4
$\begingroup$

Let $p$ be prime and let $T^p$ be the $p$-torus and $\mathbb{Z}/p$ the cyclic group of order $p$ generated by $(12\ldots p)$. Consider the semidirect product $T^p\rtimes\mathbb{Z}/p$ with the natural action.

I would like to know if the representation ring of this semidirect product is known. Alternatively, what is known about the cohomology of the classifying space $H^*(B(T^p\rtimes\mathbb{Z}/p);G)$ for $G$ either $\mathbb{Q}$ or $\mathbb{Z}$?

$\endgroup$

2 Answers 2

3
$\begingroup$

For any finite dimensional representation, you can restrict to $T^p$ to get a decomposition into a direct sum of one-dimensional characters, which are given by $p$-tuples of integers. If a given character is diagonal (i.e., of the form $(n,\ldots,n)$), then it is a one-dimensional representation of $\mathbb{Z}/p$, i.e., characterized by a $p$-th root of unity. If a character is non-diagonal, then it is contained in the $p$-dimensional representation containing all cyclic permutations of that character, so the cyclic permutation class characterizes the non-diagonal irreducible representations.

Thus, we can describe the finite dimensional representations as finitely supported sheaves on the stack quotient $[\mathbb{Z}^p_{\mathbb{C}}/(\mathbb{Z}/p)]$. Tensor product is given by lifting to the cover and convolving (i.e., the cyclic symmetry on characters of $T^p$ is preserved).

$\endgroup$
2
  • $\begingroup$ Thank you. I was also thinking about restriction to $T^p$, perhaps using induction. I'm looking for a description in terms of generators and relations. Maybe some spectral sequence argument? Bockstein? $\endgroup$ Feb 24, 2014 at 21:19
  • $\begingroup$ What does the subscript $\mathbb C$ in $\mathbb Z_{\mathbb C}^p$ mean? $\endgroup$
    – LSpice
    May 31, 2023 at 19:44
0
$\begingroup$

As to the cohomology, here is an answer, which is probably an "over-kill". Probably you can find an answer in much older literature like Borel's.

Since $BT$ has torsion-free homology, COROLLARY 4.9 of JOHN ROBERT HUNTON The complex oriented cohomology of extended powers Annales de l’institut Fourier, tome 48, no 2 (1998), p. 517-534 applies.

With rational coefficients (you can definitely find this in Borel) you get just the invariants of $Q[x_1,\ldots x_p]$ under the action of cyclic groups. When $p=2$, the cyclic group being same as the permutation group, you get the ring of symmetric polynomials $Q[c_1,c_2]$ this way (which happens to be the cohomology of $BU(2)$, and of course, in this case your semi-direct product is the Weyl group of $U(2)$).

Things get a bit more complicated for integer coefficients, but you can read it off from the above corollary.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.