An inexact interior point method for monotone NCP

被引:4
|
作者
Bellavia, S [1 ]
Macconi, M [1 ]
机构
[1] Univ Florence, Dipartimento Energet S Stecco, I-50134 Florence, Italy
来源
OPTIMIZATION METHODS & SOFTWARE | 1999年 / 11-2卷 / 1-4期
关键词
inexact interior point; nonlinear complementarity problems; polynomial complexity; rate of convergence;
D O I
10.1080/10556789908805752
中图分类号
TP31 [计算机软件];
学科分类号
081202 ; 0835 ;
摘要
In this paper we present an inexact Interior Point method for solving monotone nonlinear complementarity problems. We show that the theory presented by Kojima, Noma and Yoshise for an exact version of this method can be used to establish global convergence for the inexact form. Then we prove that local superlinear convergence can be achieved under some stronger hypotheses. The complexity of the algorithm is also studied under the assumption that the problem satisfies a scaled Lipschitz condition. It is proved that the feasible version of the algorithm is polynomial, while the infeasible one is globally convergent at a linear rate.
引用
收藏
页码:211 / 241
页数:31
相关论文
共 50 条