Consider the map $x_{i+1} = f(x_{i})$ and let $x^{*}$ be a fixed point then the fixed point is said to be Lyapunov stable if $\forall \epsilon > 0 $ there exist $\delta > 0$ such that $x_{i} \in B_{\epsilon}(x^{*}) \forall x_{0} \in B_{\delta}(x^{*}) $ and $\forall i \geq 0$ and

the fixed point is asymptotically stable if it is Lyapunov stable and there exist $\delta >0$ such that $x_{i} \rightarrow x^{*}$ as $i \rightarrow \infty \forall x_{0} \in B_{\delta}(x^{*})$.

Now we take the case of linear maps $x_{i+1} = Ax_{i}$ where $A$ is realvalued $n \times n$ matrix and for any $x_{0} \in \Bbb{R}^n$ we have $x_{i} = A^{i}x_{0}$

Then I was thinking of proving rigorously the following -

$1. $ If $|\lambda| <1$, for every eigenvalue of $A$ then $0$ is an asymptotically stable fixed point.

$2. $ If $|\lambda|>1$ for some eigenvalue $\lambda$ of $A$ then $0$ is not a Lyapunov stable fixed point $x_{i+1} = Ax_{i}$.

I think that if $x$ is an eigenvector of which $\lambda$ is a corresponding eigenvalue then due to factor coming from $\lambda^n$ which as $n \rightarrow \infty$ tends to zero so seems to be asymptotically stable but here $x_{0}$ is any real $n$ dimensional vector.

  • Consider $x_0$ in an eigenbasis. Or use if for the spectral radius $\rho(A)<1$ that then there is a vector norm so that $\|A\|<1$ in the associated operator norm. – LutzL Sep 13 at 11:42

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.

Browse other questions tagged or ask your own question.