6-1-2IterativeTechniMatrixAlgebra案例.ppt

《数值分析》 主讲教师：陈松林 Ch7 Iterative Technique in Matrix Algebra There are many directive stages for solving linear equations, however, the complexity in time and space are faced in solving large systems such as guided missile launching. §7.1 Norm of vector (or Matrix) §7.2 Eigenvalues and Eigenvectors §7.3 Iterative Techniques for Solving Linear Systems §7.4 Error Bounds and Iterative Refinement §7.5 The Conjugated Gradient method §7.1 Norm of Vector (or Matrix) Def 7.2 Suppose X is a linear space, (.,.) is a function with two arguments defined from X to C and satisfies： （1）(nonnegative) (x,x)≥0, (x,x)=0 if and only if x=0； （2）(linearity for the first element) for any a,b∈C1,(ax+by,z)= a(x,z)+ b(y,z); (3) (conjugate symmetry): Then (.,.) is called an inner (scalar) product on X. Remarks: The CBS inequality holds for any (e.x. infinite dimension space )inner space; (X;+,.; ||.||; (.,.)): set; linear space; normed linear space; inner space Minkovsikii Inequality i.e. Triangular inequality it is easy to prove it in an inner product space apart from other spaces. Homework (p428):2;4(c,d);12 §7.2 Eigenvalue 、 Eigenvector  some relative properties Def. Characteristic polynomial; Eigenvalue Eigenvector; Spectrum radius; Matlab Some theorems about induced norm Relations between spectrum radius induced norms of matrix Proof: Remark. ρ(A) be a lower bound of square matrix norms, and be the inf of induced norms, thus, it is also the inf of square matrix norms. ? Problems for consideration A useful Lemma Some related theorem* Limit of vector or matrix sequence §7.3 Iterative Techniques for Solving Linear Systems §7.3.0 Convergence Error Estimation theory §7.3.1 Jacobi Iteration §7.3.2 Gauss-Seidel Iteration §7.3.3 The Sufficient Conditions of convergence for Iterations An example diverges when so-called Jacobi iteration applied Convergence Error Estim