互补约束优化问题的一个非单调信赖域法  

A Nonmonotone Trust Region Algorithm for Mathematical Programs with Complementarity Constraints

作　　者：段翀[1] 陈国庆[2] 贾秀敏[1] 

机构地区：[1]内蒙古科技大学数理与生物工程学院,内蒙古包头014010 [2]内蒙古大学数学科学学院,呼和浩特010021

出　　处：《内蒙古大学学报（自然科学版）》2010年第5期490-495,共6页Journal of Inner Mongolia University：Natural Science Edition

基　　金：内蒙古科技大学校内科研基金资助项目

摘　　要：针对互补问题构造了一种新的价值函数,从而把互补约束优化问题等价地转化为一般光滑约束优化问题.然后,结合非单调技术给出了一个信赖域算法,在一定条件下证明了算法的全局收敛性.A new merit function is constructed for mathematical programs with complementarity constraints. Thus the mathematical programs with complementarity constraints are equivalently transformed into general smoothing constrained optimization problems. Then, a trust region algorithm with a nonmonotone techniqe is proposed,and under some conditions, the global convergence of the algorithm is proved.

关 键 词：互补约束优化问题 乘子 信赖域 全局收敛 

分 类 号：O221.2[理学—运筹学与控制论]