1. Limited time only! Sign up for a free 30min personal tutor trial with Chegg Tutors
    Dismiss Notice
Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

Homework Help: Jordan Forms

  1. Dec 14, 2009 #1
    1. The problem statement, all variables and given/known data
    find all Jordan forms of 8x8 matrices given the minimal polynomial x^2*(x-1)^3

    2. Relevant equations

    3. The attempt at a solution

    The roots are clearly 0,1 and 0 has degree 2 while 1 has degree 3. The forms would be made up of the blocks [0,0;1,0] corresponding to 0 and [1,0,0;1,1,0;0,1,1] corresponding to 1.

    So all possible 8x8 forms would be different combinations of the blocks such that they are always both included at least once and dimension 1 blocks being either of the roots such that the overall dimension of the blocks is 8.

    -I am not convinced this is the solution because of the geometric multiplicities of 1 and 0 would affect the entries next to the main diagonals...I'm just not sure how if at all.
  2. jcsd
  3. Dec 15, 2009 #2


    User Avatar
    Science Advisor

    Every such 8 by 8 matrix will have 2 "0"s and 3 "1"s on the diagonal. The "Jordan" form may or may not have "1" above each number on the diagonal. For the "0"s, then, you can have either
    [tex]\begin{array}{cc}0 & 0 \\ 0 & 0\end{array}[/tex]
    [tex]\begin{array}{cc}0 & 1 \\ 0 & 0\end{array}[/tex]

    For the "1"s there are 4 possiblilties.
  4. Dec 16, 2009 #3
    sorry I should have made this a bit more clear in the question, but can there exist a Jordan block (given the minimal polynomial above) with 1's on the main diagonal of dimension 2 such that it satisfies the 8x8 dimension req? i.e. assuming we have the dim=3 blocks obtained from 1 and the dim=2 blocks obtained from 0, can we have a dim=2 block with 1's on the main diagonal, and an appropriate number of dim=1 blocks being either 0,1 to satisfy the 8x8 dim?
  5. Dec 17, 2009 #4


    User Avatar
    Science Advisor

    Yes, it is possible, that "1" be an eigenvalue of algebraic multiplicity 3 and geometric multiplicity 2 (or any positive integer less than or equal to 3). One possiblity would be
    [tex]\begin{bmatrix}1 & 1 & 0 \\ 0 & 1 & 1 \\0 & 0 & 1\end{bmatrix}[/tex]
Share this great discussion with others via Reddit, Google+, Twitter, or Facebook