1
$\begingroup$

I have two $9\times 9$ matrices $g_1, g_2$ where

g1 = {{-1, -(1/2), 0, 0, -(1/2), -(1/2), 0, 1/2, -(1/2)}, {0, 0, 0, 1, 0, 0, 0, 0, 0}, {-1, -(1/2), 3/2, 1, -(1/2), -1, -(3/2), 1, -(1/2)}, {1, 0, -(3/2), -1, 0, 1/2, 3/2, 1/2, 0}, {0, 0, 0, 0, 0, 0, 0, 0, 1}, {1/2, 0, 1/4, 1/2, -(1/2), -(1/4), -(3/4), 1/4, -(1/2)}, {0, 0, 0, 0, 0, 0, 0, 1, 0}, {-(1/2), 1/2, 3/4, 3/2, 0, -(5/4), -(5/4), 1/4, 0}, {0, 1/2, 0, -1, -(1/2), -(1/2), 1, -(1/2), 1/2}}

g2 = {{1/2, 1/2, 1/4, -(1/2), 0, 1/4, 1/4, -(1/4), 1}, {1/2, -(1/2), -(5/4), -(1/2), 1, 3/4, 3/4, 1/4, 0}, {0, 0, 1, 0, 0, 0, 0, 0, 0}, {0, 0, 0, 0, 0, 1, 0, 0, 0}, {0, 1/2, 1/2, 0, 1/2, 0, -(1/2), 0, -(1/2)}, {0, 0, 0, 1, 0, 0, 0, 0, 0}, {0, 0, 0, -1, 0, 1, 1, 0, 0}, {0, 0, 0, 0, 0, 0, 0, 1, 0}, {1/2, 0, 1/4, 1/2, -(1/2), -(1/4), -(3/4), 1/4, -(1/2)}}

I want to find four index invariants under the action of these two matrices, i.e. \begin{equation} (g_1)_a^{\,a_1}(g_1)_b^{\,b_1}(g_1)_c^{\,c_1}(g_1)_d^{\,d_1}T_{a_1 b_1 c_1 d_1}=T_{abcd} \end{equation} and \begin{equation} (g_2)_a^{\,a_1}(g_2)_b^{\,b_1}(g_2)_c^{\,c_1}(g_2)_d^{\,d_1}T_{a_1 b_1 c_1 d_1}=T_{abcd} \end{equation}

Now, let me try to explain what I was trying to do. I wrote these equations as \begin{equation} \big[(g_1)_a^{\,a_1}(g_1)_b^{\,b_1}(g_1)_c^{\,c_1}(g_1)_d^{\,d_1}-\delta_a^{\,a_1}\delta_b^{\,b_1}\delta_c^{\,c_1}\delta_d^{\,d_1}\big]T_{a_1b_1c_1d_1}=0 \end{equation} Then use Kronecker Product to make this problem as a problem of finding the null space of a matrix. i.e.

(KroneckerProduct[g1, g1, g1, g1]) - IdentityMatrix[9^4] //NullSpace

Then I move to see whether the linear combinations of these null vectors can be the null vectors of

(KroneckerProduct[g2, g2, g2, g2]) - IdentityMatrix[9^4] //NullSpace

If I can find one solution, then the problem is solved. All I need to do then is to translate the $9^4$ vector to a 4 rank tensor.

I tried this algorithm in the case of $g_1,g_2$ are two $5\times5$ matrices and it worked well. However, it seems it takes forever to run the first command line in $9\times 9$ case. Any suggestions for me to solve this problem?

For my use, I only need to solve the problem for the case where $T_{abcd}$ is totally symmetric. Perhaps I could write down a general ansatz for the totally symmetric four rank tensor and solve it. I will try to do it.

For curiosity, I'd like to understand the general case.

Improve this question
$\endgroup$
7
  • $\begingroup$ Can you show all your code, including definitions of g1 and g2, rather than just snippets? What does each $(g_1)^{a_1}_{a}$ mean? What is $T_{abcd}$? $\endgroup$
    – MarcoB
    Jan 19 at 16:34
  • $\begingroup$ @MarcoB: The OP is using abstract index notation. $\endgroup$
    – Michael Seifert
    Jan 19 at 16:36
  • 2
    $\begingroup$ It seems plausible to me that if $W \subseteq V$ is the subspace of $V$ invariant under the actions of $g_1$ and $g_2$, then the subspace of $V \otimes V \otimes V \otimes V$ that is invariant under $g_1$ and $g_2$ will simply be $W \otimes W \otimes W \otimes W$. If that's true, it simplifies your problem greatly, since you only have to run your code on an $n$-dimensional space instead of an $n^4$-dimensional space. Proving that would be off-topic here, though. You might consider posting this conjecture as a question to Mathematics instead (if you can't readily disprove it.) $\endgroup$
    – Michael Seifert
    Jan 19 at 16:53
  • 3
    $\begingroup$ I got Mathematica to complete your code in about 15 minutes of calculation time, using $g_1$ as the rotation matrix that takes $\hat{e}_1 \to \hat{e}_9$ and $g_2$ as the rotation matrix that takes $\hat{e}_1 \to \hat{e}_2$. The resulting nullspaces are 2997-dimensional, which suggests that my conjecture above is incorrect. $\endgroup$
    – Michael Seifert
    Jan 19 at 17:01
  • 1
    $\begingroup$ @user3257842 Thanks. I have updated the question $\endgroup$
    – Vayne
    Jan 19 at 23:43

0

Reset to default

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.