Volume 51, Number 1, January-February 2017
|Page(s)||209 - 223|
|Published online||28 November 2016|
Optimized waveform relaxation methods for RC circuits: discrete case
1 Sichuan University of Science and Engineering, Zigong,
Sichuan 643000, P.R. China.
2 Department of Mathematics, Yarmouk University, 21163 Irbid, Jordan.
3 Department of Mathematics and Sciences, Khalifa University, 127788 Abu Dhabi, UAE.
Revised: 28 August 2016
Accepted: 8 September 2016
The optimized waveform relaxation (OWR) methods, benefiting from intelligent information exchange between subsystems – the so-called transmission conditions (TCs), are recognized as efficient solvers for large scale circuits and get a lot of attention in recent years. The TCs contain a free parameter, namely α, which has a significant influence on the convergence rates. So far, the analysis of finding the best parameter is merely performed at the continuous level and such an analysis does not take into account the influence of temporal discretizations. In this paper, we show that the temporal discretizations do have an important effect on the OWR methods. Precisely, for the Backward–Euler method, compared to the parameter αcopt from the continuous analysis, we show that the convergence rates can be further improved by using the one αdopt analyzed at the discrete level, while for the Trapezoidal rule, it is better to use αcopt. This conclusion is confirmed by numerical results.
Mathematics Subject Classification: 65L12 / 65L20 / 65B99
Key words: Waveform relaxation (WR) / discretization / parameter optimization / RC circuits
© EDP Sciences, SMAI 2016
Current usage metrics show cumulative count of Article Views (full-text article views including HTML views, PDF and ePub downloads, according to the available data) and Abstracts Views on Vision4Press platform.
Data correspond to usage on the plateform after 2015. The current usage metrics is available 48-96 hours after online publication and is updated daily on week days.
Initial download of the metrics may take a while.