Existence of interesting infinite subsets of the rationals whose elements are not arbitrarily close to some reals

38 Views Asked by At

It is well known that for every real number we can find a rational number arbitrarily close to it. In other words:

Let $a\in\mathbb{R}$. For each $\epsilon\gt 0$, there exist $p,q\in\mathbb{Z}$ such that $\left|a-\frac{p}{q}\right|\lt\epsilon$.

But can we let $p$ belong to some infinite subset of $\mathbb{Z}$ so that this is not true? In other words:

Does there exist an infinite subset $S\subset\mathbb{Z}$ such that there exist $a\in\mathbb{R}$ and $c>0$ so that for every $p\in S$ and $q\in\mathbb{Z}$, we have $c\lt\left|a-\frac{p}{q}\right|$?

What if we say $q$ is also in $S$?

If such subsets exist I’d naturally want to know the largest one.

1

There are 1 best solutions below

0
On BEST ANSWER

With $q \in S$ we can say even more. We can find an infinite $S$ such that for any given $c \gt 0$ we can find $a$ such that $\forall p,q \in S, |a-\frac pq| \gt c$ Just take $S$ to be the powers of $2$. Then $\frac pq$ is also a power of $2$, so given $c$ we just take $a$ large enough that powers of $2$ are far enough apart.

If we don't insist that $q \in S$ this fails. As $S$ is infinite there are arbitrarily large elements. The distance between $\frac pq$ and $\frac p{q+1}$ is about $\frac p{q^2}$, which is about $\frac aq$ and can be made arbitrarily small.