摘要
本文建立带退化线性等式与不等式约束最优化向题的一种信赖域算法,方法用一系列以原点为内点的一般紧集为信赖域.讨论了方法的收敛性,证明了迭代点列必有一聚点为原问题的 Kuhn-Tucker 点.最后,在一定的假设下,讨论了算法的超线性收敛性.
This paper presents a trust region algorithm for optimizational problems with degenerate linear equality and inequality constraints.It uses a sequence general compact sets whose interior sets contain the origion O as the trust regions.We discusses the convergence of the method,and proves that the iteration sequential points must have som limit point which is a Kuhn-Tucker point of the original problem.At last,the superlinear convergence is discussed under suitable assumptions.
基金
广西自治区青年科学基金
教委科学基金资助项目
关键词
退化约束
线性约束
依赖域法
最优化
收敛性
linear constraints,trust region method,global and superlinear convergence, degenerate constraint.
