Parareal — RBF algorithms for solving time-dependent PDEs

Nadun Kulasekera Mudiyanselage, Cecile Piret and Benjamin Ong (Michigan Technological University)

Radial basis functions (RBF) are a mesh-less approach to discretize differential operators in space. Over the past two decades, the RBF method has gained attention from numerical analysts for its ability to achieve spectral/high-order accuracy. When solving time-dependent PDEs with RBFs, choosing a time integrator that couples well with the RBF discretizations has been an important research topic within the RBF community. In this talk, we explore how the parareal framework can be used to provide a time-parallel approach to solving time-dependent PDEs discretized spatially using RBFs, focusing on how the RBF discretizations can be modified (enriched) to generate desirable coarse parareal solvers.