🌊

Matrix diagonalization

2024/02/24に公開

Let A \in \R^{N \times N}, P \in \R^{N \times N}, and \set{\lambda_n \in \R}_{n=1}^N.

The diagonalization of A is described as:

\begin{align*} A = P \Lambda P^{-1} ,\quad \Lambda = \begin{pmatrix} \lambda_1 & & & \\ & \lambda_2 & & \\ & & \ddots & \\ & & & \lambda_N \\ \end{pmatrix} \end{align*}

where \lambda_1, \lambda_2, \dots, \lambda_N are eigenvalues of P, and the n-th column vector P_n \in \R^N is the n-th eigenvector corresponding to \lambda_n.


\begin{align*} A P_n &= \lambda_n P_n \\ \end{align*}

because

\begin{align*} A P_n &= P \Lambda P^{-1} P_n \\ &= P \Lambda e_n \\ &= P \lambda_n e_n \\ &= \lambda_n P e_n \\ &= \lambda_n p_n \\ \end{align*}

Let's consider A y where A \in \R^{N \times N}, y \in \R^N.

When we have the diagonalization A = P \Lambda P^{-1}, we can calculate A y more easily, like,

\begin{align*} A y &\overset{\flat}{=} A \left( \sum_{n=1}^N w_n P_n \right) \\ &= \sum_{n=1}^N w_n A P_n \\ &= \sum_{n=1}^N w_n (\lambda_n P_n) \\ \end{align*}

\overset{\flat}{=} comes from y being represented as y = \sum_{n=1}^N w_n P_n (see: this link).

And w = A^{-1}y, \quad w = \set{w_n}_{n=1}^N.

Discussion