I have the following information: the characteristic polynomial of $A$ is $p_A(t)=(t-4)^3(t+6)^2$ and the minimal polynomial is $q_A(t)=(t-4)^2(t+6).$

I'm having problems seeing how one would work backwards with just this information.

My attempt at JCF.

$\begin{pmatrix} 4&1&0&0&0\\ 0&4&0&0&0\\ 0&0&4&0&0\\ 0&0&0&-6&0\\ 0&0&0&0&-6\\ \end{pmatrix}$

  • $\begingroup$ NB in general the minimal and characteristic polynomials of a matrix do not determine its Jordan form uniquely. For example, there are two Jordan matrices $J$ (up to rearrangement) for which $p_A(t) = t^4$ and $m_A(t) = t^2$. $\endgroup$ – Travis Willse Jul 22 at 2:44

The multiplicity of an eigenvalue as a root of the characteristic polynomial is the size of the block with that eigenvalue in the Jordan form. The size of the largest sub-block (Elementary Jordan Block) is the multiplicity of that eigenvalue as a root of the minimal polynomial.

Can you see how this determines the matrix?

Added: For example, the block with the eigenvalue $4$ in its diagonal must be $3\times3$, since that is the multiplicity of $3$ as a root of $p_A$. We also know by the definition of the Jordan form that this block is made up by smaller sub-blocks, each with $4$ in its diagonal and $1$s above, and with decreasing size as one goes down left to right.

Now, without any further information, since the larger block is of size $3\times 3$, there are three possibilities for our subblocks:

  • Three sub-blocks of $1 \times 1$, giving the following large block: $\left (\begin{matrix}4&0&0 \\0&4&0 \\0&0&4\end{matrix} \right )$

  • One subblock of size two, and since the other must not be bigger and their sizes must "add up" to $3 \times 3$ we get: $\left (\begin{matrix}4&1&0 \\0&4&0 \\0&0&4\end{matrix} \right )$.

  • A single large block: $\left (\begin{matrix}4&1&0 \\0&4&1 \\0&0&4\end{matrix} \right )$

But we know that the dimension of the largest sub-block must be the multiplicty of $4$ as a root of the minimal polynomial, hence the only possibility is: $$ \left (\begin{matrix}4&1&0 \\0&4&0 \\0&0&4\end{matrix} \right )$$

| cite | improve this answer | |
  • $\begingroup$ I have a matrix that is the eigenvalues on the diagonal, 1's above the eigenvalues, and zeroes everywhere else. I sort of see this relationship. $\endgroup$ – emka Mar 31 '15 at 2:30
  • $\begingroup$ I added to my original answer how to figure out the form of the block with $4$s in its diagonal. $\endgroup$ – Reveillark Mar 31 '15 at 2:38
  • $\begingroup$ I put an attempt above. $\endgroup$ – emka Mar 31 '15 at 2:46
  • $\begingroup$ Your attempt is correct :) I'm assuming you used that the multiplicty of $-6$ as a root of the minimal is $1$ and hence there are two sub-blocks of sizes $1\times 1$ each. $\endgroup$ – Reveillark Mar 31 '15 at 2:48
  • $\begingroup$ I was very tempted to stick a 1 in the lower part of that matrix. It took me a rather long moment to make the connection. Your prompts were very helpful. $\endgroup$ – emka Mar 31 '15 at 2:50

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.