« first day (1997 days earlier)      last day (1713 days later) » 

1:59 AM
Jan 23 at 6:58, by Martin Sleziak
Posting a dummy message to keep the room from getting frozen. Just in case it could be useful in the future.
 
 
10 hours later…
12:08 PM
Any ideas on determining whether there is a subgroup of $\text{GL}(2,\mathbb C)$ isomorphic to $D_4\times \mathbb Z_2$?
 
There are real matrices which correspond to symmetries of the square. So you have a subgroup of $GL(2,\mathbb R)$ which is isomorphic to $D_4$.
However, I am not really sure how to "add $\times\mathbb Z_2$".
 
Yeah, any dihedral group can be represented by 2-by-2 matrices. It is also trivial if we modify the 2 in the general linear group to 3.
 
12:42 PM
If we have such group of matrices, there must be matrices $A$, $B$, $C$ corresponding to $(r,0)$, $(s,0)$ and $(id,1)$.
For $A$ we know that $A^4=I$.
This implies that $A$ has to be diagonalizable, and we have $PAP^{-1}=D$ for some regular matrix $P$.
Since $X\mapsto PXP^{-1}$ is an automorphism of $GL(2,\mathbb C)$, we can w.l.o.g. assume that $A$ is diagonal, i.e., $A=\operatorname{diag}(d_1,d_2)$.
We know that $d_i^4=1$, so $d_i\in\{\pm1,\pm i\}$.
We know that $A^2\ne I$, which eliminates some possibilities.
Also a cannot be a scalar multiple of $I$, since then it would commute with each matrix. (And don't want it to commute with $B$.)
So as possibilities for $A$ we get $\operatorname{diag}(1,-1)$, $\operatorname{diag}(1,i)$, $\operatorname{diag}(1,-i)$ and the matrices which are obtained from these three by multiplying by $-1$ or $\pm i$.
Still, I am not sure whether we can get from this that there are no matrices $B$, $C$ with the required properties.
I think we get from $AC=CA$ that $C$ has to be diagonal, too. And again, we have only a few possibilities for $C$, the diagonal elements have to be $\pm1$ (since $C^2=I$).
Also $A^2\ne I$ eliminates some possibilities.
 
1:02 PM
I tried using Mathematica to solve the equations of the generating conditions of $D_4\times \mathbb Z_2$. It seems that there's no solution.
 
Also $B$ has to be diagonal, since $BC=CB$.
So all three matrices $A$, $B$, $C$ are w.l.o.g. diagonal. I'd guess that if we are working with diagonal matrices only, that should be enough information to show that this is impossible.
If we know that A, B, C are diagonal, then all elements of the group generated by these three matrices commute - which is contradiction, since the group $D_4\times\mathbb Z_2$ is not commutative.
Where does this problem come from?
Of course, it would be nicer to have some more elegant approach. (Without transforming the whole situation to working with diagonal matrices.)
 
My friend tried to find the least order of faithful complex representation of some small groups.
Well, he like collecting these strange results.
I appreciate your solution, especially the part of transforming the problem into diagonal situations. :)
 

« first day (1997 days earlier)      last day (1713 days later) »