PenalizedBoundaryComparison #
Ported from scottnarmstrong/Superdiffusion at 7f20476385d4af418f110303085ad0740c9eba46. Module imports, support namespace and local option directives adapted for Lean Pool.
Boundary-collar comparison for potential equations #
The whole-space penalization interchange compares two Dirichlet solutions on
nested cubes. This module isolates the weak maximum-principle step: if the
larger solution is at most eps on a collar of the smaller boundary, then it
is at most the smaller solution plus eps throughout the smaller domain.
Replace the pointwise representative of an H¹ function without
changing its weak gradient or its equivalence class on the domain.
Equations
- DivergenceFormProcess.Form.h1FunctionWithAeRepresentative u v hvu = { toFun := v, grad := u.grad, memL2 := ⋯, gradMemL2 := ⋯, hasWeakGradient := ⋯ }
Instances For
A scalar-forced equation is unchanged when both its solution and forcing are replaced by almost-everywhere equal representatives.
A scalar-forced weak equation restricts to an open subdomain.
Subtracting two equations with the same source produces the homogeneous potential equation for their difference.
A homogeneous potential difference with a compactly supported positive part is bounded by its boundary-collar level.
Continuous representatives upgrade the boundary-collar comparison from almost everywhere to every point of the open domain.