利用最优尺度矩阵及M-1N的某些估计量讨论了外推Gauss-Seidel迭代法的收敛性及其和H-矩阵的关系。基于外推Gauss-Seidel及Gauss-Seidel迭代法得到了H-矩阵的几个等价条件。同时也得到了严格对角占优矩阵,不可约对角占优矩阵及Stieltjes矩阵的Gauss-Seidel迭代法,外推Gauss-Seidel迭代法的相关收敛性结论。
The optimal scaling matrix and the estimates of some amounts of M^-1 N are used to discuss the convergence of the extrapolated Gauss-Seidel iteration methods and the relationship between them and H-matrices. Several equivalent conditions for the H-matrices based on the extrapolated Gauss-Seidel and the Gauss-Seidel iterative methods are obtained. And the convergence results of the Gauss-Seidel and the extrapolated Gauss-Seidel iterative methods are given for the strictly diagonally dominant matrices, the irreducible diagonally dominant matrices and the Stieltjes matrices.