Elliptic boundary value problems on corner domains. Smoothness and asymptotics of solutions (Q1188499)

The book under review is a rather complete (and thorough) treatment of elliptic boundary value problems in domains with nonsmooth boundaries. It is divided into four parts. Part 1 contains a review of previous results and a comparison to some of the author's original work which is incorporated in the book. The highlight of this part is a very careful definition of the author's ``corner domains'' (by recursion on space dimension) which is both precise and general enough so as to include domains with ``holes'' (single points removed), ``slits'', ``cracks'', ``edges'' and ``conical points''. Part 2 is devoted to Dirichlet's problem both for elliptic equations and strongly elliptic systems. Thanks to the author's careful definition of corner domains the ``splitting'' of the resulting operators L by introducing ``spherical coordinates'' can be described with unusual ease and clarity. Furthermore the author's idea of replacing ``injectivity'' by ``injectivity modulo polynomials'' for the ``split operators'' is introduced and yields improvements over the classical results of V. A. Kondrat'ev. Namely in great generality the properties of ``L is (semi-) Fredholm'', ``global regularity holds for L'',... are tied to ``injectivity modulo polynomials'' properties for the ``split operators''. Also ``subtracting off singularities'' can be described in great generality. In the second half of Part 2 the general machinery is applied to special situations namely i) domain in \({\mathbb{R}}^ 2\); ii) edges modelled over smooth cones; iii) edges in \({\mathbb{R}}^ 3\) meet in a conical point (polyhedra); iv) the Laplacian in various geometries. Part 3 is devoted to non-Dirichlet-problems in variational setting. It starts with a regularity theory in fractional order Sobolev spaces for smooth operators in smooth domains based on a-priori estimates. This theory is used in treating non-Dirichlet-operators in ``polyhedral domains''. (This means that by modelling the corner singularities by recursion on the dimension one ends up with cones with ``plane faces''.) Revising the definition of ``injectivity modulo polynomials'' for the split operators results similar to those of Part 2 are obtained. Part 4 is in the form of an appendix devoted to four topics: A) Fractional order (weighted) Sobolev spaces: Comparison of the various possible definitions of this space (in particular by using the Mellin transform); B) An abstract lemma connecting the indices of two Fredholm operators differing by ``subtracted singularities''. C) An analytic perturbation result for the range of a semi-Fredholm operator. D) An algebraic lemma on polynomials vanishing on the boundary of a cone.