The maximizer step shared by the lower-order weak maximum principles #
Both lower-order weak maximum principles — for -Δ + c in
TauCeti.Analysis.InnerProductSpace.Laplacian.ZerothOrderMaximumPrinciple and for
-Δ - b·∇ + c in TauCeti.Analysis.InnerProductSpace.Laplacian.LowerOrderMaximumPrinciple —
perturb a subsolution by a barrier, take a maximizer of the perturbation over the compact set, and
then argue that the frontier bound already holds at that maximizer. Only the last step is common:
the two differ in how they produce the maximizer and in which barrier they use.
This module holds that step, stated once for the operator Δ + b·∇ with an arbitrary barrier. The
-Δ + c principle runs it at b = 0 with the quadratic barrier ‖·‖².
It is a support module: TauCeti.le_of_isMaxOn_add_smul is support API for those two proofs rather
than a result the roadmap asks for. Both principles import this module non-publicly, so neither
re-exports it and nothing downstream of either principle sees the step in its interface.
The maximizer step of the lower-order weak maximum principles. Let z be a maximizer of
f + ε • w over K. If f z ≤ m whenever z is a frontier point, and, whenever z is interior
and f z exceeds m, the operator Δ + ∇_v is nonnegative on f at z while the barrier w is
C² and strictly positive under that operator at z, then f z ≤ m.
The direction is a single vector v, and the regularity and operator-sign hypotheses hcd,
hLf0, hwcd, hwpos are asked for at z alone and only when z is interior. A caller holding
the usual ∀ x ∈ interior K form instantiates it at z, as fun h => hcd h; a caller with a drift
field b passes b z.