• 论文 • 上一篇    

基于非二次模型的无约束极小化方法

孙麟平   

  1. 南京大学
  • 出版日期:1985-03-20 发布日期:1985-03-20

孙麟平. 基于非二次模型的无约束极小化方法[J]. 数值计算与计算机应用, 1985, 6(3): 188-193.

AN UNCONSTRAINED MINIMIZATION METHOD BASED ON NONQUADRATIC MODEL

  1. Sun Lin-ping(Nanjing University)
  • Online:1985-03-20 Published:1985-03-20
多年来,关于无约束极小化问题已经提出了许多行之有效的算法。它们大多是基于二次目标函数而建立的数值方法。这些算法所固有的良好属性也只对二次目标函数才成立。在实际计算中,如果目标函数在x~((k))的一个充分小的领域内不能很好地用一个二次函数逼次时,计算效果就会受到严重的影响,甚至导致算法的失败。为了对那些极小点附近呈现出较强非二次性的目标函数,构造出相应的有效算法,D.H.Jacobson和W.Oksman于1972年提出了以齐次函数为目标函数建立极小化算法的思想。 后来J.S.
Considered is on application of the generalized inverse matrix to the homogeneous algorithm for unconstrained optimization problems. The optimization algorithm is based on a nonquadratic model function. Some numerical tests illustrate the performance of the proposed version of the method.
()

[1] J. Dennis, J. More, Quasi-Newton methods, motivation and theory, SIAM Rev. 19 (1977) p 46-87.
[2] D. Jacobson, W. Oxman, An algonthm that minimizes homogenems functions of N variables in N+2 iterations and rapidly minimizes general functions, J. Math. Anal. Appl, 38 (1972) P. 533-552.
[3] J. Kowalik, K. Ramakrishnan, A namerically stable optimization method based on a homogenous function, Math. Prog 11 (1976) P 50-66.
[4] J. S. Wowalik, S. P. Kumar, Fast Givens Transformations Applied to the Homogeneous Optimization Method, Appl. Math. Comput, 4 (1978) , P. 239-252.
[5] J. H. Wilkinson (1977) , Some Recent Advences in Numerical Linear Algebra, The State of the Art in Numerical Analysis, Edited by D. Jacobs, Academie Press, London-New York-San Fransisco PP. 3-53.
[6] Frieder Kuhnert广义逆矩阵和正则化方法.
[7] 1974, № 2, PP. 483-486.
[8] 孙麟平,无约束极小化的自适应多信息下降法,高等学校计算数学学报2,(1982) .
[9] M. J. D. Powell, Restart Procedures for the Conjugate Gradient Method, Math. Prog. Vol. 12 1977 P. 241-254.
No related articles found!
阅读次数
全文


摘要