Loading [MathJax]/extensions/TeX/boldsymbol.js

 

 

 

Discussion of Householder's method for eigenvalues

It remains to find a recipe for determining the transformation \mathbf{S}_n . We illustrate the method for \mathbf{S}_1 which we assume takes the form \mathbf{S_{1}} = \left( \begin{array}{cc} 1 & \mathbf{0^{T}} \\ \mathbf{0}& \mathbf{P} \end{array} \right), with \mathbf{0^{T}} being a zero row vector, \mathbf{0^{T}} = \{0,0,\cdots\} of dimension (n-1) . The matrix \mathbf{P} is symmetric with dimension ( (n-1) \times (n-1) ) satisfying \mathbf{P}^2=\mathbf{I} and \mathbf{P}^T=\mathbf{P} . A possible choice which fullfils the latter two requirements is \mathbf{P}=\mathbf{I}-2\mathbf{u}\mathbf{u}^T, where \mathbf{I} is the (n-1) unity matrix and \mathbf{u} is an n-1 column vector with norm \mathbf{u}^T\mathbf{u} (inner product).