Adaptive-order rational Arnoldi-type methods in computational electromagnetism

In this paper, we will consider moment-matching methods for the application of model order reduction to Maxwell’s equations. Since the main difficulty of moment-matching methods results from the adequate determination of a set of expansion points, we will introduce an adaptive expansion point selection on the basis of a suitable approximation of an upper bound of the output moment-matching error. Furthermore, we give some remarks about structure- and passivity preserving adaptive-order rational Arnoldi methods for Maxwell’s equations. Numerical examples indicate the reliability of the proposed algorithm.