Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

Linear Algebra Matrix Eigenvalues

  1. Oct 23, 2011 #1
    1. The problem statement, all variables and given/known data

    The matrix A has 3 distinct eigenvalues t1< t2< t3. Let vi be the unique eigenvector associated to ti with a 1 as its first nonzero component. Let

    D= [t1 0 0
    0 t2 0
    0 0 t3]

    and P= [v1|v2|v3] so that the ith column of P is the eigenvector vi associated to ti

    7 2 -8
    0 1 0
    4 2 -5

    a) Find D
    b) Find P
    c) Find P-1
    2. Relevant equations

    3. The attempt at a solution
    My thought to find D was to find the characteristic equation of A which I found to be (t-3)(t+1)=0 so the eigenvalues would be t=-3, t=1 I then plugged in these values into the matrix D so D became
    3 0 0
    0 -1 0
    0 0 0

    but it was counting it wrong. What did I mess up on?
  2. jcsd
  3. Oct 23, 2011 #2

    I like Serena

    User Avatar
    Homework Helper

    Hi lina29 :smile:

    I'm afraid your characteristic equation is a little off.
    How did you get it?

    And your eigenvalues are not the solution of the equation you have.
    They are off by a minus sign.

    You're supposed to get 3 distinct eigenvalues.
    (They should be 3, -1, 1.)

    Oh, and according to your problem statement your eigenvalues should be sorted low to high on your diagonal.
  4. Oct 23, 2011 #3
    To find the determinant I did aei+bfg+cdh-ceg-bdi-afh. So I got (t-7)(t-1)(t+5)+0+0-(8)(t-1)(-4)-0-0. I simplified that to (t-7)(t-1)(t+5)-(-32)(t-1). My mistake was that I accidently crossed out the t-1 on both sides. From there how would I get to P?
  5. Oct 23, 2011 #4

    I like Serena

    User Avatar
    Homework Helper

    Good! :smile:

    Find the eigenvectors v1, v2, v3.
    P is the matrix with these eigenvectors as columns.
  6. Oct 23, 2011 #5
    So the eigenvectors would be -1, 1, and 3. Then the matrix P would be
    is that it?
  7. Oct 23, 2011 #6
    or I guess that wouldn't be right since I need to find P inverse so P would have to be a square matrix
  8. Oct 23, 2011 #7

    I like Serena

    User Avatar
    Homework Helper

    No. There's a difference between eigenvalues and eigenvectors.
    t1=-1 is your first eigenvalue.
    v1 is the vector belonging to this eigenvalue.

    To find v1 you need to solve A v1 = t1 v1.

    Note that any scalar multiple of v1 also satisfies the equation, but its direction is unique.

    Edit: I just saw that your problem statement says that v1 should have "1 as its first nonzero component".
    Last edited: Oct 23, 2011
  9. Oct 23, 2011 #8
    ohh I got it. Thank you so much!
  10. Oct 23, 2011 #9

    I like Serena

    User Avatar
    Homework Helper

Share this great discussion with others via Reddit, Google+, Twitter, or Facebook