1

Suppose $Sub(G)$ is the set of all subgroups of a group $G$. What is the asymptotic of $|Sub(S_n)|$ with $n \to \infty$?

One can get the following bounds on $|Sub(S_n)|$:

$$2^{2^{\langle \frac{n}{2} \rangle}} - 1 \leq |Sub(S_n)| \leq 2^{n!}$$

Proof:

On one hand:

$|Sub(S_n)| \leq |P(S_n)| = 2^{n!}$

On the other hand:

Suppose $k = \langle \frac{n}{2} \rangle$. Then $S_n$ has a subgroup isomorphic to $Sym(\{1, … , k\} \times \{0, 1\})$. For all $A \subset \{1, … , k\} \times \{0, 1\})$ define $\sigma_A: (p, q) \mapsto (p, q + I_A(p) \mod 2)$. It is not hard to see, that $\{\sigma_A| A \subset \{1, … , k\}\} \cong C_2^{2^k}$. And it is known, that $C_2^m$ always has $2^m - 1$ subgroups isomorphic to $C_2$.

The only information on $|Sub(S_n)|$ I was able to find online was the list of exact values of $|Sym(S_n)|$ for $n \leq 18$, calculated by Derek Holt in «Enumerating subgroups of symmetric groups».

Chain Markov
  • 15,564
  • 6
  • 36
  • 116
  • I think the sequence you're after has an OEIS entry here. I see no asymptotics information, though. – Arthur Jan 14 '20 at 18:33
  • 1
    This question has been closed, but it happens to be a topic of current interest, because a number of people are working hard to prove the conjecture that Pyber's lower bound of $2^{(\frac{1}{16}+o(1))n^2}$ on the number of subgroups of $S_n$ is tight. – Derek Holt Jan 14 '20 at 19:52

0 Answers0