Skip to main content
8 events
when toggle format what by license comment
Aug 18, 2022 at 6:50 answer added Maciej Ulas timeline score: 1
Aug 17, 2022 at 12:40 comment added Katy Thanks for the explanation
Aug 17, 2022 at 12:35 comment added Olivier Bégassat My bad, I forgot some minus signs... The eigenvalues are $(a+c) + (b+d)$, $(a+c) - (b+d)$, $(a-c) -i(b-d)$ and $(a-c) + i(b-d)$ so you'll need ($c = a \pm 1$ and $b=d$) or ($c = a$ and $d=b\pm 1$). In the first case you also need $\pm 1 + 2a + 2b = \pm 1$ and $\pm 1 + 2a - 2b = \pm 1$ so that it seems only a few possiibilities for a and b are allowed. Something analoguous holds in the second case. There are thus only finitely many possibilities.
Aug 17, 2022 at 12:16 comment added Olivier Bégassat As far as I can tell this should work because the complex eigenvalues of a 4x4 circulant matrix with first row $(a,b,c,d)$ are $(a+c) + (b+d)$, $(a+c) - (b+d)$, $(a+c) +i(b+d)$ and $(a+c) - i(b+d)$ which will all be in $\{\pm1, \pm i\}$ unless I'm mistaken.
Aug 17, 2022 at 12:05 comment added Katy I tried it for k=2 and n=3 it does not work..the determinant is not 1 or _1 so it is not invertible in Z?Am I wrong?
Aug 17, 2022 at 11:58 comment added Olivier Bégassat It seems that there are four infinite families: those with first row $(n\pm1,k,-n,-k)$ and those with first row $(n,k\pm1,-n,-k)$, with $n,k\in\Bbb{Z}$.
S Aug 17, 2022 at 11:48 review First questions
Aug 17, 2022 at 14:39
S Aug 17, 2022 at 11:48 history asked Katy CC BY-SA 4.0