第8章 矩阵特征值拥碾特征向量的计算.pptVIP

第8章 矩阵特征值和特征向量的计算;PA(?)是?的高次的多项式，它的求根是很困难的。 设法通过数值方法是求它的根。;定理1 ：A?R n?n，?1, …, ?n为A的特征值，则;定义 设矩阵A, B?R n?n，若有可逆阵P，使 则称A与B相似。;8.1 幂法和反幂法;故有 xk = Akx0 =β1?1kξ1+β2?2kξ2+…+βn?nkξn;实际计算时，常把每一步计算的迭代向量xk规范化。;所以;用乘幂法求A的按模最大的特征值和相应特征向量.;8.1.2 加速技术; 2.原点位移法; 反幂法是求矩阵按模最小的特征值和相应特征向量的方法.;也可将上式改写成;; Jacobi方法是求实对称矩阵全部特征值和特征向量的一种矩阵变换方法。;平面解析几何中的平面坐标旋转变换;称Rij(?)为平面旋转矩阵或Givens变换矩阵. ;Evaluation only. Created with Aspose.Slides for .NET 3.5 Client Profile 5.2.0.0. Copyright 2004-2011 Aspose Pty Ltd.;Evaluation only. Created with Aspose.Slides for .NET 3.5 Client Profile 5.2.0.0. Copyright 2004-2011 Aspose Pty Ltd.; 设实对称矩阵A=(apq)n?n ,记B=RijT(?)ARij(?)=(bpq)n?n则它们元素之间有如下关系:;所以有;只需角?满足;非对角线元素的平方和，; ; ; ;的全部特征值.;所以;从而A的特征值可取为 ?1?2.125825, ?2?8.388761, ?3?4.485401 ; 为了减少搜索非对角线绝对值最大元素时间, 对经典的Jacobi方法可作进一步改进.; ;8.3 QR方法; QR算法的收敛性 ;QR方法收敛性 ;QR方法收敛性; QR方法运算量很大，为了减少运算量，常在使用QR方法之前把矩阵A简化为拟上三角矩阵。或称之为海森伯格矩阵(次对角元以下的元素全为零)。;定义8.2;x;由此可得;;与平面旋转变换不同的是，镜面反变换可成批的消去向量的非零元;Evaluation only. Created with Aspose.Slides for .NET 3.5 Client Profile 5.2.0.0. Copyright 2004-2011 Aspose Pty Ltd.;Evaluation only. Created with Aspose.Slides for .NET 3.5 Client Profile 5.2.0.0. Copyright 2004-2011 Aspose Pty Ltd.;Evaluation only. Created with Aspose.Slides for .NET 3.5 Client Profile 5.2.0.0. Copyright 2004-2011 Aspose Pty Ltd.;Evaluation only. Created with Aspose.Slides for .NET 3.5 Client Profile 5.2.0.0. Copyright 2004-2011 Aspose Pty Ltd.;Evaluation only. Created with Aspose.Slides for .NET 3.5 Client Profile 5.2.0.0. Copyright 2004-2011 Aspose Pty Ltd.;Evaluation only. Created with Aspose.Slides for .NET 3.5 Client Profile 5.2.0.0. Copyright 2004-2011 Aspose Pty Ltd.;Evaluation only. Created with Aspose.Slides for .NET 3.5 Client Profile 5.2.0.0. Copyright 2004-2011 Aspose Pty Ltd.;;;２、用 Givens变换对上Hessenberg阵作QR分解;Evaluation only. Created with Aspose.Slides for .NET 3.5 Client Profile 5.2.0.0. Copyright 2004-2011 Asp