Support-function subgradients as exposed faces
ID: support-function-subgradients-as-exposed-faces
Let be a nonempty closed convex set in finite dimensions. At any where its support function is finite,Indeed for the indicator functional of a constraint set. The equality case of the Fenchel–Young inequality says exactly when , which is the displayed maximizing condition.
If is compact, the maximum is attained and this set is nonempty. For an unbounded closed convex set, a finite supremum need not be attained. Take , the epigraph of an exponential function, and . Then , approached as , but no point of has . The subdifferential at that direction is therefore empty. The support function must be considered at its finite-valued directions, rather than assumed finite everywhere for every closed convex set.
New to topics? Read the docs here!