Analysis of a Krylov subspace enhanced parareal algorithm for linear problems

Abstract

The parareal algorithm is a numerical method to integrate evolution problems on parallel computers. The performance of the algorithm is well understood for diffusive problems, and it can have spectacular performance when applied to certain non-linear problems. Its convergence properties are however less favorable for hyperbolic problems. We present and analyze in this paper a variant of the parareal algorithm, recently proposed in the PITA framework for systems of second order ordinary differential equations.