一线性系统投影迭代解法的推广和比较结果.pdf

A ∈ Cm ×n b ∈ Cm Æ x ∈ Cn  Ax = b Æ ℄  Æ Krylov Æ Krylov ÆÆÆ YOUSEF SAAD [1996] (cf. [22]) [27] Æ Gauss-Seidel Gauss- Seidel Æ [27] Æ Gauss-Seidel Æ Æ Petrov-Galerkin(Galerkin) Abstract In [27], the authors obtained a new iterative method for solving linear sys- tems. This method can be considered as a projection method which uses a two- dimensional space at each step. In this paper, we generalize this method to a three-dimensional projection process. And a different approach is established, which is both theoretically and numerically proven to be better than (or at least the same as) [27]’s. In Chapter 2, firstly we illustrate Gauss-Seidel using projection steps, and give further analysis for the new iterative method of [27] in the Section 3 of Chapter 2. From the projection technique point of view, this method can be considered as a generalizing of Gauss-Seidel method which uses a two-dimensional space at each step. At the end of this chapter, we generalize this method to a three-dimensional projection process. As the theory in this chapter indicates, the new iterative process gives better (or at least the same) reduction of the error than the iterative method proposed in [27]. In Chapter 3, the presented numerical examples show that, in most cases, the convergence rate of the new iterative process is better than (or at least the same as) that of the iterative method proposed in [27]. Keywords: Projection technique; Linear system; Petrov-Galerkin (G