Let $A \subseteq \mathbb{R}$ be a bounded above set.

Prove that exists a sequence $\{a_n\}_{n=1}^{\infty}$ such that $\forall n \in \mathbb{N}: a_n \in A$ and $$\lim_{n \to \infty}a_n = sup(A)$$.

I thought about: for every $n \in \mathbb{N}$ taking $a_n \in A - (- \infty, a_{n-1}]$ Is that a good idea?

  • 2
    $\begingroup$ My guess is that you meant “sequence”, not “series”. $\endgroup$ Nov 16 '19 at 18:31

Close but not quite.

There are a few things you overlooked.

1) Its possible that $(a_n, \sup A)$ need not have any values at all. This can only happen if $\sup A \in A$. For example take a finite set $A$ or the set $(0,1) \cup \{2\}$.

In the case that $\sup A \in A$ then $A$ has a max element. There's nothing in the definition of sequence that the terms be unique. In this case just take $a_k = \sup A$ for all $k$. That'll do. This is not an interesting case.

MUCH more interesting is when $\sup A \not\in A$. In this case $\max A$ does not exist and $A$ is infinite.

In this case $(a_n, \sup A)$ will never be non empty and you can inductively find $a_n < a_{n+1} < a_{n+2} < .......... < \sup A$.


2) there's no reason to assume $\lim a_n = \sup A$. Instead you might have $\lim a_n = m$ for some $m < \sup A$.

For example. Suppose $A = (0, 2)$ and you chose $a_k = 1 - \frac 1k$. Then $\lim a_n = 1 < \sup A = 2$.

Can you work around this?

Let $a_k \in (\sup A - \frac 1k)$. Because $\sup A$ is least upper bound $\sup A - \frac 1k$ is not an upper bound and so an $a_k\in (\sup A- \frac 1k, \sup A); a_k\in A$ will always exist. And for any $\epsilon > 0$ if $n > \frac 1\epsilon$ then $\sup A - \epsilon < \sup A - \frac 1n < a_n <\sup A$. so $\lim a_n = \sup A$.


Note: your idea that $a_{n} < a_{n+1} < \sup A$, was an EXCELLENT idea and very good first step to solving. But ultimately wasn't necessary. But could still be used. Maybe, let $a_{n+1} \in (\frac {\sup A + a_n}2, \sup A)$ (that is, $a_{n+1}$ is between $\sup A$ and the midpoint between $\sup A$ and $a_n$.)

  • $\begingroup$ I dont understand what do you mean by$a_k \in (\sup A - \frac{1}{k})$ the right hand side is a number, not a group so why you using in $\in$ $\endgroup$ Nov 16 '19 at 19:03
  • $\begingroup$ Oops, a typo. I meant $a_k \in (\sup A -\frac 1k, \sup A)$. I was trying to use the OP's language and notation. $\endgroup$
    – fleablood
    Nov 16 '19 at 20:52

Put $\;M:=\sup A\;$ , then by the very definition of supremum of a real set, we have that for any natural number $\;n\in\Bbb N\;$ there exists $\;a_n\in A\;$ s.t.

$$M-\frac1n<a_n\le M$$

Now make your choice and use the squeeze theorem...


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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