解信赖域子问题的隐式分段折线算法

王希云; 李亮; 于海波

doi:10.3879/j.issn.1000-0887.2014.06.003

解信赖域子问题的隐式分段折线算法

doi: 10.3879/j.issn.1000-0887.2014.06.003

太原科技大学应用科学学院, 太原 030024

基金项目: 山西省自然科学基金(2008011013)；山西省“131”领军人才工程项目

详细信息

作者简介:
王希云(1964—)，女，山西临汾人，教授(E-mail: tykdwxy@126.com)

中图分类号: O221
计量
- 文章访问数: 1136
- HTML全文浏览量: 106
- PDF下载量: 668
- 被引次数: 0
出版历程
- 收稿日期: 2013-11-21
- 修回日期: 2014-05-07
- 刊出日期: 2014-06-11

An Implicit Piecewise Dogleg Algorithm for Solving Trust-Region Subproblems

School of Applied Science,Taiyuan University of Science and Technology, Taiyuan 030024, P.R.China

摘要

摘要: 在Hessian矩阵正定的前提下,建立了一种最优曲线的微分方程模型.针对此微分方程模型，构造了一条隐式分段折线，从而提出了一种求解信赖域子问题的隐式分段折线算法，并且分析和证明了隐式分段折线路径的合理性。数值结果表明新算法是有效且可行的.
- 隐式分段折线算法 /
- 微分方程模型 /
- 信赖域子问题
Abstract: Based on the premise that Hessian matrix was positive definite, a differential equation model was established for the optimal curve. Then an implicit piecewise dogleg was constructed according to the differential equation. In turn, the implicit piecewise dogleg algorithm for solving trust-region subproblems was presented. And the rationality of the implicit piecewise dogleg path was analyzed and demonstrated. Numerical results indicate that the new algorithm is effective and practicable.
- implicit piecewise dogleg algorithm /
- differential equation model /
- trust-region subproblem

HTML全文

参考文献(13)

[1]	徐成贤, 陈志平, 李乃成. 近代优化方法[M]. 北京: 科学出版社, 2002.(XU Cheng-xian, CHEN Zhi-ping, LI Nai-cheng.Modern Optimization Method [M]. Beijing: Science Press, 2002.(in Chinese))
[2]	后六生, 孙文瑜. 三项预处理共轭梯度法与信赖域子问题[J]. 南京师大学报(自然科学版), 2001,24(3): 1-6.(HOU Liu-sheng, SUN Wen-yu. Three-term preconditioned conjugate gradient method and trust region subproblem[J].Journal of Nanjing Normal University (Natural Science),2001,24(3): 1-6.(in Chinese))
[3]	赵英良, 徐成贤. 信赖域子问题使用重新开始策略的共轭梯度法[J]. 高校应用数学学报(A辑), 2003,18(3): 341-349.(ZHAO Ying-liang, XU Cheng-xian. Conjugate gradient method using restart strategy for solving trust region subproblems[J].Applied Mathematics: A Journal of Chinese Universities(Ser A),2003,18(3): 341-349.(in Chinese))
[4]	陈争, 马昌风. 求解信赖域子问题的一个光滑牛顿法[J]. 福建师范大学学报(自然科学版), 2011,27(4): 31-35.(CHEN Zheng, MA Chang-feng. A smoothing Newton method for trust region subproblem[J].Journal of Fujian Normal University(Natural Science Edition),2011,27(4): 31-35.(in Chinese))
[5]	张立, 唐志强. 解信赖域子问题的混合折线法[J]. 南京师大学报(自然科学版), 2001,24(1): 28-32.(ZHANG Li, TANG Zhi-qiang. The hybrid dogleg method to solve subproblems of trust region[J].Journal of Nanjing Normal University (Natural Science),2001,24(1): 28-32.(in Chinese))
[6]	PU Ding-guo, HAN Bo-shun, YAO Lin, ZHENG Guang-hua. A class of dogleg trust region methods[J].Operations Research Transactions,2003,7(1): 1-10.
[7]	CHEN Jun, SUN Wen-yu. Nonmonotone adaptive trust region algorithms with indefinite dogleg path for unconstrained minimization[J].Northeast Math Journal,2008,24(1): 19-30.
[8]	Powell M J D. A hybrid method for nonlinear equations[C]//Rabonowitz P.Numerical Methods for Nonlinear Algebraic Equations . London: Gordon and Breach, 1970: 87-114.
[9]	Dennis J E, Mei H H W. Two new unconstrained optimization algorithms which use function and gradient values[J].Journal of Optimization Theory and Applications,1979,28(4): 453-482.
[10]	赵英良, 徐成贤. 解信赖域子问题的切线单折线法[J]. 数值计算与计算机应用, 2000,21(1): 77-80.(ZHAO Ying-liang, XU Cheng-xian. Tangent single dogleg method for trust region subproblem[J].Journal of Numerical Methods and Computer Applications,2000,21(1): 77-80.(in Chinese))
[11]	颜庆津. 数值分析[M]. 北京: 北京航空航天大学出版社, 2006.(YAN Qing-jin.Numerical Analysis [M]. Beijing: Beihang University Press, 2006.(in Chinese))
[12]	袁亚湘, 孙文瑜. 最优化理论与方法[M]. 北京: 科学出版社, 2007.(YUAN Ya-xiang, SUN Wen-yu.Optimization Theory and Methods [M]. Beijing: Science Press, 2007.(in Chinese))
[13]	李董辉, 童小娇, 万中. 数值最优化算法与理论[M]. 北京: 科学出版社, 2010.(LI Dong-hui, TONG Xiao-jiao, WAN Zhong.Numerical Optimization Algorithms and Theory [M]. Beijing: Science Press, 2010.(in Chinese))