7

A set $S\subset \mathbb{Z}$ is Diophantine if there is an integer polynomial $P(n, \bar{m})$ such that$$n\in S \iff (\exists \bar{m} \in \mathbb{Z}^{k})(P(n,\bar{m})=0).$$A set $S\subset \mathbb{Z}$ is bounded Diophantine if there is an integer $N$ and an integer polynomial $P(n, \bar{m})$ such that for any $n\in \mathbb{Z}$ there is at most $N$ vectors $\bar{m}\in\mathbb{Z}^k$ such that $P(n, \bar{m})=0$ and$$n\in S \iff (\exists \bar{m} \in \mathbb{Z}^{k})(P(n,\bar{m})=0).$$A set $S\subset \mathbb{Z}$ is uniquely Diophantine if there is an integer polynomial $P(n, \bar{m})$ such that for any $n\in \mathbb{Z}$ there is at most one $\bar{m}\in\mathbb{Z}^k$ such that $P(n, \bar{m})=0$ and$$n\in S \iff (\exists \bar{m} \in \mathbb{Z}^{k})(P(n,\bar{m})=0).$$It's a theorem that Diophantine sets are precisely the recursively enumerable ones.

What can be said about the bounded Diophantine and the uniquely Diophantine sets? Are they the same class of sets? Are non-negative integers bounded Diophantine?

Apjoo
  • 79
  • Related (to $N=2$ case of your question). – Wojowu May 03 '21 at 13:34
  • I think that at best you would expect bounded Diophantine sets to be d.c.e. sets, i.e., a set of the form $A \setminus B$ where $A$ and $B$ are computably enumerable. – James Hanson May 03 '21 at 16:15
  • @JamesHanson isn't every recursively enumerable set d.c.e. (by taking $B=\emptyset$)? Then what is the advantage? – Apjoo May 04 '21 at 09:28
  • @Apjoo I'm just saying that you won't get the same characterization in one direction. It may be that every c.e. set is bounded Diophantine, but it's probably not true that every bounded Diophantine set is c.e. – James Hanson May 04 '21 at 16:00
  • @JamesHanson I see. Your comment made me realize I wanted a different definition. – Apjoo May 04 '21 at 16:14
  • 1
    Would 'boundedly diophantine' maybe be a better term? Bounded diophantine sounds like bounded and diophantine. – Arno Fehm May 04 '21 at 18:42
  • 1
    These questions have been around for quite some time already. I believe that Matiyasevich was the first to consider them. You will find some background in the introduction of https://arxiv.org/abs/1706.03302 – Pasten May 04 '21 at 19:11
  • So there it's called 'finite-fold diophantine'... – Arno Fehm May 04 '21 at 19:53

0 Answers0