https://cnrs.hal.science/hal-03882984Maierhofer, GeorgGeorgMaierhoferLJLL (UMR_7598) - Laboratoire Jacques-Louis Lions - SU - Sorbonne Université - CNRS - Centre National de la Recherche Scientifique - UPCité - Université Paris CitéIserles, AAIserlesPeake, NNPeakeRecursive moment computation in Filon methods and application to high-frequency wave scattering in two dimensionsHAL CCSD2022highly oscillatory integralsnumerical integrationFilon quadraturewave scattering[MATH.MATH-NA] Mathematics [math]/Numerical Analysis [math.NA]Maierhofer, Georg2022-12-02 17:13:522023-03-24 14:53:292022-12-13 15:21:18enJournal articleshttps://cnrs.hal.science/hal-03882984/document10.1093/imanum/drac067application/pdf1We study the efficient approximation of highly oscillatory integrals using Filon methods. A crucial step in the implementation of these methods is the accurate and fast computation of the Filon quadrature moments. In this work we demonstrate how recurrences can be constructed for a wide class of oscillatory kernel functions, based on the observation that many physically relevant kernel functions are in the null space of a linear differential operator whose action on the Filon interpolation basis is represented by a banded (infinite) matrix. We discuss in further detail the application to two classes of particular interest, integrals with algebraic singularities and stationary points and integrals involving a Hankel function. We provide rigorous stability results for the moment computation for the first of these classes and demonstrate how the corresponding Filon method results in an accurate approximation at truly frequency-independent cost. For the Hankel kernel, we derive error estimates which describe the convergence behaviour of the method in terms of frequency and number of Filon quadrature points. Finally, we show how Filon methods with recursive moment computation can be applied to compute efficiently integrals arising in hybrid numerical-asymptotic collocation methods for high-frequency wave scattering on a screen.