线性方程组异步并行迭代法的舍入误差分析

本文对求解大型线性方程组的异步并行迭代法进行了浮点运算的舍入误差分析,给出了算法是向前稳定的充分条件.     

线性方程组的异步松弛迭代法*

本文考虑解线性方程组经典迭代法的异步形式,对系数矩阵为H矩阵,给出了异步迭代过程收敛性的充分条件,这不仅降低了文献[3]对系数矩阵的要求,而且收敛区域比文献[3]的大.     

线性方程组二级迭代法的收敛性

线性方程组二级迭代法的收敛性曹志浩（复旦大学）ＣＯＮＶＥＲＧＥＮＣＥＯＦＴＷＯ－ＳＴＡＧＥＩＴＥＲＡＴＩＶＥＭＥＴＨＯＤＳＦＯＲＴＨＥＳＯＬＵＴＩＯＮＯＦＬＩＮＥＡＲＳＹＳＴＥＭＳ￥ＣａｏＺｈｉ－ｂａｏ（ＦｕｄａｎＵｎｉｖｅｒｓｉｔｙ）Ａｂｓｔｒａｃ...     

线性方程组迭代法的几何解释

以2阶矩阵为例,对线性方程组的迭代法进行了深入分析.以矩阵的特征值与谱半径作为分类原则,对迭代法的收敛或发散从几何上进行了解释,旨在加深学生从几何上理解线性方程组的迭代法.     

解线性方程组迭代法的若干几何研究

从解线性方程组迭代法入手,提出了两个迭代法的基本几何过程,揭示了著名的Jacobi迭代法、Gauss-Seidel迭代法和SOR方法等迭代法的几何实质、重新认识了这些经典的迭代过程,同时揭示了解线性方程组的克兰姆法则与迭代法的关系.同时从几何出发设计了一种解线性方程组的迭代方法.     

非齐次线性方程组的快速迭代法

本给出了一种解线性代数方程组的行正交化处理的新方法,它对任意初始向量x0,利用格式x1=x0-∑i=1^kui^T/uiui^TSi(其中ui=Ai-∑j=1^k-1AiujT/ujujTSj,si=Aix0-bi-∑j=1^i-1Aiuj^T/ujuj^TSj）可逐步逼近精确解。这种方法在计算机上操作,则更显示出它的优越性。     

MATLAB中大型线性方程组的非定常迭代法

科学研究和大型工程设计中很多问题以非线性数学模型来描述,而这些数学模型求解常常归结为各种大型线性方程组的求解,因而能否有效地求解大型线性方程组,特别是病态的方程组,是非常关键的.本文介绍了MATLAB中求解大型线性方程组常用的非定常迭代法,并以GMRES算法为例介绍了算法的数学描述.     

奇异线性方程组的并行多分裂迭代法

改进了奇异M-矩阵的线性方程组的并行多分裂法的一些最近结果,给出了并行多分裂迭代方法的一些收敛性的理论结果。     

解线性方程组的一种新方法

本文提出一种求解线性方程组的快速Ｊａｃｏｂｉ迭代方法，该方法在通常的串行计算机上比Ｇａｕｓｓ－Ｓｅｉｄｅｌ方法快，而且精度高，它对收敛慢的大型线性计算特别有效。     

GAOR迭代法的收敛性

当A为实对称矩阵时,[1]中在D_i选取较特殊的条件下,证明了GAOR迭代法收敛的充要条件为A是正定矩阵. 设A为Hermite矩阵,进一步讨论GAOR迭代法收敛的充要条件. 以下记 B=D_1~(-1)(C_L+C_U).     

Iterative Gradient Descent Methods for Solving Linear Equations

Computational Mathematics and Mathematical Physics - The paper presents the results on the use of gradient descent algorithms for constructing iterative methods for solving linear equations. A...     

Optimal Control of Iterative Solution Methods for Linear Systems of Equations

Iterative solution methods for linear systems of equations can be regarded as discrete-time control systems, for which a stabilizing feedback control has to be found. Well known algorithms such as GMRES(m) may exhibit unstable dynamics or sensitive dependence on initial conditions, thus preventing the algorithm to converge to the desired solution. Based on linear system feedback design techniques a new algorithm is proposed that does not suffer under such shortcomings. Global convergence to the desired solution is shown for any initial state. (© 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim)     

Superfast Iterative Solvers for Linear Matrix Equations

Superfast algorithms for solving large systems of linear equations are developed on the basis of an original method for multistep decomposition of a linear multidimensional dynamical system. Examples of analytical synthesis of iterative solvers for matrices of the general form and for large numerical systems of linear algebraic equations are given. For the analytical case, it is shown that convergence occurs at the second iteration.     

Homotopy Iterative Methods for Polynomial Equations

Globally convergent iterative methods for polynomial equationsf(z)= 0 are obtained. They are derived by modifying iterative functionssuch as Newton, Steffensen, etc., so as to avoid the criticalpoints and ensure the convergence. The modification is realized by the use of a one-parameter embeddingoperator (homotopy) associated with the polynomial.     

Iterative Methods for Boundary Integral Equations

We review some iterative methods for solving boundary integral equations which arise in Dirichlet and Neumann problems for the Helmholtz and Laplace equations. In particular we show how these integral equations may be transformed so that they may be solved by Neumann-Poincare Picard iteration.     

Convergence of Nonstationary Iterative Methods for Solving Singular Linear Equations with Index One

In this paper, we investigate the nonstationary iterative scheme for solving consistent singular linear system with index one. We utilize the group inverse to present a sufficient condition for the convergence of the nonstationary iterative method. Our result extends the known results of the stationary iterative scheme. Finally, we present a sufficient condition for the multisplitting algorithm and provide numerical examples to illustrate the advantages of nonstationary method.     

一类推广的拟牛顿多进程异步并行方法

本文通过近似雅可比矩阵Bk代替雅可比矩阵F′(xk),运用多进程异步并行方法求解非线性方程组。该方法在保持解的精度的情况下,缩短了运行时间和迭代步数。文中给出了算法收敛性的证明及八个非线性方程组的数值测试结果,表明该算法是可行的和快速的。     

非Hermite线性方程组的若干预处理迭代算法

非Hermite线性方程组在科学和工程计算中有着重要的理论研究意义和使用价值,因此如何高效求解该类线性方程组,一直是研究者所探索的方向.通过提出一种预处理方法,对非Hermite线性方程组和具有多个右端项的复线性方程组求解的若干迭代算法进行预处理,旨在提高原算法的收敛速度.最后通过数值试验表明,所提出的若干预处理迭代算法与原算法相比较,预处理算法迭代次数大大降低,且收敛速度明显优于原算法.除此之外,广义共轭A-正交残量平方法(GCORS2)的预处理算法与其他算法相比,具有良好的收敛性行为和较好的稳定性.