The prime end capacity of inaccessible prime ends, resolutivity, and the Kellogg property (Q2332889)

The paper is a continuation of [the authors, Adv. Math. 238, 459--505 (2013; Zbl 1297.30056)] and studies the Marzurkiewicz boundary value problem for \(p\)-Laplace type equations in metric measure spaces. The Marzurkiewicz boundary value problem is more general than the Dirichlet problem allowing different boundary values on the opposite sides of the boundary ``wall''. The authors use the Marzurkiewicz metric \(d_M(x,y) = \inf_E \operatorname{diam}(E)\) where \(x\) and \(y\) are points in a bounded domain \(\Omega \subset X\) and \(E\) is a continuum joining \(x\) to \(y\). The prime end boundary \(\partial\Omega_P\) of \(\Omega\) is then defined as a collection of sequences of sets suitably identified in terms of the distance \(d_M\) and approaching the boundary of \(\Omega\). Important subsets of \(\partial\Omega_P\) are singletons, the subset of singletons which are rectifiably accessible along a rectifiable curve in \(\Omega\) and the complements of these sets. The authors define a natural topology on the prime end closure \(\Omega \cup \partial\Omega_P\) and then in an usual way the Newtonian type \(p\)-capacity \(\overline{C}^P_p(E)\) for sets \(E \subset \Omega \cup \partial\Omega_P\). The main result states that if \(F\) consists of non-singleton prime ends and singleton prime ends that are rectifiably inaccessible, then \(\overline{C}^P_p(F) = 0\). Another result says that if \(f: \partial\Omega_P \rightarrow \mathbb{R}\) is such that its restriction to the collection of rectifiably accessible prime ends is Lipschitz continuous with respect to \(d_M\), then \(f\) is resolutive and the Perron solution with boundary values \(f\) and the solution \(H_{\Omega}f\) defined by the variational principle coincide. In general, sufficient and necessary boundary conditions for the the Marzurkiewicz and the Dirichlet boundary value problems are not known in metric measure spaces and little is known about resolutivity, however, a sufficient condition for resolutivity is given in the important Lipschitz case. There is also a Kellogg-type result showing that certain irregular set in \(\partial\Omega_P\) has \(\overline{C}^P_p\)-capacity zero. Several examples illustrate the boundary behavior of domains (mostly subsets of \(\mathbb{R}^2\)) in \(X\) and the rather complicate concepts in prime end theory.