arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2005.07499math.NAcs.NAmath.FA
收藏

基于双分裂的交替定常迭代方法

Alternating Stationary Iterative Methods Based on Double Splittings

Ashish Kumar Nandi, Nachiketa Mishra, Debasisha Mishra

AI总结:

针对实非奇异(矩形)线性系统求解,提出两种交替双分裂(ADS)迭代格式,分析其收敛性与矩阵适用类别的互斥性,数值实验验证其性能优于现有方法。

摘要:

矩阵双分裂迭代在求解实非奇异(矩形)线性系统时实现简单。本文提出两种交替双分裂(Alternating Double Splitting,ADS)格式,其由两种双分裂构造并交替执行各自的迭代。文中讨论了收敛条件并进行了对比分析。每种ADS格式所用的双分裂集合会诱导出一个预条件系统,该系统有助于证明ADS格式的收敛性。我们还证明,一种ADS格式优于另一种ADS格式的矩阵类别是互斥的。数值实验证实,所提出的ADS格式在实际实现中优于现有方法。尽管这些问题是在矩形矩阵设定下研究的,但同类问题在非奇异矩阵设定下也属于新问题。

英文摘要:

Matrix double splitting iterations are simple in implementation while solving real non-singular (rectangular) linear systems. In this paper, we present two Alternating Double Splitting (ADS) schemes formulated by two double splittings and then alternating the respective iterations. The convergence conditions are then discussed along with comparative analysis. The set of double splittings used in each ADS schemes induce a preconditioned system which helps in showing the convergence of the ADS schemes. We also show that the classes of matrices for which one ADS scheme is better than the other are mutually exclusive. Numerical experiments confirm the proposed ADS schemes are superior to the existing methods in actual implementation. Though the problems are considered in the rectangular matrix settings, the same problems are even new in non-singular matrix settings.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