Krylov Subspace Acceleration of Waveform Relaxation

From MaRDI portal





In recent years, certain very old ideas of iterative algorithms traditionally used to linear algebraic problems, come back with the changed name of waveform relaxation when applied in a changed context, for example to problems of ordinary differential equations (ODEs) or partial differential equations (PDEs). In fact these methods were applied in the field of ODEs or PDEs, also for eigenproblems for ODEs, many years ago with some success (see years 60 of the 20-th century). But then, people was not interested in it. This return of good old ideas is probably due to some possibility to apply them sometimes for nonlinear problems and to possibilities of parallelisation. The authors define and analyse algorithms of this kind: Jacobi iteration, Gauss-Seidel iteration, general Krylov methods, conjugate gradient methods, generalized minimal residual (GMRES) methods, and so on. The most interesting are methods of acceleration proposed here: this is a fresh look at this subject: convergence acceleration. I recommend this paper to people interested in numerical analysis.











This page was built for publication: Krylov Subspace Acceleration of Waveform Relaxation

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4443580)