您的位置:山东大学 -> 科技期刊社 -> 《山东大学学报(理学版)》

J4 ›› 2012, Vol. 47 ›› Issue (10): 81-88.

• 论文 • 上一篇    下一篇

单调线性互补问题基于核函数的满-Newton步不可行内点算法

陈月姣,张明望*   

  1. 三峡大学理学院, 湖北 宜昌 443002
  • 收稿日期:2011-08-21 出版日期:2012-11-20 发布日期:2012-10-26
  • 通讯作者: 张明望(1959- ),男,教授,主要研究方向为最优化理论及应用. Email:zmwang@ctgu.edu.cn
  • 作者简介:陈月姣(1987- ),女,硕士,主要研究方向为最优化理论及应用. Email:mlchenyuejiao@163.com
  • 基金资助:

    湖北省自然科学基金资助项目(2008CDZ047)

A full-newton step infeasible interior-point algorithm for monotone linear  complementarity problems based on a kernel function

CHEN Yue-jiao, ZHANG Ming-wang*   

  1. College of Science, China Three Gorges University, Yichang 443002, Hubei, China
  • Received:2011-08-21 Online:2012-11-20 Published:2012-10-26

摘要:

 针对单调线性互补问题设计了一种基于核函数的满Newton步不可行内点算法,算法的主迭代由一个可行步和几个中心步构成。通过建立和应用一些新的分析工具,证明了算法的多项式复杂性为Onlogmax{(x0)Ts0,‖r0‖}n,这与当前单调线性互补问题的不可行内点算法最好的迭代界一致。

关键词: 单调线性互补问题;不可行内点算法;满-Newton步;核函数;多项式复杂性

Abstract:

 A full-newton step infeasible interior-point algorithm for monotone linear complementarity problems based on a kernel function was presented. The main iteration of the algorithm consists of a feasibility step and several centrality steps. By using some new analysis tools, the polynomial complexity bound for the algorithm was obtained, namely, Onlogmax{(x0)Ts0,‖r0‖}n, which is consistent with the currently best iteration bound for infeasible interior-point methods of monotone linear complementarity.

Key words: linear complementarity problem; infeasible interior point algorithm; full-newton step; kernel function; polynomial complexity

No related articles found!
Viewed
Full text


Abstract

Cited

  Shared   
  Discussed   
No Suggested Reading articles found!