1
$\begingroup$

In my linear algebra course I have a problem which goes as follows:

Suppose A is an nxn matrix over field (R) And J(A) is the jordan form of A.

Given α belongs to field R, what is the jordan form of αA?

I have worked out, through trial and error, that the jordan form of αA is simply J(A) with the elements (the different eigenvalues) along the diagonal multiplied by α.

I couldnt think of a formal proof for this problem, I think it might be some simple characteristic I may have overlooked.

Any help is appreciated Thanks!

$\endgroup$

2 Answers 2

0
$\begingroup$

$P^{-1}(\alpha A)P=\alpha P^{-1}AP=\alpha J(A)$ which is similar to a Jordan matrix which is the same as $J(A)$ but the eigenvalues multiplied by $\alpha$.


ok GitGud, take a matrix with $\lambda$ on the diagonal and $\alpha$ on the super diagonal, say the matrix is $J$. Then $P=$Dg$[1,1/\alpha,1/\alpha^2,1/\alpha^3,\ldots]$ is a matrix such that $P^{-1}JP$ is in Jordan form. I hope this is sufficient information for you to continue the proof...(referring to the OP)

$\endgroup$
7
  • $\begingroup$ You skipped the relevant part. $\endgroup$
    – Git Gud
    May 15, 2015 at 13:13
  • $\begingroup$ If I were to multiply P^(−1)AP by a wouldnt I then also multiply the ones in the super diagonal? Then the matrix would no longer be in Jordan form... $\endgroup$ May 15, 2015 at 13:34
  • $\begingroup$ correct, hence the additional statement: "which is similar to..." and the section under the line...so you have to prove that the alphas in the super-diagonal can be transformed through a similarity transformation to ones $\endgroup$ May 15, 2015 at 13:46
  • $\begingroup$ ok I understand. Thanks! $\endgroup$ May 15, 2015 at 14:17
  • $\begingroup$ I realize that I said I understood, but I am still having problems.... I honestly do understand what you said, about P=Dg[1,1/α,1/α2,1/α3,…], I tried it out in a small 4x4 matrix and saw that it worked. But I obviously cant just write "there exists P so that P=Dg[1,1/α,1/α2,1/α3,]" like you said without showing how I arrived at that conclusion. Not to mention I really want to understand thisI've been thinking about this for a couple of days and I still cant think of the similarity transformation you referenced, so I'd appreciate it if you could help me further. Thanks @Christiaan Hattingh $\endgroup$ May 17, 2015 at 20:06
0
$\begingroup$

Since you are given $J(A)$, Jordan form of $A$, then there must exist an invertible $P$ such that:

$$P^{-1}AP = J(A)$$

where $J(A)$ is a Jordan Matrix and thus is composed of smaller Jordan blocks, $J(x)$ for each eigenvalue, $x$, of A.

For each block in $J(x)$ of size $n \times n$, where $x$ is a eigenvalue, $P$ will have $n$ columns which are 'responsible' for the block $J(x)$ in $J(A)$.

We can then take each $i$'th column for $1 \le i le n$ and multpiply it by $α^{i-1}$.

Do this for each Jordan block in $J(A)$ and you will obtain the desired $P$ for $J(αA) = P^{-1}(αA)P$.

$\endgroup$

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .