-covariant local triple convolution #
This file closes the one analytic predicate deliberately isolated in
Continuum/CellSingular.lean. The proof uses the following
four-dimensional scaling argument.
- the
3/2power of the inverse-square kernel has local massO(r); - pointwise AM--GM bounds a product of three inverse-square kernels by
the sum of the three products of two
3/2powers; - in the all-near cell configuration, every relevant cell is contained
in a ball of radius
6rabout the adjacent endpoint.
Thus each of the three separated products costs O(r) · O(r) = O(r²).
The local 3/2-power estimate is proved by transporting to the
fundamental cube and evaluating the radial integral
∫_{‖x‖<r} ‖x‖⁻³ dx = 4 vol(B₁) r in dimension four.
The subcritical 3/2 power #
The 3/2 power of the inverse-square kernel, i.e. the model
|z|⁻³ singularity.
Equations
- Anderson4D.invSqKerThreeHalf z = Anderson4D.invSqKer z ^ (3 / 2)
Instances For
Cube transport for the local mass #
The |z|⁻³ kernel is integrable on the torus.
Translation invariance of the total |z|⁻³ mass.
Uniform local |z|⁻³ mass. A radius-r torus ball
about the singularity has mass O(r), uniformly in its centre.
Pointwise AM--GM and cell geometry #
Closure of the all-near predicate #
The local three-edge cell convolution has the sharp O(r²) scale
whenever its four consecutive cell centres are 4r-near.
The isolated analytic predicate from CellSingular.lean is true.