Ported for Lean Pool: imports relocated and code adapted to Lean/Mathlib v4.34.0.
Exposed faces #
This module defines the project-local exposed-face and strictly exposed-point predicates and proves their basic compactness, convexity, and extremality properties.
A vector normal exposes the point p of a set s.
This is the exact form used in Rockafellar 25.6: the exposed face in direction normal is the
singleton {p}.
Equations
Instances For
The exposed face of s in the direction normal.
This is Rockafellar's notation ∂f(x)_y in set-theoretic form: the points of s where the
linear functional r ↦ normal · r attains its supremal value on s, written without mentioning
the supremum.
Equations
Instances For
The project-local exposed face is a Mathlib exposed set.
Exposed faces of compact sets are compact.
Exposed faces of convex sets are convex.
Exposed faces are extreme subsets of the original set.
A nonempty compact set has a nonempty exposed face in every direction.
Extreme points of an exposed face are extreme points of the original set.
If an exposed face is a singleton, its point is a Mathlib exposed point.