0
$\begingroup$

I have a question where I am trying to find the general solution of a linearised system, which I have linearised. I am just having difficulty obtaining the correct corresponding eigenvectors to my eigenvalues. I am usually fine at these but I am not getting the correct answer (I have solutions). I will highlight my issue below.

My matrix A for my linearised system is given by

$$A=\begin{bmatrix}-1 & 0 \\-1 & 0\end{bmatrix}$$

and my eigenvalues are $\lambda_1=0$ and $\lambda_2=-1$ which I have found with no problem

However, I then know eigenvectors $v_1$ and $v_2$ are found as follows, again usually fine with:

$(A-\lambda_1 I)v_1=0$ and $(A-\lambda_2 I)v_2=0$

i.e $(A-\lambda_1 I)v_1= \left( \begin{array}{ccc} 0 & 0 \\ -1 & 1 \end{array} \right)$$v_1=0$

So, we have $\left( \begin{array}{ccc} 0 & 0 \\ -1 & 1 \end{array} \right)\left( \begin{array}{ccc} v_1 \\ v_2 \end{array} \right)=0$

So $0v_1+0v_2=0$ and $-1v_1+1v_2=0$

i.e $-1v_1=-1v_2 \implies v_1=v_2=1$ so corresponding eigenvector to eigenvalue $\lambda_2=-1$ is $\left( \begin{array}{ccc} 1 \\ 1 \end{array} \right)=0$

This is all fine. Here is where I am stuck. Now to find the corresponding eigenvector to eigenvalue $\lambda_2=0$

We want to find $v_2$ such that $(A-\lambda_1 I)v_2= Av_2= \left( \begin{array}{ccc} -1 & 0 \\ -1 & 0 \end{array} \right)$$v_2=0$

So this is where I am stuck:

trying to find eigenvector $v_2$ we have $-1v_1+0v_2=0$ and $-1v_1+0v_2=0$

The solution says $v_2=\left( \begin{array}{ccc} 0 \\ 1 \end{array} \right)$ However I cannot see how the writer has got this. Please forgive my stupidity but obviously if $-1v_1+0v_2=0$ then $v_1=0$ but how do they find $v_2$? the fact that both $v_2$ coefficients are zero is confusing me, any help much appreciated! Thanks

$\endgroup$
  • $\begingroup$ Note: $\lambda_2=-1$ $\endgroup$ – EPS May 2 '15 at 0:12
  • $\begingroup$ typo at beginning - fixed now - answer still same though $\endgroup$ – Bernard.Mathews May 2 '15 at 0:15
  • $\begingroup$ You are confused because you are using $v_2$ for a vector as well as for a scalar at the same time! $\endgroup$ – EPS May 2 '15 at 0:16
  • $\begingroup$ i know i dont know why he has done that - but no I am confused as to how to work out, say $c_2$ instead if this is the case, i.e $-1c_1+c_2=0$ and $-1c_1+c_2=0$ $\endgroup$ – Bernard.Mathews May 2 '15 at 0:18
1
$\begingroup$

First of all you have confusing notation, as you use $v_2$ for both the vector and the component of the vector.


That said, in my answer, $v$ is a vector with components $v_1$ and $v_2$

You have a system of equations, the first $$-v_1= 0$$

and the second $$-v_1 =0$$

Now, any pair $v_1, v_2$ which satisfies both equations is an eigenvector. Clearly, $v_1$ must be $0$, and clearly, if $v_1=0$, the pair will satisfy the system for ANY value $v_2$. Therefore,

$$(A-0\cdot I) v = 0$$

if and only if $v$ is a multiple of $(0,1)$.

$\endgroup$
  • $\begingroup$ Why $(0,1)$? Im just confused about the $1$ to be honest, because when I look at it, it could also be 0 surely? $\endgroup$ – Bernard.Mathews May 2 '15 at 0:19
  • 1
    $\begingroup$ @Bernard.Mathews For any $v_2$, the vector $(0,v_2)$ is a multiple of $(0,1)$. $\endgroup$ – 5xum May 2 '15 at 0:21
  • 1
    $\begingroup$ @Bernard.Mathews No, not just a positive integer. $(0,-0.89498345703984)$ is also an eigenvector. $\endgroup$ – 5xum May 2 '15 at 0:22
  • 1
    $\begingroup$ @Bernard.Mathews The way I understand it is that really, matrices don't have eigenvectors, they have eigenspaces. For each eigenvalue $\lambda$, the matrix $A$ has an eigenspace which is equal to all vectors for which $Ax=\lambda x$. This space is the nullspace of $Ax - \lambda x$, and it may have more than one dimension. In your particular case, the nullspace is one-dimensional, and the basis vector of the nullspace is $(0,1)$. So, in a way, when you say "the eigenvector is $(0,1)$", you are saying "the basis of the nullspace is $\{(0,1)\}$" $\endgroup$ – 5xum May 2 '15 at 0:28
  • 1
    $\begingroup$ @Bernard.Mathews And no, it's not stupid. It's a very important concept and it's vital that you fully grasp it. $\endgroup$ – 5xum May 2 '15 at 0:28

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.