# Asymptotics for the number of abelian groups of order at most $x.$

The number of abelian groups of order $n$ (call it $a(n)$ is a studied subject (see http://oeis.org/A000688), but I can't seem to find any asymptotic results. Obviously, there is no asymptotic for $a(n)$ itself (it is much too irregular), but there should an asymptotic for $\sum_{n\leq x} a(n),$ but I can't seem to find a reference.

Roberto answered the above, but another question is whether one has any distributional results (how high are the maxima, is there a limiting distribution, etc).

• Out of curiosity, would there be a nice formula for counting abelian groups $G$ with coefficient $1/|{\rm Aut}(G)|$? – Lev Borisov Feb 12 '16 at 20:42
• @quid that's a great reference, thanks! – Igor Rivin Feb 12 '16 at 22:11
• @LevBorisov If you look at quid's reference, the first lemma tells us (I believe) that the number of groups mod automorphism is again a multiplicative function of $n,$ so there should be nice asymptotics, but I think the summatory function should grow like $\log n.$ – Igor Rivin Feb 12 '16 at 22:15
• @GregMartin true, but I think there is enough information to get the asymptotic (for square fee numbers, you basically get the sum of $1/n,$ which makes the logarithmic estimate plausible). – Igor Rivin Feb 13 '16 at 17:41
• Order of magnitude, certainly, probably both upper and lower bounds. Might be tough to get the exact leading constant though. – Greg Martin Feb 13 '16 at 22:42

The problem was studied quite a bit; a complete summary will be complicated to give (in any case I cannot). A standard reference for classical results on this is A. Ivić "The Riemann Zeta-function: Theory and Applications" (1985); it seems there is a recent Dover edition. Ivić has various papers on this problem, too. The book has a chapter on this subject (14.5 Non-isomorphic abelian groups of a given order). At the moment I cannot recall what exactly is in there.

Some information related to this type of problem:

The correct asymptotics for $\sum_{n\in \mathbb{N}} a(n)$ were obtained by Erdős and Szekeres (1934), they proved: $$\sum_{n\le x} a(n) \sim A x + O(\sqrt{x})$$ with $A= \prod_{n \ge 2}\zeta(n)$.

By now there are more precise results known. For example the paper by Kendall and Rankin (1947), mentioned in another answer, gives $Ax + B x^{1/2} + O(x^{1/3} \log x )$, and developments continued, like the result mentioned by Roberto Pignatelli.

The to my knowledge latest improvement there is by Sargos and Wu (2000) $$Ax + Bx^{1/2} + Cx^{1/3} + O(x^{55/219} (\log x)^7 )$$ The first to get the third term of the main term was Richert (1952). For the historical development of the error see the introduction of a paper by Calderón (2003).

Related questions were also studied. For example, Kendall and Rankin (1947) showed that the "local densities" of $a(n)$ exist, that is $\sum_{n \le x, \, a(n)=k}1 \sim d_k x$ (good error terms are known). There are also estimates "in short intervals" so on $\sum_{ x \le n \le y, \, a(n)=k}1$.

For two recent papers on this see:

Emre Alkan, On the enumeration of finite abelian and solvable groups, J. Number Theory 101 (2003), no. 2, 404--423.

Ekkehard Krätzel, The distribution of values of the enumerating function of finite, non-isomorphic abelian groups in short intervals, Arch. Math. (Basel) 91 (2008), no. 6, 518--525.

The introductions and references there will lead to various additional paper. (The first is freely accessible.)

References

C. Calderón: Asymptotic estimates on finite abelian groups. Publications De L’institut Mathematique, Nouvelle série, 74, 57-70 (2003).

P. Erdős, G. Szekeres: Über die Anzahl der Abelschen Gruppen gegebener Ordnung und über ein verwandtes zahlentheoretisches Problem (in German), Acta Litt. Sci. Szeged 7 (1934), 95--102; Zentralblatt 10,294. Free PDF

D. G. Kendall and R. A. Rankin, On the number of Abelian groups of a given order, Quart. J. Math., Oxford Ser. 18 (1947), 197--208.

Hans-Egon Richert, Über die Anzahl Abelscher Gruppen gegebener Ordnung. I, Math. Z. 56 (1952), 21--32.

P. Sargos and J. Wu, Multiple exponential sums with monomials and their applications in number theory, Acta Math. Hungar. 87 (2000), no. 4, 333--354.

One reference where the asymptotic result you are asking for was first established (I think), as well as some reasonable growth estimates for $a_n$, is

D.G.Kendall and R.A.Rankin, "On the number of Abelian groups of a given order", Quart. J. Math., Oxford Ser. 18 (1947), 197–208.

According to http://mathworld.wolfram.com/AbelianGroup.html it is a theorem of Srinivasan (1973), click on the link for details.

• There are earlier results giving asymptotics as far as I know thus to say "it is a theorem of Srinivasan (1973)" is a bit misleading. – user9072 Feb 12 '16 at 13:49