解线性系统的几种预条件迭代法-应用数学专业论文.docxVIP

摘 摘 要 摘 摘 要 摘 要 数值计算是科学计算的一个重要环节, 而在数值计算中, 一类很重要的问题就是 线性方程组的求解. 线性方程组的求解在工程与科学的许多领域中都有广泛应用. 对 于大型线性方程组的求解, 迭代方法己取代直接法成为最重要的一类求解方法. 而迭 代方法好坏的标准通常是通过收敛速度来刻画. 提高迭代算法的收敛速度, 在迭代算 法研究中有重要的理论价值和实用意义. 引进非奇异预条件因子和通过该因子作用加 快迭代法的收敛速度是本研究领域的重要热点. 本文主要讨论的是用预条件迭代法求解线性方程组, 特别是在迭代法收敛的情况 下, 如何加快迭代法的收敛速度. 主要内容和创新点包括: 第一, 提出了一种新的预条件因子, 并把该预条件因子应用到 Gauss-Seidel 、Jacobi 这两种迭代法中. 证明了这两种预条件迭代法可以加速对应的预条件迭代法或经典迭 代法, 并得到相应的比较定理和给出了收敛最快时的系数取值, 通过数值例子说明这 些方法是有效的. 第二, 针对某些特殊矩阵提出两种新预条件因子, 并把它们分别运用到经典 AOR 迭代法中. 然后证明了这两种预条件方法可以改善经典 AOR 方法, 通过数值实验说 明这些方法的有效性. 最后, 总结了全文的工作和指出了有待进一步研究的一些问题. 关键词: 特殊矩阵; 迭代法; 预条件因子; 预条件迭代法; 谱半径 万方数据 I – – – PAGE IV – 万方数据 – – PAGE III – 万方数据 Abstract Calculation of numerical value is a significant link solving the linear system of equations, which is a very important question in calculation of numerical value. Solving the linear system of equations has a wide range of applications in science and engineer. To solve large linear system, iterative methods have become one of the most important methods instead of direct method. The rule whether the iterative is good is usually described by convergence rate. Accelerating convergence rate of iterative method has important academic value and practical significance in research of iterative algorithms. Quoting a nonsingular pre-conditioner and improving the convergence rate of iterative method by the pre-conditioner are important hot spot of the field of study. The thesis mainly deals with the preconditioned iterative method to solve the linear system of equations and discusses how to accelerate the rate of convergence in the condition of precondition. The main contents and innovations are as follows: Firstly, the writer gives a new pre-conditioner, and the pre-conditioner is applied to Gauss-Seidel and Jacobi methods. That two preconditioned methods can accelerate the corresponding preconditioned iterative method or iterative method is proved. Some comparison theorems and the optimal parameter are obtained and illustr