Results 1 to 2 of 2

Math Help - Parametrizing problem

  1. #1
    Member Mauritzvdworm's Avatar
    Joined
    Aug 2009
    From
    Pretoria
    Posts
    122

    Parametrizing problem

    I am wondering if it is possible to parametrize the space of all unitary 3 by 3 matrices with complex entries.

    U=\left(\begin{array}{ccc}a&b&c\\d&e&f\\g&h&j\end{  array}\right) with U being any unitary 3 by 3 matrix.

    We know that the column vectors have to be orthonormal such that we have the unitary property U^*U=1.

    So let a=a_1+ia_2, ~ d=d_1+id_2, ~ g=g_1+ig_2 then we can write

    a^2_1+a^2_2+d^2_1+d^2_d+g^2_1+g^2_2=1

    this is just a sphere in \matbb{R} and we can easily parametrize it as follows

    \begin{array}{ccc}<br />
g_2&=&\cos\phi_1\\<br />
g_1&=&\sin\phi_1\cos\phi_2\\<br />
d_2&=&\sin\phi_1\sin\phi_2\cos\phi3\\<br />
d_1&=&\sin\phi_1\sin\phi_2\sin\phi_3\cos\phi_4\\<br />
a_2&=&\sin\phi_1\sin\phi_2\sin\phi_3\sin\phi_4\cos  \phi_5\\<br />
a_1&=&\sin\phi_1\sin\phi_2\sin\phi_3\sin\phi_4\sin  \phi_5<br />
\end{array}

    using these we can find the corresponding parametrization of the first column. My question is how would we parametrize the remainder of the matrix?

    If v_1=\left(\begin{array}{c}a\\d\\g\end{array}\right  ) and v_2,v_3 are respectively the second and third columns of U, then we know that

    v_i\cdot v_j=0 if i\neq j and 1 otherwise

    we would have to consider the space orthogonal to v_1, this space must be some two dimensional plane. The vectors v_2,v_3 must then lien in this plane.

    I can solve this problem for the case of two by two matrices with complex entries and wanted to see if it was at all possible for higher dimensions, however the parametrization is a little tricky.
    Follow Math Help Forum on Facebook and Google+

  2. #2
    Member Mauritzvdworm's Avatar
    Joined
    Aug 2009
    From
    Pretoria
    Posts
    122
    If we can find one vector which will always be orthogonal to v_1, we can just find the crossed product of that vector with v_1 to determine the last vector.

    However, we have to take rotations into account for the parametrization, and I'm not sure as how to go about solving this problem.
    Follow Math Help Forum on Facebook and Google+

Similar Math Help Forum Discussions

  1. Re-parametrizing the curve
    Posted in the Calculus Forum
    Replies: 1
    Last Post: October 12th 2011, 09:19 PM
  2. Parametrizing
    Posted in the Differential Geometry Forum
    Replies: 2
    Last Post: March 14th 2011, 01:12 PM
  3. Parametrizing a line segment
    Posted in the Calculus Forum
    Replies: 1
    Last Post: September 7th 2009, 05:46 PM
  4. Question About Parametrizing in 3 Dimensions
    Posted in the Calculus Forum
    Replies: 1
    Last Post: June 20th 2009, 12:52 PM
  5. Parametrizing the Intersection of Surfaces
    Posted in the Calculus Forum
    Replies: 3
    Last Post: May 9th 2008, 10:33 PM

Search Tags


/mathhelpforum @mathhelpforum