In this paper, the modifications of the Hermitian-Normal splitting iteration methods for solving a class of complex symmetric linear systems are presented. Theoretical analysis shows that the modified iteration methods of Hermitian-normal splitting are unconditionally convergent; the coefficient matrices of the two linear systems solved in each iteration of the methods are real symmetric positive definite. Inexact version of the methods employs the Krylov subspace method as an internal iteration to accelerate. Numerical examples from two model problems are given to illustrate the effectiveness of the modified iteration methods.

中文翻译：

---

一类复杂对称线性系统的修正Hermitian-Normal分裂迭代方法

本文提出了求解一类复杂对称线性系统的Hermitian-Normal分裂迭代方法的改进方法。理论分析表明，改进的Hermitian-Normal分裂迭代方法是无条件收敛的。在方法的每次迭代中求解的两个线性系统的系数矩阵是实对称正定的。方法的不精确版本采用Krylov子空间方法作为内部迭代来加速。给出了两个模型问题的数值例子，以说明改进的迭代方法的有效性。