A dwindling filter trust region algorithm for nonlinear optimization

作者：Gu, Chao^*; Zhu, Detong

来源：Applied Mathematics and Computation, 2014, 240: 72-81.

DOI：10.1016/j.amc.2014.04.101

• Nonlinear optimization
• Trust region
• Dwindling filter
• Feasibility restoration phase
• Convergence

In this paper, we develop dwindling filter idea and trust region strategy for solving nonlinear optimization. Compared with traditional filter trust region algorithm, the new filter algorithm has more flexibility for the acceptance of the trial step and the feasibility restoration phase is not needed. Under mild conditions, the global convergence of the new algorithm is analyzed. Preliminary numerical results are reported.

• 出版日期2014-8-1
• 单位上海师范大学; 上海立信会计金融学院