0
$\begingroup$

Let $T$ be a linear operator on a finite dimensional vector space $V$ over a field $F$. Suppose that the characteristic polynomial of $T$ splits into linear factors over $F$. Let $\lambda_1,\ldots, \lambda_k$ be all the distinct eigen values of $T$ and $E_{\lambda_1} , . . . , E_{\lambda_k}$ be their corresponding eigen spaces. How to prove $T$ is diagonalizable if and only if the multiplicity of $\lambda_i$ is equal to the $\dim(E_{\lambda_i})$ for every $i$, where $i = 1 , \ldots, k$? Further how to prove that the sum $E_{\lambda_1}+\cdots+ E_{\lambda_k}$ is always direct and equals the subspace of $V$ spanned by $\{ x \in V \mid x \mbox{ is an eigen vector of } T\}$?

$\endgroup$
1
  • $\begingroup$ @Solumilkyu thanks for the edit ☺ $\endgroup$
    – Shona
    May 9 '16 at 8:53
0
$\begingroup$

I suppose you must know that the multiplicity of every linear factor $\;(x-\lambda)\;$ in the minimal polynomial equals the maximal size a Jordan block corresponding to $\;\lambda\;$ has in the Jordan form of the matrix. Since this is $\;1\;$ in our case this means the Jordan form of the matrix is diagonal, i.e. our matrix is diagonalizable.

$\endgroup$
8
  • $\begingroup$ Its if and anly if case so wont we have to prove it by the given that the multiplicity of mi is equal to the dim(E lambda i) for every i , where i = 1 , . . . , k ? And E_lambd_i are not used here $\endgroup$
    – Shona
    May 9 '16 at 8:50
  • 1
    $\begingroup$ Yes, of course it is an iff statement, and I don't need the eigenspaces for the part I addressed if you know what I wrote in the lines 1-2 $\endgroup$
    – DonAntonio
    May 9 '16 at 8:55
  • $\begingroup$ I know . But i tried by taking basis of eigen vectors whose dimension is equal to the multiplicity of λi is equal to the dim(Eλi) for every i where i=1,…,k $\endgroup$
    – Shona
    May 9 '16 at 9:06
  • 1
    $\begingroup$ @Shona Exactly...and then you find that $\;\dim E_\lambda=m_\lambda\;$ , where $\;m_\lambda=$ the algebraic multiplicity of $\;\lambda\;$ , right? Well, this means precisely that there's a basis for the whole space which is all eigenvectors of the matrix, and this is equivalent to be diagonalizable ! $\endgroup$
    – DonAntonio
    May 9 '16 at 9:12
  • 1
    $\begingroup$ @Shona There you go. $\endgroup$
    – DonAntonio
    May 9 '16 at 9:16

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.