I was studying Markov chain and in it, it is useful to have a transition matrix $M$ that $\lim_{n \rightarrow \infty}M^n$ exist.

So I thought about the existence of $\lim_{n \rightarrow \infty}A^n$ in general($A$ is a square matrix). and it was easy to prove:

for every diagonalizable matrix $A= QDQ^{-1}$, the $\lim_{n \rightarrow \infty}A^n$ exists if and only if elements of $D$ be in the range $(-1,1]$.

but what about not diagonalizable matrices? what is the condition for them?

EXAMPLES: $E_1$ is not diagonalizable and $\lim_{n \rightarrow \infty}E_1^n$ exists, $E_2$ is not diagonalizable and $\lim_{n \rightarrow \infty}E_2^n$ does not exists. $$ E_1 = \begin{bmatrix}0.1 & 0.1 \\0 & 0.1 \end{bmatrix}, E_2 = \begin{bmatrix}1 & 1 \\0 & 1 \end{bmatrix}$$


You can basically do the same thing with the Jordan normal form. You need all eigenvalues to be either in the interior of the unit disk or $1$. Additionally, if $1$ is an eigenvalue, it cannot be defective (defective eigenvalues of $1$ lead to polynomial growth).


Your Answer

By clicking "Post Your Answer", you acknowledge that you have read our updated terms of service, privacy policy and cookie policy, and that your continued use of the website is subject to these policies.

Not the answer you're looking for? Browse other questions tagged or ask your own question.