88

Let $f(n)$ be the number of groups of order $n$ up to isomorphism. We want to prove that: $$f(a) \cdot f(b) \leq f(a \cdot b)$$ for all nonnegative integers $a$ and $b$. Our progress:

  1. If $a \cdot b \leq 8191$, then our conjecture holds.

  2. If $a + b \leq 848$, then our conjecture holds.

  3. If $\gcd(a, b) = 1$, then our conjecture holds.

  4. If $a \cdot b$ has less than 5 primes in its integer factorization, then our conjecture holds.

  5. If $f(a \cdot b) \leq 3$, then our conjecture holds.

  6. Let $c$ and $d$ nonnegative integers. Let $p$ a prime number. If $c + d \leq 9$ or $c=d$ or $c\cdot(5\cdot c + 24) + d\cdot(5\cdot d + 24) \leq 17 \cdot c \cdot d + 9$ then $$f(p^{c}) \cdot f(p^{d}) \leq f(p^{c + d})$$

Methods:

A lower bound of $f(2048)$ was used. What are the tightest known bounds for the number of groups of order $2048$?

For all positive integers $a$ and $b$: $$f(a) \leq f(a \cdot b)$$

  1. If $a \cdot b = 0$, then our conjecture holds. If $1\leq a \cdot b \leq 8191$: 37592 tests in an algorithm.

  2. 180625 tests in an algorithm. If $f(65274) = 36$, then $36 \leq f(130548)$.

  3. Thanks @testaccount.

  4. Let $p$, $q$, $r$, $s$ distinct prime numbers. Study of 11 cases for $a \cdot b$: $p$, $p^{2}$, $p\cdot q$, $p^{3}$, $p^{2}\cdot q$, $p \cdot q \cdot r$, $p^{4}$, $p^{3} \cdot q$, $p^{2} \cdot q^{2}$, $p^{2} \cdot q \cdot r$, $p \cdot q \cdot r \cdot s$.

  5. Our conjecture is valid for all squarefree $a \cdot b$.

  6. https://www.math.auckland.ac.nz/~obrien/research/gnu.pdf

Let $p$ a prime number and $m$ a positive integer.

$$p^{\frac{2}{27}m^{2}(m-6)} \leq f(p^{m}) \leq p^{\frac{1}{6}(m^{3}-m)}$$

Source:

https://groups.quendi.de/

Known bounds for the number of groups of a given order.

https://arxiv.org/abs/1702.02616

  • 20
    @WhileIAm It's not so simple: $G_1 \times H_1$ could be isomorphic to $G_2 \times H_2$ although $G_1$ is not isomorphic to $G_2$ and $H_1$ not isomorphic to $H_2$. Most notably, $G \times H$ is isomorphic to $H \times G$. – Robert Israel Apr 12 '24 at 03:16
  • 8
    If, after a few days, no one has solved the question here, you might consider posting it to MathOverflow (making sure to link each post to the one on the other site). – Gerry Myerson Apr 13 '24 at 00:43
  • 6
    The function $f$ has often been named ${\rm gnu}$, meaning "group number". – Derek Holt Apr 16 '24 at 08:08
  • 5
    A nice article on the gnu function is here: https://www.math.auckland.ac.nz/~obrien/research/gnu.pdf – KenWSmith Apr 16 '24 at 19:23
  • This article above is really nice, and should definitely help you with this problem. – SomeCallMeTim Apr 19 '24 at 10:41
  • The conjecture was tested and found valid for a.b<50000 except the unknown values running 503330 tests on this database : https://github.com/olexandr-konovalov/gnu/blob/master/data/gnu50000.txt Of course the unknown values are the most interesting ones. – A. Random Apr 20 '24 at 12:32
  • Let $\mathcal{A}$ and $\mathcal{B}$ be representatives for the isomorphism classes of groups of orders $a$ and $b$, respectively. Maybe the set of all possible semidirect products ${A \rtimes B : A \in \mathcal{A}, B \in \mathcal{B}}$ contains at least $|\mathcal{A}| \cdot |\mathcal{B}|$ distinct groups, up to isomorphism. This would prove that $f(a)f(b) \leq f(ab)$. – testaccount Apr 20 '24 at 12:57
  • Can you prove this when $a = p^2$ for a prime $p$? Surely $f(p^2 b) \geq 2 f(b)$? – testaccount Apr 20 '24 at 13:11
  • @testaccount, I'm sorry. I don't know how to prove when $a=p^{2}$. – Jorge Rael Apr 21 '24 at 11:43
  • If you can't see the article by @KenWSmith, here is an internet archive link: https://web.archive.org/web/20240726071313/https://www.math.auckland.ac.nz/~obrien/research/gnu.pdf – Robin Apr 18 '25 at 12:46
  • Because of the similar structure maybe you could use some of the tricks shown on the following Youtube video – Joako Jun 19 '25 at 06:24

0 Answers0