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

Trace and Determinant

  1. Jan 29, 2009 #1
    can anyone give me an hint to proof the relation

    det(I + x A) = 1 + x tr A + O(x^2)

    thank you.
  2. jcsd
  3. Jan 29, 2009 #2

    Ben Niehoff

    User Avatar
    Science Advisor
    Gold Member

    Write out the general form of (I+xA). The determinant must be some polynomial in x, right? So, carefully pick out the constant- and linear-order terms of this polynomial by choosing the appropriate factors in the determinant.
  4. Jan 29, 2009 #3
    ok that's the idea behind....but what about some explicit calculations? Are them possible? I thought i can use the general determinant expansion in terms of the total antisymmetric tensor which, for a nxn matrix A must be something like

    det A = \frac{1}{n!} \sum_{i_1, i_2, ... , i_n = 1}^{n} \epsilon_{i_1\;i_2\;...\;i_n} A_{1\;i_1} A_{2\;i_2} A_{3\;i_3} ... A_{n\;i_n}

    the expression of the fact that the determinant is the average of all possible products of the elements picked one per column keeping them in different rows, thanks to totally antisymmetric tensor.

    Any idea to be a bit formal?
  5. Jan 29, 2009 #4

    Ben Niehoff

    User Avatar
    Science Advisor
    Gold Member

    There should be no 1/n! in there. The determinant is the anti-symmetrized sum of all the products; not the average.

    If you want to do the proof formally, the simplest way would be to diagonalize A. Or in general, bring A to upper-triangular form by a similarity transformation. Note that the identity matrix is invariant under similarity transformations. The determinant and trace are also invariant.
Share this great discussion with others via Reddit, Google+, Twitter, or Facebook