含非对称正定系数矩阵方程的解法研究

Algorithm for Equations Including Non-symmetric Definite Matrix

在求解有限元方程中有时会遇到系数矩阵是非对称正定的，如在变分法中引入拉格郎日乘数时，即出现这种情况。此类方程有时会出现病态，其求解具有一定特点，本文提出了一种求解方法。

This paper describes a procedure for solving equations which include non-symmetric definite matrix.The ill-condition in this matrix occurs in finite elment method such as using a least squares functional minimization approach.The matrix has it's own characters.The efficiency of the process is demonstrated by a test case.