兰州大学机构库 >数学与统计学院
Chebyshev多项式加速不同分块的SOR迭代方法
Alternative TitleChebyshev polynomial acceleration of different SOR iterative method
段利英
Thesis Advisor郑兵
2008-05-24
Degree Grantor兰州大学
Place of Conferral兰州
Degree Name硕士
KeywordChebyshev多项式 2-block SOR迭代方法 3-block SOR迭代方法 最优外推迭代法 4-block SOR迭代方法
Abstract许多科学计算问题最终将转化为线性系统的求解问题. 而线性最小二乘问题[26]是计算数学一个重要的领域, 也是一个非常活跃的研究领域. 它在大地测量、摄影测量、结构分析、分子结构学等的科学计算中均有广泛的应用.   继古典的迭代法如:Jacobi方法、GS方法、SOR方法、SSOR方法后, 人们又提出了它们的分块迭代法如: BJ方法、BGS方法、BSOR方法、SBSOR方法及AOR方法([10], [39], [41])GSOR方法[42]、SOR-like方法([23],[24],[27])、AG-SOR方法[15]等. 分裂迭代法来解决线性最小二乘问题是非常有效的. 但对于病态的方程组, 一些迭代法就有些不足. 在迭代法中, 迭代格式收敛性及收敛速度就成为关键问题. 下面我们就对一些分裂迭代法进行加速, 通过讨论加速后迭代矩阵的收敛性来证明我们的方法优于加速前的迭代法, 并对病态的迭代矩阵同样有效.   本文用Chebyshev多项式加速不同分块的SOR迭代法, 给出了三种新的迭代格式来解决线性最小二乘问题. 通过讨论它们的收敛性, 并与加速前的迭代法以及最优外推法的收敛性做比较, 我们发现加速后的方法优于加速前的方法, 并且收敛速度是相应最优外推法的二倍. 最后, 我们比较了两种新迭代格式, 并给出了数值例子.
Other AbstractMany scientific computing problems eventually turn into a linear system to solve the problems. And linear least squares problems [26] is a important areas in mathematical calculation, also is a very active research field. It is applied to many scientific computing areas such as geodesy, photogrammetry, structural analysis, and structural molecular.  Following the classic method such as: Jacobi, GS, SOR, SSOR methods, people put forward their block iterative methods such as: BJ method, BGS method, BSOR method, SBSOR method, AOR ([10], [39], [41])method, GSOR method [42], SORlike approach ([23], [24], [27]), AGSOR method [15], etc. Splitting iterative method is very effective to solve linear least squares problems, but for morbid equation, a number of iterative methods have their deficient. In the iterative method, the convergence and convergence rate play a major role. Below we shall accelerate some different iterative methods. Through the discussions of the convergence of accelerated iterative matrix, we prove that our method is superior to the iterative method before, and is efficient for sick equations.    In this paper, we apply the Chebyshev polynomials to the different SOR iterations, and present three new iterative format to solve linear least squares problems. Through discussing of the convergence, and comparing with the iteration method before and the extrapolation method, we found that the new method is superior to the method before. And the corresponding optimal convergence rate is twice the extrapolation method. Finally, we compare the two new methods and give numerical examples.
URL查看原文
Language中文
Document Type学位论文
Identifierhttps://ir.lzu.edu.cn/handle/262010/225730
Collection数学与统计学院
Recommended Citation
GB/T 7714
段利英. Chebyshev多项式加速不同分块的SOR迭代方法[D]. 兰州. 兰州大学,2008.
Files in This Item:
There are no files associated with this item.
Related Services
Recommend this item
Bookmark
Usage statistics
Export to Endnote
Altmetrics Score
Google Scholar
Similar articles in Google Scholar
[段利英]'s Articles
Baidu academic
Similar articles in Baidu academic
[段利英]'s Articles
Bing Scholar
Similar articles in Bing Scholar
[段利英]'s Articles
Terms of Use
No data!
Social Bookmark/Share
No comment.
Items in the repository are protected by copyright, with all rights reserved, unless otherwise indicated.