4
$\begingroup$

A well known result about the natural numbers $\mathbb{N}$ says that for any finite subset $A \subset \mathbb{N}$ there exists $R \ge 0$ such that if $n$ is in the subgroup of $\mathbb{Z}$ generated by $A$ and if $n \ge R$ then $n$ is in the semigroup generated by $A$.

Are there any references to a higher dimensional version of this result?

The version I want goes like this.

  • Take a finite subset $U$ of $\mathbb{N}^n$. Let $C_U$ be the smallest closed cone in $\mathbb{R}^n$ containing $U$, i.e. all non-negative real linear combinations of $U$. Let $G_U$ be the subgroup of $\mathbb{N}^n$ generated by $U$, i.e. all integer linear combinations. Let $S_U$ be the subsemigroup generated by $U$, i.e. all non-negative integer linear combinations. Then there exists $R>0$ such that for every $v \in G_U$, if the ball around $v$ of radius $R$ is contained in $C_U$ then $v \in S_U$.
$\endgroup$
6
  • 1
    $\begingroup$ I don't have any literature in front of me. The right buzzword is affine semigroups because that is what people who study them call them. $\endgroup$ Aug 2, 2012 at 19:58
  • $\begingroup$ emis.de/journals/SC/2002/6/pdf/smf_sem-cong_6_43-127.pdf seems to have some info on affine semigroups, the group they span and the polyhedral cone they span. $\endgroup$ Aug 2, 2012 at 20:10
  • 2
    $\begingroup$ A higher dimensional analogue of the result you mention in dimension 1 can be found as Exercise 7.15 of Miller and Sturmfels combinatorial commutative algebra book but it is not quite what you want, I think. Let $N=C_U\cap G_U$. Then $N$ is a finitely generated semigroup and there exists according to exercise 7.15 an element $a$ of $S_U$ with $a+N\subseteq S_U$. $\endgroup$ Aug 2, 2012 at 20:42
  • $\begingroup$ That would do it. Take $R > |a|$. If $B_R(v) \subset C_U$ then $v-a \in C_U \cap G_U = N$ so $v \in a+N \subset S_U$. Great! $\endgroup$
    – Lee Mosher
    Aug 2, 2012 at 21:07
  • $\begingroup$ Glad to be of help. Perhaps we should copy this into the answer box so that the software knows it is answered? $\endgroup$ Aug 3, 2012 at 0:38

1 Answer 1

6
$\begingroup$

This combines comments of myself and Lee Mosher.

Exercise 7.15 of Miller and Sturmfels combinatorial commutative algebra book proves the following. Let $N=C_U\cap G_U$. Then $N$ is a finitely generated semigroup and there exists according to exercise 7.15 an element $a$ of $S_U$ with $a+N\subseteq S_U$. Now take $R>|a|$. If $B_R(v)\subseteq C_U$ then $v−a\in C_U\cap G_U=N$ so $v\in a+N\subseteq S_U$.

$\endgroup$
1
  • $\begingroup$ By the way, my original purpose in asking this question was to provide a reference for Thurston's paper "Entropy in Dimension 1"; see Lemma 11.6 of that paper. The paper has now been published in "Frontiers in Complex Dynamics: In Celebration of John Milnor's 80th Birthday". $\endgroup$
    – Lee Mosher
    Jul 17, 2014 at 16:49

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.