求解非对称线性方程组的松弛混合算法

Relaxation Hybrid Algorithms for Nonsymmetric Linear Systems

求解大型稀疏非对称线性方程组的混合迭代算法通常会由于系数矩阵的谱分布较广而导致收敛失败。本文通过在迭代多项式中加入变化的松弛因子定义了一类松弛混合算法。

The hybrid iterative algorithms for solving large sparse nonsymmetric systems of linear equations generally result in fail on convergence if the coefficient matrix has a wide spectrum distribution.A class of relaxation hybrid algorithm is defined to add varying relaxation factors into the iteration polynomials.With proper choice of the relaxation factors the new algorithms improve the convergence.