有限元线性方程组改进的全稀疏存贮方法

有限元线性方程组的系数矩阵一般具有稀疏性和对称性的特点,全稀疏存贮方法就是利用这些特点,只存贮对称部分的非零元素,采用链表式管理,即节省存贮空间,又便于动态更改;但在实际求解计算时,查找元素花费的时间较多;这里给出了一种改进的方法,是在链表的基础上生成相应的线性表,能够方便快速查询元素,使存贮结构同时具有链表和线性表的优点.数值算例证明该存贮方案在时间和存贮上都较为占优,可靠高效,能够应用于有限元线性方程组的求解.

Modified Fully Sparse Strategy of FEM Linear Equations

Coefficient matrix of linear equations from FEM is sparse and symmetrical.Using these characters,fully sparse strategy method stores only nonzero elements of symmetrical part with chain pattern management.Not only storage scale is small,but also storage structure is convenient for dynamic change.But in solution,too much time is expended in searching elements.A modified method is introduced for searching elements quickly.It establishes linear table on the basis of chain,yet has advantages of chain and linear table simultaneously.Numerical illustrations prove that the method is predominant for time and storage,and can improve efficiency of solution.Therefore modified fully sparse strategy can be applied to solve systems of linear equations from FEM.