# Thread: Is there a way to find matrix from characteristic polynomial?

1. ## Is there a way to find matrix from characteristic polynomial?

i.e., p(lambda) = (1 - lamba)^3. Find A^25.
It was kind of my instinct to choose [1, 0, 0; 0, 1, 0; 0, 0, 1] as my answer, but I'm not sure the proper way of going about it.

2. ## Re: Is there a way to find matrix from characteristic polynomial?

Hey biggerleaffan.

You need to know the eigen-vectors in addition to the eigen-values to know the original matrix.

The eigen-values give the information about how the matrix scales a vector with respect to some eigen-value but it doesn't give the direction.

3. ## Re: Is there a way to find matrix from characteristic polynomial?

Besides the size of the matrix (3x3), no other information was given.

4. ## Re: Is there a way to find matrix from characteristic polynomial?

If the matrix A were diagonalizable then we could write $A= B^{-1}DB$ where D is the diagonal matrix having the eigenvalues of A along its diagonal and B is the matrix havig the eigenvectors of A as columns. That is why we cannot determine A itself from its eigenvalues alone- we need to know the corresponding eigenvectors also.

However, whatever B is, in that case we can write $A^{25}= (B^{-1}DB)^{25}= (B^{-1}AB)(B^{-1}AB)...(B^{-1}AB)= B^{-1}D^{25}B$ which, if D is the identity matrix, which is certainly possible since the three eigenvalues are "1", then $A^{25}= B^{-1}IB= B^{-1}B= I$.

However this is, as I said, "If the matrix A were diagonalizable". Since 1 is a triple eigenvalue it is possible that there do NOT exist three independent eigenvectors. The "eigenspace" may be a subspace of dimension 1 or a subspace of dimension 2 in which we cannot diagonalize A. The best we could do would be the "Jordan Normal Form"

$\begin{bmatrix}1 & 1 & 0 \\ 0 & 1 & 1 \\ 0 & 0 & 1\end{bmatrix}$
if the eigenspace is of dimension 1 or

$\begin{bmatrix}1 & 1 & 0 \\ 0 & 1 & 0 \\ 0 & 0 & 1\end{bmatrix}$
if the eigenspace is of dimension 2.

And finding powers of those is much harder.

Added: Well, not that much harder! A little calculation shows that
$\begin{bmatrix}1 & 1 & 0 \\ 0 & 1 & 1 \\ 0 & 0 & 1\end{bmatrix}^n= \begin{bmatrix}1 & n & 1 \\ 0 & 1 & n \\ 0 & 0 & 1\end{bmatrix}$
and
$\begin{bmatrix}1 & 1 & 0 \\ 0 & 1 & 0 \\ 0 & 0 & 1\end{bmatrix}^n= \begin{bmatrix}1 & n & 0\\ 0 & 1 & 0\\ 0 & 0 & 1\end{bmatrix}$

Of course you would still have the problem that, since these are NOT the identity matrix, $B$ and $B^{-1}$ will not cancel so the $A^n$ will still depend on the eigenvectors.

5. ## Re: Is there a way to find matrix from characteristic polynomial?

Makes sense, thank you!