Article
Keywords:
accelerated overrelaxation method; AOR method; successive overrelaxation; rate of convergence; relaxation parameter; interval of convergence; iterative process
Summary:
The convergence of the Accelerated Overrelaxation (AOR) method is discussed. It is shown that the intervals of convergence for the parameters $\sigma$ and $\omega$ are not always of the following form: $0\leq \omega \leq \omega_1, -\sigma_1\leq\sigma\leq\sigma_2, \sigma_1, \sigma_2\geq 0$.
References:
[1] G. Avdelas A. Hadjidimos:
Some theoretical and computational results concerning the accelerated overrelaxation (AOR) method. Anal. Numer. Theor. Approx. 9 (1980), 5-10.
MR 0617249
[2] Lj. Cvetkovič D. Herceg:
Some sufficient conditions for convergence AOR-method. In: Numerical Methods and Approximation Theory, G. V. Milovanič, ed., Faculty of Electronic Engineering, Niš, 1984, 143-148.
MR 0805793
[3] Lj. Cvetkovič D. Herceg: Convergence theory for AOR method. Journal of Computational Mathematics (in print).
[4] Lj. Cvetkovič D. Herceg:
An improvement for the area of convergence of the AOR method. Anal. Numer. Theor. Approx. 16 (1987), 109-115.
MR 0986095
[7] M. Martins:
An improvement for the area of convergence of the accelerated overrelaxation iterative method. Anal. Numer. Theor. Approx. 12 (1983), 65 - 76.
MR 0743917 |
Zbl 0527.65023