The approximation of the Maxwell eigenvalue problem using a least-squares method

From MaRDI portal





The authors analyze a new approach for computing the eigenvalues of Maxwell's equation in a three-dimensional domain. First, the problem is related to a block system involving the solution of two very weakly formulated div-curl systems. However, the existence of a solution of the discrete problems requires compatability conditions which are difficult to deal with. Instead, an approach which is related to least squares in a dual norm is used. Analytical results include the approximation property of the discrete operators, estimates of the rate of convergence of the eigenvalues and eigenvectors, and an improved rate of convergence in case of smooth data. The method can be implemented in such a way that the discrete problems become symmetric and hence fast solvers for large symmetric eigenvalue problems can be applied. Numerical studies are provided, ranging from academic test examples for supporting the analysis to a real-world problem.



Cites work


Cited in
(34)


Describes a project that uses

Uses Software






This page was built for publication: The approximation of the Maxwell eigenvalue problem using a least-squares method

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q5315408)