4
$\begingroup$

The motive is to evaluate the following limit:

$$\lim_{n\to \infty}\sum_{r=1}^{n} \frac{r}{n^2 + n + r}$$

I wrote it as

$$ \lim_{n\to \infty}\sum_{r=1}^{n}\frac{r/n}{1 + 1/n + r/n^2} \approx ^{?} \int_{0} ^{1} x \, dx = \frac{1}{2}$$

Now is this correct? Doesn't seem very correct to me. Thanks for your thoughts :)


Oliver Oloa gave a hint on Sandwich theorem but removed answer.

$$\sum_{r=1}^{n} \frac{r}{n^2 + n + n} \le \sum_{r=1}^{n} \frac{r}{n^2 + n + r} \le \sum_{r=1}^{n} \frac{r}{n^2 + n + 1}$$

Using this I think we get $1/2 \le L \le 1/2$ so limit is $1/2$.

$\endgroup$
  • $\begingroup$ Oops I made a mistake in my answer... $\endgroup$ – Olivier Oloa Jan 1 '18 at 12:02
  • $\begingroup$ You can convert into an integral only if you get something $(1/n)\sum f(r/n) $. You should first get a factor of $1/n$ outside the sum. I will give it a try and post an answer if I succeed. $\endgroup$ – Paramanand Singh Jan 1 '18 at 12:06
  • $\begingroup$ After you last update the limit should be $1/2$ which matches your integral answer. $\endgroup$ – Paramanand Singh Jan 1 '18 at 12:09
  • 1
    $\begingroup$ Just note that the sum can be written as $(1/n)\sum (r/n) (1+1/n+r/n^2)^{-1}=(1/n)\sum (r/n) +o(1)$ so that the integral also works fine. $\endgroup$ – Paramanand Singh Jan 1 '18 at 12:12
  • $\begingroup$ @Paramanand Do we have rules about when this ignoring works and when it doesn't ? $\endgroup$ – samjoe Jan 1 '18 at 12:13
2
$\begingroup$

Hint. One may also write $$ \sum_{r=1}^{n} \frac{r}{n^2 + n + r}=(n^2+n)\left(H_{n^2+n}-H_{n^2+2n}\right)+n $$ and conclude, as $n \to \infty$, with the asymptotics $$ H_n=\ln n+\gamma+\frac1{2n}+O\left(\frac1{n^2}\right). $$

$\endgroup$
4
$\begingroup$

Oliver Oloa gave a hint on Sandwich theorem but removed answer.

$$\sum_{r=1}^{n} \frac{r}{n^2 + n + n} \le \sum_{r=1}^{n} \frac{r}{n^2 + n + r} \le \sum_{r=1}^{n} \frac{r}{n^2 + n + 1}$$

Using this we get:

$$ \frac{n(n+1)}{2(n^2 + n + n)} \le \sum_{r=1}^{n} \frac{r}{n^2 + n + r} \le \frac{n(n+1)}{2(n^2 + n + 1)}$$

So as $n\to \infty$ the limit is $1/2$

$\endgroup$
  • $\begingroup$ You should accept your answer above which is OK for future refence. Thank you. $\endgroup$ – Olivier Oloa Jan 1 '18 at 12:40
2
$\begingroup$

The limit is $\frac{1}{2}$, since $\sum_{r=1}^{n}r = \frac{1}{2}n^2+\frac{1}{2}n$ and both $n^2+2n$ and $n^2+n+1$ are $n^2+O(n)$.

$\endgroup$
  • $\begingroup$ Ok Jack but there is error in the first method, what is that (answer is same by fluke) $\endgroup$ – samjoe Jan 1 '18 at 12:11
  • $\begingroup$ @samjoe: the first method is not incorrect: the wanted sum is not a Riemann sum strictly speaking, but it can be approximated by Riemann sums. It is enough to justify $\approx$ properly. $\endgroup$ – Jack D'Aurizio Jan 1 '18 at 12:13
  • $\begingroup$ Yes I mean how do we do that $\endgroup$ – samjoe Jan 1 '18 at 12:14
  • 3
    $\begingroup$ @samjoe: by approximating $\frac{r}{n^2+n+r}$ with $\frac{r}{n^2}$ and $\frac{r}{(n+1)^2}$, for instance. $\endgroup$ – Jack D'Aurizio Jan 1 '18 at 12:16
  • $\begingroup$ last comment was very useful for me! Thanks a lot :) $\endgroup$ – samjoe Jan 2 '18 at 14:32
0
$\begingroup$

$\newcommand{\bbx}[1]{\,\bbox[15px,border:1px groove navy]{\displaystyle{#1}}\,} \newcommand{\braces}[1]{\left\lbrace\,{#1}\,\right\rbrace} \newcommand{\bracks}[1]{\left\lbrack\,{#1}\,\right\rbrack} \newcommand{\dd}{\mathrm{d}} \newcommand{\ds}[1]{\displaystyle{#1}} \newcommand{\expo}[1]{\,\mathrm{e}^{#1}\,} \newcommand{\ic}{\mathrm{i}} \newcommand{\mc}[1]{\mathcal{#1}} \newcommand{\mrm}[1]{\mathrm{#1}} \newcommand{\pars}[1]{\left(\,{#1}\,\right)} \newcommand{\partiald}[3][]{\frac{\partial^{#1} #2}{\partial #3^{#1}}} \newcommand{\root}[2][]{\,\sqrt[#1]{\,{#2}\,}\,} \newcommand{\totald}[3][]{\frac{\mathrm{d}^{#1} #2}{\mathrm{d} #3^{#1}}} \newcommand{\verts}[1]{\left\vert\,{#1}\,\right\vert}$ \begin{align} \lim_{n\to \infty}\sum_{r = 1}^{n}{r \over n^{2} + n + r} & = \lim_{n\to \infty}\bracks{% {1 \over n}\sum_{r = 1}^{n}{r \over n} + \sum_{r = 1}^{n}\pars{{r \over n^{2} + n + r} - {r \over n^{2}}}} \\[5mm] & = \lim_{n\to \infty}\bracks{% {1 \over n}\sum_{r = 1}^{n}{r \over n} - \sum_{r = 1}^{n}{nr + r^{2} \over \pars{n^{2} + n + r}n^{2}}} \end{align}

Note that

$$ 0 < \sum_{r = 1}^{n}{nr + r^{2} \over \pars{n^{2} + n + r}n^{2}} < \sum_{r = 1}^{n}{2n^{2} \over \pars{n^{2} + n + 1}n^{2}} = {2n \over n^{2} + n + 1}\,\,\,\stackrel{\mrm{as}\ n\ \to\ \infty}{\to}\,\,\,{\large 0} $$ such that \begin{align} \lim_{n\to \infty}\sum_{r = 1}^{n}{r \over n^{2} + n + r} & = \lim_{n\to \infty}\pars{% {1 \over n}\sum_{r = 1}^{n}{r \over n}} = \int_{0}^{1}x\,\dd x = \bbx{1 \over 2} \end{align}

$\endgroup$

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.