Comparison results for parallel multisplitting methods with applications to AOR methods

From MaRDI portal





For the iterative solution of large linear systems, several splittings of the same matrix are used in parallel to perform one step ahead, and the obtained results are added in a weighted way to define the next iterate. The authors compare the convergence speeds of such procedures, generalizing former results (and classical M-matrix splitting theory) in dependence on properties of the splittings and of the weighting matrices. Typical results are of Stein-Rosenberg type: depending on iteration parameters, one of two compared multisplitting methods is either faster converging or faster diverging than the other.



Cites work









This page was built for publication: Comparison results for parallel multisplitting methods with applications to AOR methods

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q5943031)