On the generalized Stepanov theorem

From MaRDI portal





Let \(\Omega\) be a bounded domain in \({\mathbb R}^n\) and \(u:\Omega \rightarrow {\mathbb R}\) a bounded function. We say that \(u\in T_\infty^2(x)\) (resp. \(u\in t_\infty^2(x)\)) if there exists an affine function \(P_x\) (resp. a quadratic function \(Q_x\)) such that \[ \sup_{y\in B_r(x)\cap\Omega} | u(y)-P_x(y)| \leq O(r^2) \qquad\text{(resp.}\quad \sup_{y\in B_r(x)\cap\Omega} | u(y)-Q_x(y)| \leq o(r^2)) \] as \(r\rightarrow 0\). The generalized Stepanov theorem, due to \textit{A. P. Calderón} and \textit{A. Zygmund} [Stud. Math. 20, 171-225 (1961; Zbl 0099.30103)], states that if \(u\in T_\infty^2(x)\) for almost every \(x\in\Omega\), then \(u\in t_\infty^2(x)\) for almost every \(x\in\Omega\). Here the authors give a proof of this result using A. D. Aleksandrov's theorem on the almost everywhere twice differentiability of convex functions, rather than Whitney's extension theorem which was used by Calderón and Zygmund. In addition, a parabolic version of the result is proved. These results have applications to the study of differentiability properties of viscosity solutions of elliptic and parabolic partial differential equations.











This page was built for publication: On the generalized Stepanov theorem

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4372278)