Singular integral operators, quantitative flatness, and boundary problems (Q2673450)

The book describes results for Calderon-Zygmund singular integrals with a certain algebraic structure, for \(\delta\)-AR domains in \(\mathbb{R}^n\), also known as \(\delta\)-flat Ahlfors regular domains. These are a subcategory of the class of two-sided nontangentially accessible (NTA) Ahlfors regular domains, and completes work for SKT (Semmes-Kenig-Toro) domains. SKT was designed for domains with compact boundaries, while their results apply to the case of a connected unbounded domain with a connected unbounded boundary and an unbounded connected complement. They summarize it by saying they explore the implications of demanding that the Gauss map \(\partial \Omega \ni x \mapsto \nu(x) \in S^{n-1}\) has small BMO norm. \textit{C. E. Kenig} [Harmonic analysis techniques for second order elliptic boundary value problems: dedicated to the memory of Professor Antoni Zygmund. Providence, RI: American Mathematical Society (1994; Zbl 0812.35001)] asked for a proof that double layer potentials are invertible in appropriate spaces in suitable subclasses of uniformly rectifiable domains. He had noted that there were some very general classes of open sets \(\Omega \subseteq \mathbb{R}^n\) such that if \(\sigma = \mathcal{H}^{n-1}(\partial \Omega)\), where \(\mathcal{H}^{n-1}\) is the \((n - 1)\)-dimensional Hausdorff measure in \(\mathbb{R}^n\) (this means throughout that \(\sigma\) is the measure induced by Hausdorff measure), layer potentials are bounded operators on \(L^p(\Omega, \sigma)\) for each \(p \in (1, \infty)\), whenever \(\Omega\) is an open set with a uniformly rectifiable boundary [\textit{G. David} and \textit{S. Semmes}, Singular integrals and rectifiable sets in \(\mathbb{R}^ n\). Au-delà des graphes lipschitziens. Montrouge: Société Mathématique de France (1991; Zbl 0743.49018)]. The authors define the double layer potential \(K_A\) (\(A\) are the tensor coefficients in a representation of \(L\), a constant coefficient weakly elliptic \(M \times M\) system; details are in the book, but note weak ellipticity., which means the characteristic matrix is assumed invertible rather than that the characteristic form is bounded below). \textit{S. Hofmann} et al. [Int. Math. Res. Not. 2010, No. 14, 2567--2865 (2010; Zbl 1221.31010)] showed how to use Fredholm theory and boundedness on \(L^p(\Omega, \sigma)\) to prove that for every \(\epsilon > 0\), there is a \(\delta > 0\), such that if \[ \| \nu \|_{ [VMO(\partial \Omega,\sigma)]^n } < \delta, \ \operatorname{dist} (K_A, C_p(\partial \Omega, \sigma)) < \epsilon, \] where \(C_p\) are the compact operators and choosing \(\epsilon = 1/2\), \(\frac12 + K_A\) is a Fredholm operator of index zero, which was a key step in proving that \(K_A\) is invertible. This led to the development of a theory of boundary layer potentials in \(\delta\)-SKT domains, a subclass of the bounded uniformly rectifiable domains, inspired by the work of \textit{S. Semmes} [Adv. Math. 85, No. 2, 198--223 (1991; Zbl 0733.42015); ibid. 88, No. 2, 170--199 (1991; Zbl 0733.42016)] and \textit{C. E. Kenig} and \textit{T. Toro} [Duke Math. J. 87, No. 3, 509--551 (1997; Zbl 0878.31002); Ann. Math. (2) 150, No. 2, 369--454 (1999; Zbl 0946.31001); Ann. Sci. Éc. Norm. Supér. (4) 36, No. 3, 323--401 (2003; Zbl 1027.31005)]. It proceeds similarly but with \(\| \nu \|_{ [VMO(\partial \Omega,\sigma)]^n } < \delta\) replaced by \(\| \nu \|_{ [BMO(\partial \Omega,\sigma)]^n } < \delta\). This was used in [Hofmann et al., loc. cit.] to prove well-posedness of the Dirichlet, Regularity, Neumann and Transmission Problems for the Laplacian on \(\delta\)-regular SKT domains. The SKT domains are in the nature of best possible for close to compactness results; it is shown in [Hofmann et al., loc. cit.] that if \(\Omega\) is a uniformly rectifiable domain with a compact boundary and \(K_{\Delta}\) and certain commutators are compact, then \(\nu \in [VMO(\partial \Omega,\sigma)]^n\). Their main work is therefore on unbounded domains. Explicit solutions of the Laplacian \(\Delta\) in \(\mathbb{R}^n_+\) are known and \(K_{\Delta} = 0\) there. This leads to the idea that if a domain is nearly a half-space, \(\|K_{\Delta}\|\) will be small. One needs a way to quantify the idea that a domain is nearly a half space, and it is known that a uniformly rectifiable domain is a half-space if \(\|\nu \|_{ [BMO(\partial \Omega,\sigma)]^n } =0\). It should be nearly a half-space if the norm is sufficiently small. Given \(\delta > 0\), an open, nonempty, proper subset \(\Omega \subset \mathbb{R}^n\) is said to be a \(\delta\)-flat Ahlfors regular domain (or \(\delta\)-AR domain, for short) if \(\partial \Omega\) is an Ahlfors regular set, and if \(\sigma = \mathcal{H}^{n-1}(\partial \Omega)\), then the geometric measure theoretic outward unit normal \(\nu\) to \(\partial \Omega\) is well defined at \(\sigma\)-a.e. point on \(\partial \Omega\) and satisfies \[ \| \nu \|_{ [BMO(\partial \Omega,\sigma)]^n } < \delta. \] Throughout the work, the authors allow different tensor representations \(A\) of the elliptic system. For the next result, they require \(\mathcal{U}^{dis}_L \ne \emptyset\), which means that the tensor representing the operator has a distinguished coefficient tensor, which I will not define, but any strongly elliptic system has a distinguished coefficient tensor. They are able to answer Kenig's question for any constant coefficient second-order elliptic system with a distinguished coefficient tensor in \(\delta\)-AR domains \(\Omega \subseteq \mathbb{R}^n\) with \(\delta \in (0,1)\) sufficiently small in ordinary Lebesgue spaces, Lorentz spaces, Muckenhoupt weighted Lebesgue spaces, Morrey spaces, and Sobolev spaces, suitably defined. In the final chapter, they are able to extend the results to some weighted Banach function spaces by making a link with Muckenhoupt weighted Lebesgue spaces.