方程组系数矩阵的递归分块分解算法的设计实现

Implementation of Factorization Algorithm for Matrix

将方程组系数矩阵的LR分解算法过程加以改进,加入递归算法,使矩阵产生分块。用MATLAB和C语言对2种算法进行比较,发现递归分解法比LR分解算法速度快,占用内存少,适合当今分层多级存储的计算机结构。

This paper analyses the LR factorization of matrix, and improves the algorithm based on recursive algorithm and iteration, and brings forth the developed method of matrix factorization, which accelerates the convergence speed of the factorization. After comparing these two algorithms using MATLAB and C-program, it＇s found that the recursive algorithm is faster than the currently used algorithm, and it is fit to multiple level memories in the computer structure.