1
$\begingroup$

Let $H$ be a multiplicatively written monoid with identity $1_H$. An atom of $H$ is an element $x \in H \setminus H^\times$ such that $a \ne xy$ for all $x, y \in H \setminus H^\times$, where $H^\times$ is the group of units of $H$. Accordingly, we define $\mathsf L_H(1_H) := \{0\} \subseteq \mathbf N$ and, for $x \in H \setminus \{1_H\}$, we take ${\sf L}_H(x)$ to be the set of all $k \in \mathbf N^+$ such that $x = a_1 \cdots a_k$ for some atoms $a_1, \ldots, a_k \in H$.

It follows by Proposition 1 in Question #269229 and Fekete's lemma that the function $$ \ell_H: H \to [0,\infty]: x \mapsto \lim_{n \to \infty} \frac{|\mathsf L_H(x^n)|}{n} $$ is well defined, insomuch as the limit in the above exists, and is either a non-negative real number or $\infty$. My question is as follows:

Q. Let $H$ be the multiplicative monoid of the ring of integers, $\mathbf Z_K$, of a number field $K$. Is it true that $\ell_H(H)$ is a compact subset of $\mathbf R$ (with the usual topology)?

Of course, $0 \in \ell_H(H)$, and it can be proved (and, as far as I can say, it is not so obvious) that, under the assumptions of this question, $\ell_H(H)$ is a bounded subset of $\mathbf R$, so the real point is whether or not $\ell_H(H)$ is closed.

Also, let me note that the answer is yes if $\mathbf Z_K$ is half-factorial (in particular, a UFD), namely, $|\mathsf L_H(x)| = 1$ for all $x \in H \setminus H^\times$: This is trivial, though it covers many non-trivial cases. In particular, it is well known (at least in some circles) that $H$ is half-factorial iff the class number of $K$ is $\le 2$, see [2, Theorem 1.7.3.5]: The result was first established in [1], though the term "half-factorial" was introduced only later in [3].

Bibliography

[1] L. Carlitz, A characterization of algebraic number fields with class number two, Proc. Amer. Math. Soc. 11 (1960), 391-392.

[2] A. Geroldinger and F. Halter-Koch's, Non-Unique Factorizations. Algebraic, Combinatorial and Analytic Theory, Pure Appl. Math. 278, Chapman & Hall/CRC, 2006.

[3] A. Zaks, Half factorial domains, Bull. Amer. Math. Soc. 82 (1976), 721-723.

$\endgroup$
4
  • $\begingroup$ Are the effects of obvious constructions known? For example, if you measure the disjoint union (or appropriate amalgam) of two monoids, do you get the max of the measures of each one? You might then get closure by taking the disjoint union of all of them. If not union, then product maybe? Gerhard "Maybe The Category Is Closed?" Paseman, 2017.05.15. $\endgroup$ Commented May 15, 2017 at 15:40
  • $\begingroup$ @GerhardPaseman: Many, many things are known, the basic reference for the cancellative, commutative case being Geroldinger and Halter-Koch's monograph. E.g., Proposition 1.4.5.1 in the book carries over to the following: If $(H_i)_{i \in I}$ is a family of Dedekind-finite monoids and $H:=\coprod_{i \in I} H_i$, then ${\sf L}_H(a)=\sum_{i \in I} {\sf L}_{H_i}(a_i)$ for every $a=(a_i)_{i \in I} \in H$. Does this answer your question, at least in part? (It's not so clear to me how you'd like to use this kind of results for the present problem.) $\endgroup$ Commented May 15, 2017 at 17:37
  • $\begingroup$ The idea would be to show that limits are preserved in passing from the collection of monoids to the real numbers. If one has a sub collection of monoids M each with real number valuation v (v is easier for me to read than l), then a sub sequence of the v's converging to a limit w may point to a related sequence of monoids M. Hopefully one can build a monoid N out of the M's whose valuation is w. Gerhard "Then This Would Show Closure" Paseman, 2017.05.15. $\endgroup$ Commented May 15, 2017 at 17:40
  • $\begingroup$ So, I seem to understand that you have in mind some kind of convergence in a (hypothetical) "space of monoids", with the convergence being compatible, in some sense, with the "structure behind my question". If so, I'm skeptical that something like this can be made to work, but I would be more than happy to be proved wrong! $\endgroup$ Commented May 15, 2017 at 17:50

0

You must log in to answer this question.