求解绝对值线性互补问题的一种广义牛顿法

A Generalized Newton Method for Solving Absolute Complementarity Problems

本文研究了在绝对值互补问题的矩阵A-Dx正定的条件下,求解绝对值互补转化为求解凸二次函数极小值问题,并且利用该转化提出了一个求解绝对值互补问题的广义牛顿算法,证明了该算法的全局收敛性,并通过数值实验表明本文所提出的算法的有效性.

In this paper ,when the matrix A-Dx is positive definite in absolute complementarity problem , we transform solution of absolute value complementarity into solution of convex quadratic function minimi-zation problem .we propose a generalized Newton iterative algorithm for solving absolute value complemen-tarity problem.The global convergence of the algorithm was proved and numerical experiments show the ef -fectiveness of the proposed method .