Searcharxiv⌕ Search

arXiv subjects

Kathrin Völkner

Publications and source records attributed to Kathrin Völkner.

2 recordsLinked to original sources

Structure theory of $BV$ functions and finite perimeter sets on Riemannian manifolds

We develop the theory of functions of bounded variation and the structure theory of finite-perimeter sets on arbitrary Riemannian manifolds without relying on global curvature bounds or completeness of the manifold. To this end, we build a localization framework that permits a synthesis of techniques from Euclidean geometric measure theory and analysis on metric measure spaces while preserving genuinely Riemannian features, such as polar and normal vector fields, reduced boundaries, and approximate tangent spaces. As a consequence, we recover key results of the Euclidean theory, such as a differentiation theorem for a Riemannian generalization of vector-valued measures and the structure theorems by De Giorgi and Federer, formulated intrinsically on Riemannian manifolds. This makes a large portion of the classical Euclidean $BV$ theory available to the Riemannian setting. We demonstrate this by providing the theoretical background for boundary value problems on domains in manifolds, including trace and Gauss-Green theorems. Finally, we prove an approximation result for finite-perimeter sets in a strict sense that respects a prescribed Dirichlet boundary portion of a given ambient domain and apply this to prove Gamma-convergence of a family of energy functionals for capillarity problems with mixed boundary conditions.

math.DG↗

First-order conditions for the optimal control of learning-informed nonsmooth PDEs

In this paper we study the optimal control of a class of semilinear elliptic partial differential equations which have nonlinear constituents that are only accessible by data and are approximated by nonsmooth ReLU neural networks. The optimal control problem is studied in detail. In particular, the existence and uniqueness of the state equation are shown, and continuity as well as directional differentiability properties of the corresponding control-to-state map are established. Based on approximation capabilities of the pertinent networks, we address fundamental questions regarding approximating properties of the learning-informed control-to-state map and the solution of the corresponding optimal control problem. Finally, several stationarity conditions are derived based on different notions of generalized differentiability.

math.OC↗