Skip to main content
11 events
when toggle format what by license comment
Jul 30, 2015 at 0:31 vote accept Daniel Soltész
Jul 11, 2015 at 1:25 answer added Igor Rivin timeline score: 7
Jul 11, 2015 at 0:25 comment added Daniel Soltész @EricWofsey Well I can't prove it I just read it at OEIS. Here is the proof, (jstor.org/stable/2946623?seq=8#page_scan_tab_contents) I think the lower bound suggests that we look at the cases where $k$ is a power of two. (Maybe this effect is the reason Gerhard Paseman had two in his mind.)
Jul 11, 2015 at 0:22 comment added Alex R. @EricWofsey: It seems like a difficult question actually, see the top answer here: math.stackexchange.com/questions/76176/…
Jul 10, 2015 at 23:44 comment added Eric Wofsey How do you prove there are $c^{n^2}$ subgroups? It seems likely to me that by inspecting such a proof it would not be too hard to find the $k$ that maximizes $a_k$.
Jul 10, 2015 at 23:25 comment added Daniel Soltész @GerhardPaseman Well I don't have any OEIS sequence for a_k, but I have one for the total number of subgroups (oeis.org/A005432). But k can not be small by a counting argument. The number of $2$ element subgroups is trivially at most $n!$, but the total number of subgroups has order of magnitude $c^{n^2}$, so the most frequent order by the pidgeon-hole must be at least $c^{n^2}/n!$ which is larger.
Jul 10, 2015 at 23:03 comment added Gerhard Paseman I would guess k=2. Do you have any data or OEIS sequences? Gerhard "Yet Another Use For Mode" Paseman, 2015.07.10
Jul 10, 2015 at 23:03 history edited Daniel Soltész CC BY-SA 3.0
added 110 characters in body
Jul 10, 2015 at 23:00 comment added Daniel Soltész In terms of number of subgroups. I will edit the question to make this obvious.
Jul 10, 2015 at 22:59 comment added Stefan Kohl "Most frequent" in terms of number of conjugacy classes of subgroups, or really in terms of number of subgroups?
Jul 10, 2015 at 22:56 history asked Daniel Soltész CC BY-SA 3.0