摘要
本文首先将半定规划转化为一个变分不等式问题,在满足单调性和Lipschitz连续的条件下,提出了一种基于Korpelevich-Khobotv算法的新的预测-校正算法,并给出算法的收敛性分析.
In this paper,we first transform a semidefinite programming problem into a variational inequality problem,and then propose a novel prediction-correction algorithm based on Korpelevich-Khobotv method under monotone and Lipschitz continuous conditions.The convergence of the algorithm is also given and the optimal step-size is employed in the correction step to accelerate convergence rate.The numerical experiment indicates that this method is effective.
出处
《应用数学》
CSCD
北大核心
2005年第S1期5-9,共5页
Mathematica Applicata
基金
教育部跨世纪优秀人才基金项目
陕西省自然科学研究项目(2002A13)
关键词
半定规划
预测-校正算法
变分不等式
Semidefinite programming
Prediction-correction algorithm
Variational inequality
