1
$\begingroup$

Suppose $m_A(\lambda)$ is the minimal polynomial of $A_{n\times n}$.

  1. Show that if $\lambda$ is an eigenvalue of $A$, then $m_A(\lambda)=0$

  2. Show that $m_A(x)$ of a diagonalizable matrix $A$ divides the characteristic polynomial $f_A(x)$.

For the second question, I was thinking we could use the fact that if $p(x)$ is a polynomial such that $p(A)=0$, then $m_A(x)$ divides $p(x)$. I think this requires a proof of number 1, though, since a diagonalizable matrix $A$ will have the eigenvalues on the diagonal, but I'm not really sure.

$\endgroup$
  • $\begingroup$ The second one is actually true in general, by Cayley-Hamilton. But for a diagonalizable matrix, the characteristic polynomial and minimal polynomial are easy to find, so you can check it directly. For 1, what happens when you apply $m_A(A)$ to an eigenvector? $\endgroup$ – Nishant Dec 3 '14 at 3:59
0
$\begingroup$

Hints: For the first question: let $x$ be an eigenvector. Consider $m_A(A)x$. For the second: yes, that will work.

$\endgroup$
  • $\begingroup$ What does $m_A(A)x$ represent? $\endgroup$ – Carley Dec 3 '14 at 4:10
  • $\begingroup$ $m_A(A)$ is the matrix that comes out of plugging $A$ into the polynomial $m_A$. Take that matrix, and multiply it by $x$. $\endgroup$ – Omnomnomnom Dec 3 '14 at 4:26
  • $\begingroup$ OK, I think I got #1. For #2, what is $m_A(x)$ of a diagonal matrix $A$? Why is it zero? $\endgroup$ – Carley Dec 3 '14 at 4:36
  • $\begingroup$ Note that for any diagonal matrix $$ D = \pmatrix{d_1\\&\ddots\\&&d_n} $$ and any polynomial $p$, we have $$ p(D) = \pmatrix{p(d_1)\\&\ddots\\&&p(d_n)} $$ $\endgroup$ – Omnomnomnom Dec 3 '14 at 4:41
  • $\begingroup$ When we say that "matrix $A$ is diagonalizable," are we really saying that $A$ can be written as a diagonal matrix where its eigenvalues are on the diagonal? I thought it just meant that $A$ could be factored into $SBS^{-1}$, where $B$ is the diagonal matrix with eigenvalues on the diagonal. $\endgroup$ – Carley Dec 3 '14 at 4:44

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.