解P_0非线性互补问题的光滑Levenberg-Marquardt方法

A SMOOTHING LEVENBERG-MARQUARDT METHOD FOR P_0 NONLINERA COMPLEMENTARITY PROBLEM

本文研究了一个P0非线性互补问题.利用信赖域技术获得了求解该问题的光滑Levenberg-Marquardt算法,该算法在一定条件下具有全局性.利用局部误差界还获得了该算法的超线性和二次收敛.数值结果表明该算法是有效的.

In this paper, we study a P0 nonlinear complementarity problem. By using the techniques of trust-region, we acquire a smoothing Levenberg-Marquardt algorithm for nonlinear complementarity problem. Under suitable condition the gobal convergence properties of this algorithm are proved. With the local error condition, the local superlinear convergence of this algorithm is also obtained. This algorithm is efficient by the numerical experiments.