[Math] suprema and infima of sets without proof – Abbott 1.3.8

proof-verificationreal-analysissupremum-and-infimum

This problem comes from Understanding Analysis by Abbott.

1.3.8 Compute, without proofs, the suprema and infima (if they exist) of the following sets:

(a) $ \{ m/n : m,n \in \mathbb{N} \text{ with } m<n \} $

(b) $ \{ (-1)^m / n : m,n \in \mathbb{N} \} $

(c) $ \{ n/(3n+1) : n \in \mathbb{N} \} $

(d) $ \{ m/(m+n) : m,n \in \mathbb{N} \} $

Assume $A$ represents the set in context. My answers:

(a) I believe this should be $\sup A = 1$, because $m \ngeq n$ then $m/n$ can't be $1$ or greater, but it continuously gets closer. I wasn't sure here either $\inf A = 0$ or $\inf A = 1/2$ I think depending on the inclusion of $0$.

(b) This has $\min \{A\} = -1$ and $\max \{A\} = 1$, so $\inf A = -1$ and $\sup A = 1$.

(c) The values seem to be approaching $1/3$, so I answered $\sup A= 1/3$, and again I'm not sure if $\inf A = 0$ or $\inf A = 1/4$.

(d) $\inf A = 0$ (think huge $n$ and small $m$) and $\sup A = 1/2$ or $\sup A = 1$ depending on whether $n$ can be $0$ or not.

Any corrections?

Best Answer

Your reasoning for (a) is incorrect: if you fix $m=1$ and let $n \rightarrow \infty,$ we get $1/n \rightarrow 0.$ So $\inf A = 0.$

The others seem fine. Usually in analysis $\mathbb N$ does not include $0$ and judging by (b) I would assume this convention is taken here. It is likely that the author lists his notational conventions near the start of the book, so you should check there.