2

Show that the sequences $$x_n=\sum_{k=0}^n\frac1{k!}\quad\text{and}\quad y_n=\left(1+\frac1n\right)^n$$ are convergent and have the same limit.

I do not know how to show that either of these are bounded sequences, but they are both pretty clearly monotone.

Lutz Lehmann
  • 126,666

2 Answers2

1

For $k \ge 2,k! > 2^{k-1}$ so $$\frac{1}{k!} \le \frac{1}{2^{k-1}}$$

i.e. terms of the first sum are all (after a while) bounded by 1/2 + 1/4 + 1/8 + ... = 1.

oks
  • 1,345
1

Although I have provided a link to an answer (of a more general question) in the comments, it would be better to give you the main ideas related to your specific question.

Note that the sequence $x_{n}$ is actually the $n$-th partial sum of a series which is convergent (easily proved via ratio test). Another approach to proving that $x_{n}$ converges is to show that it increases as $n$ increases and is bounded. Clearly as $n$ increase the number of terms in sum defining $x_{n}$ increases and hence the sequence $x_{n}$ increases. Again we can see that if $n > 2$ then $$\begin{aligned}x_{n} &= 1 + 1 + \frac{1}{2!} + \frac{1}{3!} + \cdots + \frac{1}{n!}\\ &\leq 1 + 1 + \frac{1}{2} + \frac{1}{2^{2}} + \cdots + \frac{1}{2^{n - 1}}\\ &= 1 + \dfrac{1 - \dfrac{1}{2^{n}}}{1 - \dfrac{1}{2}}\text{ (sum of a finite GP)}\\ &\leq 1 + \dfrac{1}{1 - \dfrac{1}{2}} = 3\end{aligned}$$ Hence $x_{n}$ is bounded and therefore $x_{n}$ tends to a limit.

Next we can focus on $y_{n}$. Clearly if $n > 3$ then we can see that $$\begin{aligned}y_{n} &= \left(1 + \frac{1}{n}\right)^{n}\\ &= 1 + n\cdot\frac{1}{n} + \frac{n(n - 1)}{2!}\left(\frac{1}{n}\right)^{2} + \cdots + \text{ (upto }n\text{ terms)}\\ &= 1 + 1 + \dfrac{1 - \dfrac{1}{n}}{2!} + \dfrac{\left(1 - \dfrac{1}{n}\right)\left(1 - \dfrac{2}{n}\right)}{3!} + \cdots\end{aligned}$$

Now it is easy to see that as $n$ increases number of terms in the above expression increases as well as each term of the above sum also increases. It follows that $y_{n}$ increases as $n$ increases. Again we can note that $$\begin{aligned}y_{n} &= 1 + 1 + \dfrac{1 - \dfrac{1}{n}}{2!} + \dfrac{\left(1 - \dfrac{1}{n}\right)\left(1 - \dfrac{2}{n}\right)}{3!} + \cdots\\ &\leq 1 + 1 + \frac{1}{2!} + \frac{1}{3!} + \cdots + \frac{1}{n!}\\ &= x_{n} \leq 3 \text{ (as proved before)}\end{aligned}$$ therefore $y_{n}$ is also bounded and $\lim_{n \to \infty}y_{n}$ also exists.

