Subdifferentials of convex symmetric functions: An application of the inequality of Hardy, Littlewood, and Pólya (Q1118131)

The authors consider Schur functions \(f:L^ p(\mu)\to {\mathbb{R}}\), where \(\mu\) is a nonatomic probability measure. They prove that any subgradient of a convex Schur function is ``arranged similarly'' to the point where it is taken, thus extending Schur's classical result for differentiable Schur functions on \({\mathbb{R}}^ n\). They go on to compute the subdifferential of the ``basic convex Schur function'', so called because any weakly lower semicontinuous convex Schur function can be written as the pointwise supremum of a family of its translates. This computation enables them to characterize the extreme points of certain subdifferential sets. References to the economic literature on electricity pricing and investment are given.