Documentation

LeanPool.Superdiffusion.Algsuperdiff.Section5.Support.EssentialSupremum

EssentialSupremum #

Ported from scottnarmstrong/Superdiffusion at 7f20476385d4af418f110303085ad0740c9eba46. Module imports, support namespace and local option directives adapted for Lean Pool.

On an open set the essential supremum of a continuous function is its supremum #

The L^∞ norm of a Sobolev function on an open set is an essential supremum: it does not see a null set, so it is attached to the almost-everywhere class and not to any pointwise choice. When the class has a representative continuous on the open set, the essential supremum is the pointwise supremum of that representative: the two inequalities are the almost-everywhere bound in one direction and, in the other, the fact that a nonempty open subset of ℝ^d has positive volume, so no continuous function can exceed its essential supremum on a null set only.

This is what lets the localized error of Section 5.1 be stated as the honest essential supremum of the Sobolev functions while still being computed, where a continuous representative exists, from pointwise values.

Main results #

References #

On an open set a continuous function does not exceed its essential supremum. Otherwise it would exceed it on a nonempty open set, which has positive volume.

The L^∞ norm on an open set is the supremum of a continuous representative.

The same, reading the L^∞ norm of a function through a continuous function equal to it away from a null set.