Also we have seen above that $y_{n} \leq x_{n}$ so that $\lim_{n \to \infty}y_{n} \leq \lim_{n \to \infty}x_{n}$. To establish equality we need to bound $x_{n}$ by another sequence. That sequence is given by $z_{n} = (1 - (1/n))^{-n}$. Clearly using general binomial theorem we have $$\begin{aligned}z_{n} &= 1 + n\cdot\frac{1}{n} + \frac{n(n + 1)}{2!}\left(\frac{1}{n}\right)^{2} + \cdots\\ &= 1 + 1 + \dfrac{1 + \dfrac{1}{n}}{2!} + \dfrac{\left(1 + \dfrac{1}{n}\right)\left(1 + \dfrac{2}{n}\right)}{3!} + \cdots\\ &\geq 1 + 1 + \frac{1}{2!} + \cdots + \frac{1}{n!} = x_{n}\end{aligned}$$ and thus we finally arrive at $$y_{n} \leq x_{n} \leq z_{n}$$ The next thing is to show that both $y_{n}$ and $z_{n}$ tend to same limit as $n \to \infty$ and then by Squeeze theorem $x_{n}$ would also tend to same limit. To proceed further we can note that $$\begin{aligned}\frac{y_{n}}{z_{n}} &= \left(1 - \frac{1}{n^{2}}\right)^{n}\\ &= 1 - n\cdot\frac{1}{n^{2}} + \frac{n(n - 1)}{2!}\left(\frac{1}{n^{2}}\right)^{2} - \cdots + \text{ (upto }n\text{ terms)}\\ &= 1 - \frac{1}{n} + \dfrac{1 - \dfrac{1}{n}}{2!}\left(\frac{1}{n}\right)^{2} - \cdots\\ &= 1 - \phi(n)\end{aligned}$$ where the function $\phi(n)$ is a finite sum defined by $$\phi(n) = \frac{1}{n} - \dfrac{1 - \dfrac{1}{n}}{2!}\left(\frac{1}{n}\right)^{2} + \dfrac{\left(1 - \dfrac{1}{n}\right)\left(1 - \dfrac{2}{n}\right)}{3!}\left(\frac{1}{n}\right)^{3}- \cdots + $$ Now we can estimate $\phi(n)$ as $$\begin{aligned}0\leq|\phi(n)| &\leq \frac{1}{n} + \dfrac{1 - \dfrac{1}{n}}{2!}\left(\frac{1}{n}\right)^{2} + \dfrac{\left(1 - \dfrac{1}{n}\right)\left(1 - \dfrac{2}{n}\right)}{3!}\left(\frac{1}{n}\right)^{3} + \cdots\\ &\leq \frac{1}{n} + \frac{1}{2!}\cdot\frac{1}{n^{2}} + \cdots\\ &\leq \frac{1}{n} + \frac{1}{2n^{2}} + \frac{1}{2^{2}n^{3}} + \cdots\\ &= \frac{1}{n}\left(1 + \frac{1}{2n} + \frac{1}{(2n)^{2}} + \cdots\right)\\ &= \frac{1}{n}\cdot\dfrac{1 - \dfrac{1}{(2n)^{n}}}{1 - \dfrac{1}{2n}}\\ &\leq \frac{1}{n}\cdot\dfrac{1}{1 - \dfrac{1}{2n}} \to 0\text{ (as }n \to \infty)\end{aligned}$$

It follows that $\lim_{n \to \infty}\phi(n) = 0$ and hence $\lim_{n \to \infty}y_{n}/z_{n} = 1$ and hence $\lim_{n \to \infty}y_{n} = \lim_{n \to \infty}z_{n}$.

  • You are a genius, I would not be able to come up with something like this on my own, thank you. – terrible at math Feb 22 '14 at 06:49
  • I am bit flattered by your appreciation, but to be frank there are many users in this website who deserve much more appreciation than me. I do know some stuff and I know them well enough. Thats all I can say. – Paramanand Singh Feb 22 '14 at 07:00
  • I have simplified the last past related to equality of limit of $y_{n}, z_{n}$ so that my earlier linked answer is no longer needed. – Paramanand Singh Feb 22 '14 at 07:01
  • I dunno man, you seem like an extremely intelligent person who really knows his stuff. That is admirable and deserves respect. I am just barely passing undergrad. – terrible at math Feb 22 '14 at 07:01
  • By the way I can let you in on my secret of knowing such things well enough. Try to find the book "A Course of Pure Mathematics" 10th edition by G. H. Hardy. Most of what I have learnt comes from this book. – Paramanand Singh Feb 22 '14 at 07:05
  • One book? that's impressive that you could learn so much from just that. – terrible at math Feb 22 '14 at 07:39
  • There are many other books too but the best is from G H Hardy. A list of my favorite 7-8 books is available at http://paramanands.blogspot.in/p/bibliography.html – Paramanand Singh Feb 22 '14 at 07:47