摘要
对框式凸二次规划问题提出了一种非精确不可行内点算法 ,该算法使用的迭代方向仅需要达到一个相对的精度 .在初始点位于中心线的某邻域内的假设下 。
An inexact-infeasible-interior algorithm for convex quadratic programming problem with box constraints is developed.The algorithm uses the inexact search directons that are computed with only moderate accuracy.Under assumption that initial point is in neighborhood of the centeral path,the global convergence of the algorithm is proved.
出处
《应用数学》
CSCD
北大核心
2004年第2期315-321,共7页
Mathematica Applicata
基金
湖北省教育厅重点科研项目资助 (2 0 0 2 0 5 30 12 )
关键词
框式凸二次规划
非精确不可行内点
全局收敛性
对偶规划
半正定矩阵
Convex quadratic programming problem with box constraints
Infeasible-interior-point algorithm
Inexact search direction
Global convergence
