I was reading about how completeness is required for limits. And I came across this:

the sequence $a_n=(1+1/n)^n$ is monotonically increasing and bounded by 3 and so we expect it to converge, but that it does not converge within $\mathbb{Q}$. More generally it stands to reason that any sequence of real numbers which is increasing and bounded must converge to some real number. This is a consequence of completeness of $IR$

My question is: How is the mentioned sequence monotonically increasing and bounded by $3$ ?

  • $\begingroup$ you can prove it by induction $\endgroup$ – hchengz Oct 5 '13 at 6:41
  • $\begingroup$ You cannot generally prove convergence statements using induction. Induction only shows things to be true for finite n. $\endgroup$ – Jebruho Oct 5 '13 at 7:02
  • $\begingroup$ @Jebruho: If $(a_n)$ is a sequence, the statements $a_{n+1}\geq a_n$ and $a_n\leq 3$ are statements about integers, hence may in principle be proved by induction. For convergence in $\mathbb R$ one then needs to invoke completeness, but GTX OC is not asking about that part. $\endgroup$ – Jonas Meyer Oct 5 '13 at 7:37
  • $\begingroup$ Related: math.stackexchange.com/questions/167843/…, math.stackexchange.com/q/220723 $\endgroup$ – Jonas Meyer Oct 5 '13 at 7:38

The monotonicity follows from the AM-GM inequality for $n+1$ points. Taking $x_1 = x$ and $x_2 = \cdots = x_{n+1} = y$, we get $$ \sqrt[n+1]{xy^n} \leq \frac{x+ny}{n+1}. $$ In particular, taking $x = 1$ and $y = 1+\frac{1}{n}$ yields $a_n \leq a_{n+1}$ ($n \geq 1$).

Now as another special case, take $x = 1$ and $y = 1-\frac{1}{n}$ to get $$ b_n \geq b_{n+1} \qquad (n \geq 1), $$ where $b_n := \left( 1-\frac{1}{n}\right)^{-n}$. We let you verify that $$ b_n = a_{n-1} \left( 1+\frac{1}{n-1}\right) \geq a_{n-1}, \qquad (n \geq 2). $$ Hence $$ a_1 \leq a_2 \leq a_3 \leq \cdots \leq b_3 \leq b_2 \leq b_1 $$ and we deduce that $a_n \leq b_m$ for all $n, m \in \mathbb{N}$. In particular, taking $m = 6$, we get $$ a_n \leq \left( \frac{6}{5} \right)^6 \leq 3. $$

  • $\begingroup$ Can't we show that it is increasing monotonically by differentiating $a_n$ with respect to n and checking whether it is always positive for all n $ \leq 1$ ? $\endgroup$ – GTX OC Oct 5 '13 at 7:49
  • $\begingroup$ Well you can't differentiate with respect to $n$ because $n \in \mathbb{N}$. However you could try and differentiate the function $(1+1/x)^x$ on $x > 0$. Then if you can show it is always $\geq 0$ on $x > 0$ that would do it. $\endgroup$ – Amateur Oct 5 '13 at 7:57

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.