This article is concerned with adaptive numerical frame methods for elliptic operator equations. We show how specific noncanonical frame expansions on domains can be constructed. Moreover, we study the approximation order of best n-term frame approximation, which serves as the benchmark for the performance of adaptive schemes. We also discuss numerical experiments for second order elliptic boundary value problems in polygonal domains where the discretization is based on recent constructions of boundary adapted wavelet bases on the interval.
Abstract. A multiplicative Schwarz overlapping domain decomposition method is considered for solving elliptic boundary value problems. By equipping the relevant Sobolev spaces on the subdomains with wavelet bases, adaptive wavelet methods are used for approximately solving the subdomain problems. The union of the wavelet bases forms a frame for the Sobolev space on the domain as a whole. The resulting method is proven to be optimal in the sense that, in linear complexity, the iterands converge with the same rate as the sequence over N ∈ N of the best approximation from the span of the best N frame elements. Numerical results are given for the method applied to Poisson's equation.
Adaptive wavelet algorithms for solving operator equations have been shown to converge with the best possible rates in linear complexity. For the latter statement all costs are taken into account, i.e., also the cost of approximating entries from the infinite stiffness matrix with respect to the wavelet basis using suitable quadrature. A difficulty is the construction of a suitable wavelet basis on the generally non-trivially shaped domain on which the equation is posed. In view of this, recently corresponding algorithms have been proposed that require only a wavelet frame instead of a basis. By employing an overlapping decomposition of the domain, where each subdomain is the smooth parametric image of the unit cube, and by lifting a wavelet basis on this cube to each of the subdomains, the union of these collections defines such a frame. A potential bottleneck within this approach is the efficient approximation of entries corresponding to pairs of wavelets from different collections. Indeed, such wavelets are piecewise smooth with respect to mutually non-nested partitions. In this paper, considering partial differential operators and spline wavelets on the subdomains, we propose an easy implementable quadrature scheme to approximate the required entries, which allows the fully discrete adaptive frame algorithm to converge with the optimal rate in linear complexity.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.