2
$\begingroup$

The number of linearly independent eigenvectors for eigenvalue $1$ for the given matrix ?

$\begin{bmatrix} 1 & 3 & 2 \\ 0 & 4 & 2\\ 0 &-3 & -1 \end{bmatrix}$


Eigenvalues are $1,1,2$ for the above matrix .So, $1$ has multiplicity $2$ and $2$ has multiplicity $1$.

After putting the value $1$, I am getting $ i= $any value, and the relation $3j = -2k$.


Now, How to proceed further ?

$\endgroup$
  • 3
    $\begingroup$ Hint: You do not have a deficient matrix here and can find three linearly independent eigenvectors, even given the repeated eigenvalue. $\lambda_{1,2,3} = 1,1,2$, with eigenvectors $v_1 = (1, 0, 0), v_2 = (0,-2,3), v_3 = (-1,-1,1)$. $\endgroup$ – Moo Feb 5 '17 at 18:19
  • $\begingroup$ @copper.hat Can you explain ? $\endgroup$ – Jon Garrick Feb 5 '17 at 18:21
  • 2
    $\begingroup$ @JonGarrick: As Moo pointed out, the are two li. eigenvectors of $A$ for the eigenvalue 1. Look at $\ker (A-I)$, it has dimension 2. $\endgroup$ – copper.hat Feb 5 '17 at 18:23
  • 1
    $\begingroup$ @JonGarrick: I'm not sure what you are asking, $\operatorname{dim}\ker(A-\lambda I)$ will be the number of linearly independent eigenvectors. $\endgroup$ – copper.hat Feb 5 '17 at 18:34
  • 1
    $\begingroup$ @copper.hat Thanks :) Just wanted that .. $\endgroup$ – Jon Garrick Feb 5 '17 at 18:36
2
$\begingroup$

Find the rank of $A-I$:

$$\begin{pmatrix} 0 && 3 && 2 \\ 0 && 3 && 2 \\ 0 && - 3 && -2 \end{pmatrix} $$

which is clearly $1=3-2$, so the size of the eigenspace is $2$-dimensional

$\endgroup$
  • $\begingroup$ Why take $A -I$ ? $\endgroup$ – Jon Garrick Feb 5 '17 at 18:28
  • 1
    $\begingroup$ @JonGarrick because the eigenvalue in question is $1$. If it were $2$ I would take $A-2I$ and if it were $\lambda$ I would look at $A-\lambda I$. $\endgroup$ – Adam Hughes Feb 5 '17 at 18:28
  • $\begingroup$ Okk !! Got now. One more thing, Is this a rule of thumb to find rank of $A−λI$ ? $\endgroup$ – Jon Garrick Feb 5 '17 at 18:30
  • 1
    $\begingroup$ @JonGarrick indeed this is more than that, it's essentially the definition of the co-dimension of the eigenspace for $\lambda$ (the "co" means you have to take the total dimension minus the rank, i.e. $3-1=2$). $\endgroup$ – Adam Hughes Feb 5 '17 at 18:32
  • 1
    $\begingroup$ Well, in general, find the rank of $A-\lambda I$. $\endgroup$ – Michael Hoppe Feb 5 '17 at 18:32

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.