HESSIAN MEASURES OF CONVEX FUNCTIONS AND APPLICATIONS TO AREA MEASURES

From MaRDI portal





Let \(u:\Omega\to \mathbb R\) be a convex function on an open convex set of \(\mathbb R^d\). Its Hessian measures \(F_k(u;\cdot)\), \(k=0,1,\dots,d\), are defined through a local Steiner formula presenting the \(d\)-measure of the set \[ P_t(u;\omega):=\{x+ty: x\in \omega\text{ and }y\in\partial u(x)\} \] as a polynomial in \(t>0\) of degree \(d\). Here \(\omega\subset\mathbb R^d\) is a Borel set and \(\partial u(x)\) denotes the subdifferential of \(u\) at \(x\). For \(u\in C^2(\Omega)\) there is an alternative way to present \(F_k\)'s using the eigenvalues of the Hessian matrix \(D^2u\). The authors present a geometric description of \(\text{supp}\,F_k\) in terms of \(k\)-extreme points of \(u\), and investigate the Radon-Nikodym derivative and absolute continuity of \(F_k\) with respect to the \(d\)-measure of \(\mathbb R^d\).











This page was built for publication: HESSIAN MEASURES OF CONVEX FUNCTIONS AND APPLICATIONS TO AREA MEASURES

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4673207)