3
$\begingroup$
  1. For a square complex/real matrix $A$, $A$ and $A^T$ have the same set of eigenvalues, each with same algebraic multiplicities, since their characteristic polynomials are the same.

    I wonder for each eigenvalue, are its geometric multiplicities for $A$ and for $A^T$ the same?

  2. Similar question for $A$ and $A^H$, where $H$ means conjugate and transpose, and the relation between their eigenvalues is conjugate.

Thanks!

$\endgroup$
  • $\begingroup$ Note that if $J$ is the Jordan form for $A$, then $A^t$ is similar to $J^t$. $\endgroup$ – Gerry Myerson Nov 27 '12 at 3:14
  • $\begingroup$ @Gerry: Thanks! Although $J^t$ or $J^H$ may not be a Jordan form, it seems that geometric multiplicity stays the same for the same/conjugate eigenvalue? $\endgroup$ – Tim Nov 27 '12 at 3:24
  • $\begingroup$ Think about why $J$ gives you information about multiplicities, and whether you can extract the same information from $J^t$ (or $J^h$). $\endgroup$ – Gerry Myerson Nov 27 '12 at 4:49
  • $\begingroup$ Hmmm, I am not sure. $\endgroup$ – Tim Nov 27 '12 at 5:12
2
$\begingroup$

Another attempt:

First note that for a square matrix $M$, we have $\dim \ker M = \dim \ker M^T =\dim \ker M^*$. Hence $\dim \ker (\sigma I-M) = \dim \ker (\sigma I-M)^T = \dim \ker (\sigma I-M^T)$, and $\dim \ker (\sigma I-M) = \dim \ker (\sigma I-M)^* = \dim \ker (\overline{\sigma} I-M^*)$.

The geometric multiplicity of an eigenvalue $\lambda$ of $A$ is $\dim \ker (\lambda I -A)$.

It follows that $A$ has an eigenvalue $\lambda$ of geometric multiplicity $k$ iff $A^T$ has an eigenvalue $\lambda$ of geometric multiplicity $k$.

It follows that $A$ has an eigenvalue $\lambda$ of geometric multiplicity $k$ iff $A^*$ has an eigenvalue $\overline{\lambda}$ of geometric multiplicity $k$.

$\endgroup$

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.