In Wang et al. (J. Optim. Theory Appl., \textbf{181}: 216--230, 2019), a class of effective modified Newton-tpye (MN) iteration methods are proposed for solving the generalized absolute value equations (GAVE) and it has been found that the MN iteration method involves the classical Picard iteration method as a special case. In the present paper, it will be claimed that a Douglas-Rachford splitting method for AVE is also a special case of the MN method. In addition, a class of inexact MN (IMN) iteration methods are developed to solve GAVE. Linear convergence of the IMN method is established and some specific sufficient conditions are presented for symmetric positive definite coefficient matrix. Numerical results are given to demonstrate the efficiency of the IMN iteration method.
翻译:在Wang等人(J.Optim.Theory Appl.,\textbf{181}:216-230,2019)一案中,为解决普遍绝对值方程(GAVE),提出了一类有效的牛顿-tpye(MN)迭代法,发现MN迭代法涉及典型的皮卡迭代法,这是一个特殊案例。在本文件中,将声称道格拉斯-拉克福德对AVE的分解法也是MN方法的一个特殊案例。此外,还开发了一类不完全的MN(IMN)迭代法,以解决GAVE问题。建立了IMN方法的线性趋同,并为正数固定系数矩阵提出了一些具体的条件。提供了数字结果,以证明IMN的迭代法效率。