1. Not finding help here? Sign up for a free 30min 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!

Tensor problem

  1. Oct 30, 2007 #1
    1. The problem statement, all variables and given/known data

    Let {e1, e2, e3} be a basis for vector space V. Show that the rank 2 tensor g defined by g=2E1*E2 + 2E2*E1+E1*E3+E3*E1 (where Ei are dual vectors and * is the tensor product) is symmetric and nondegenerate. Caculate g inverse.

    2. Relevant equations

    Um. lots of tensor stuff. I have the properties of a tensor, that's about it.

    3. The attempt at a solution

    I am wondering how I would show that it is non-degenerate. I know that if a bilinear form B is non-degenerate then: ker[Bv1]={x|B(x,y)=0 for all y} and ker[Bv2]={y|B(x,y)=0 for all x}.So we have to show that the kernels of the two vector spaces generating the bilinear form are 0 for all y and all x, respectively. But how on earth is this done, especially when all I know is that we are taking two vectors from the vector space V for our tensor?

    I don't get it! I CAN show that it is symmetric, I think, by evaluating g on the basis vectors one at a time with tensor properties (is that the right thing to do?) So to show symmetric I just took g(e1), g(e2), g(e3).

    I haven't the foggiest how to find the inverse components.
  2. jcsd
  3. Oct 30, 2007 #2


    User Avatar
    Science Advisor

    It seems to me that your problem is equivalent to showing that the matrix

    0 2 0
    2 0 1
    0 1 0

    is symmetric, and does not have a zero eigenvalue.
  4. Oct 31, 2007 #3
    aha, I have figured it out. There was a typo in the problem, which I resolved with the instructor. THe problem was that it was symmetric, but not non-degenerate (the determinant is zero. This is a problem to show non-degeneracy!) Once the problem is fixed, I guess the answer is to write out the matrix for it, show symmetric, show invertibility (this should show us non-degeneracy?), get the inverse components from that.

Know someone interested in this topic? Share this thread via Reddit, Google+, Twitter, or Facebook

Have something to add?

Similar Discussions: Tensor problem
  1. A tensor problem (Replies: 2)

  2. Fluids/tensors problem (Replies: 3)