It's a fact that similar matrices have the same eigenvalues. Prove that matrices with the same eigenvalues aren't necessarily similar.
0__________________________________________________________________0
By the way, \(A\) is similar to \(B\) if there is some matrix \(P\) that relates them like this: \[A = P^{-1}BP\]
i don't get this, is this a question or a lesson?
A fun challenge question for anyone who's interested >:D
count me in with my algebra 1 education
if the matrices \(A,\;B\) have the same eigenvalues, then, the subsequent condition holds: \[\Large \det \left( {A - \lambda I} \right) = \det \left( {B - \lambda I} \right)\]
Yup, that's true. Can you use that to show: \[A \ne P^{-1} B P\]
taco @ShadowLegendX
for example we can consider two similar matrices A, B such that A is nilpotent and B is not nilpotent
How will that show that two matrices with the same eigenvalues are NOT similar though? I think you might have more success if you start the other way round with matrices that have the same eigenvalues, then show they lead to matrices that are not similar.
if: \[\Large \begin{gathered} {A^k} = 0 \hfill \\ B = {N^{ - 1}}AN \hfill \\ \end{gathered} \] then: \[\Large {B^k} = {N^{ - 1}}{A^k}N = 0\] which is false by hypothesis
This shows that nilpotent matrices are not similar to a matrix that's not nilpotent. It's a nice proof I like this, but it doesn't have anything to do with what I'm asking.
I have to think better to this question, I will answer tomorrow, now I have to go thanks for your suggestion @Kainui
Ahhh I think I see your reasoning now, but I think it needs a little more before it works. I think you're saying A and B have the same eigenvalues and are similar. Then you show that they can't be similar by the fact that A is nilpotent and B is not. The only thing left to show is that you can construct A and B with the same eigenvalues, but with one being nilpotent and the other not. I am not sure if this is possible or not.
there is a trivial example...you are working too hard
If B is the 2x2 identity matrix, what does A have to be if A and B were similar?
I agree that there's a simple way to do this, since we just have to find a single counter example. But I kinda like this more general answer that we're looking for it's kinda fun. Here's the proof I had in mind: \[\begin{pmatrix} 1 & 1 \\ 0 & 1 \end{pmatrix} = A\] A has the same eigenvalues as the identity matrix but can't be similar since this is a contradiction: \[A=P^{-1}IP = I \]
Join our real-time social learning platform and learn together with your friends!