科研之友
微信
新浪微博
Facebook
分享链接
摘要
In this paper, a new filter trust region algorithm is proposed for solving unconstrained nonlinear optimization problems. It modifies the retrospective ratio to a convex combined ratio for updating the trust-region radius. Moreover we use the filter technique to relax the acceptance of the trial point. The new algorithm is shown to be globally convergent to a first-order critical point. Numerical experiments on CUTEr problems indicate that the new algorithm is competitive.
- 出版日期2016-10
- 单位南京晓庄学院; 南京师范大学
