一种新型压缩预处理CGS算法

CGS算法是求解大型非对称线性方程组的常用算法，然而该算法无极小残差性质，因此它常因出现较大的中间剩余向量而出现典型的不规则收敛行为．本根据IRA方法提出了一种压缩预处理CGS方法，数值实验表明这种算法在一定程度上减小了迭代算法在收敛过程中的剩余问题，从而使得算法具有更好的稳定性，该法构造简单，减少了收敛次数，加快了收敛速度．

A new deflated preconditioning CGS algorithm

CGS algorithm is popular for solving large scale nonsymmetric linear equation.However,it has no minimal residual property and typically exhibit a rather irregulate convergence behavior with wild oscillations in the residual norm.Here a new preconditioned deflated CGS algorithm based on IRA method is presented.Numerical experiments show that the algorithm can work well to lower the norms of the residuals,which make the algorithm performs more stable.And a faster convergent rate of the iterative method is schieved.