29.2: Diagonalizable Matrix
- Page ID
- 69445
In class we will be using matrix diagonalization to solve some problems.
Matrix \(A\) is diagonalizable if there exists a diagonal matrix \(D\) that is similar similar to \(A\):
\[ D = C^{-1}AC \nonumber \]
If matrix \(A\) has linearly independent eigenvectors (\(v_1, \ldots, v_n\)) then \(A\) is diagonalizable with the following solution:
\[C = \left[ v_1^T, \ldots, v_n^T \right] \nonumber \]
In other words, each column of \(C\) is a linearly independent eigenvector of \(A\). The diagonal matrix \(D\) is
\[\begin{split} D =
\left[
\begin{matrix}
\lambda_1 & 0 & 0 \\
0 & \ddots & 0 \\
0 & 0 & \lambda_n
\end{matrix}
\right]
\end{split} \nonumber \]
In other-other words, \(D\) consists of the corresponding eigenvalues.
Using numpy
, Diagonalize (i.e. calculate \(C\) and \(D\)) the following matrix:
Verify that \(A\) is in fact Diagonalizable by calculating \(D2 = C^{-1} AC\) and comparing it to your original \(D\) using np.allclose
.
Diagonalization of Symmetric Matrices
One special case is Symmetric Matrices. It can be shown that symmetric Matrices are Diagonalizable and the resulting eigenvectors are not only linearly independent but also orthogonal. Since this is true, the equation changes to:
\[ D = C^{T}AC \nonumber \]
Why do we care if \(C\) is orthogonal? What advantages does the above equation give us?