The optimal \(\omega\) is not best for the SOR iteration method (Q808162)

The author proposes a new method for improving the convergence of the successive overrelaxation (SOR) method by using a polynomial acceleration together with a suboptimal relaxation factor \((\omega <\omega_ b)\) as an alternative to the unaccelerated SOR iteration with optimal relaxation factor \(\omega_ b\). Practical polynomial accelerations and numerical results of the run using the better polynomial are presented.