• 论文 • 上一篇    下一篇

一个无约束优化问题并行算法的异步执行

郑芳英,韩丛英,贺国平   

  1. 浙江理工大学数学科学系;上海交通大学数学系;山东科技大学信息科学与工程学院 杭州310018;上海200240 山东科技大学信息科学与工程学院;山东青岛266510;山东青岛266510
  • 出版日期:2007-01-20 发布日期:2007-01-20

郑芳英,韩丛英,贺国平. 一个无约束优化问题并行算法的异步执行[J]. 数值计算与计算机应用, 2007, 28(1): 63-70.

AN ASYNCHRONOUS EXECUTION OF PARALLEL VARIABLE TRANSFORMATION ALGORITHM WHICH WAS FOR UNCONSTRAINED OPTIMIZATION

  1. Zheng Fangying (Department of Mathematical Science,Zhejiang Sci-Tech University,Hangzhou 310018,China) Han CongYing (Department of Mathematics,Shanghai Jiao Tong University,Shanghai 200240,China;School of Information Science and Engeering,Shandong University of Science and Technology,Qingdao 266510,Shandong,China) He Guoping (School of Information Science and Engeering,Shandong University of Science and Technology,Qingdao 266510,Shandong,China)
  • Online:2007-01-20 Published:2007-01-20
在这篇文章里,我们对Fukushima提出的关于无约束优化问题的PVT算法作了改进,提出利用PVD算法中的PVD-方向来构造的PVT-变换矩阵,得到一个更适合于异步执行的PVT算法,从而减少各处理机之间的等待时间,提高并行机的并行效率.文中证明算法具有线性收敛速度,且其线性收敛比与处理机个数无关,该结果改进了中的结果,更适合于并行计算.
In this paper,we proposed an asynchronous execution for parallel variable transformation(PVT)algorithm which was for unconstrained optimization.We employed PVD-direction to construct the transformation matrix and obtained a specified PVT algorithm in which we need not to wait between processors.Also we obtained the linear convergence rate which was independent of the number of processors which cannot be obtained in[1].
()

[1]M.Fukushima,Parallel variable transformation in unconstrained optimization
[J],SIAM,J.on Optimization,8(1998),658-672.
[2]M.C.Ferris and O.L.Mangasarian,Parallel variable distribution
[J],SIAM Journal on Optimiza- tion,4(1994),815-832.
[3]M.Solodov,New inexact parallel variable distribution algorithms
[J],Computational Optimization and Applications,7(1997),165-182.
[4]LiPing Pang and ZunQuan Xia,A modified PVT algorithm
[J],OR Transactions,6:4(2002),8-14.
[5]D.P.Bertsekas and J.N.Tsitsiklis,Parallel and Distribution Computation
[M],Numerical Methods, Prentice-Hall:Englewood Cliffs,NJ,1989.
[6]E.Yamakawa and M.Fukushima,Testing Parallel Variable Transformation on Vpp500
[J],Com- putational Optimization and Applications,13(1999).,253-274.
[7]J.J.Moré,B.S.Grabow and K.E.Hillatrom,Testing Unconstrained Optimization Software
[J], ACM Trans.Math.Software,7(1981),17-41.
[8]S.G.Nash,Newton-type minimization via the Lanczos Method
[J],SIAM J.on Numerical Analysis, 7(1984),770-788.
[9]S.S.Oren,Self-scaling variable metric(SSVM)algorithms Part 2:Implementation and experi- ments.Management Science
[J],20(1974),863-874.
No related articles found!
阅读次数
全文


摘要