18
$\begingroup$

So I found this on the Wolfram website today:

cartoon

So I was wondering about how one might be able to (if possible) solve those four problems by hand. Here are the problems, $\LaTeX$ed:

  1. $ \lim_{x \to +\infty} \dfrac {\sqrt{x^3-x^2+3x}}{\sqrt{x^3}-\sqrt{x^2}+\sqrt{3x}} $
  2. $ \displaystyle\sum_{k=1}^{\infty} \dfrac {(-1)^{k+1} k^2}{k^3+1} $
  3. $ \dfrac {\mathrm{d}}{\mathrm{d}u} \left[ \dfrac {u^{n+1}}{(n+1)^2} \cdot \left[ (n+1) \ln u - 1 \right] \right] $
  4. $ \displaystyle\int_0^{2\pi}\displaystyle\int_0^{\frac{\pi}{4}}\displaystyle\int_0^4 \left( \rho \cos \phi \right) \rho^2 \sin \phi \, \mathrm{d}\rho \mathrm{d}\phi \mathrm{d}\theta$

Ideas

  1. The degree of the numerator is $\frac{3}{2}$. The degree of the denominator is $\frac{3}{2}$. The expression is of an indeterminate form, namely $\frac{\infty}{\infty}$, so we use l'Hoptial's Rule.
  2. No ideas, really.

  3. Derivatives are always pretty easy, although this one is a bit bashy. Basically bash Product and Chain Rules, etc.

  4. Integration by Parts bash? The $\mathrm{d}\theta$ part is very trivial. I mean for the $\rho$ and $\phi$.

$\endgroup$
  • 6
    $\begingroup$ Man, I love when Foxtrot does this. In some strips, they show an Into Physics text which has the famous green background/yellow wave of Halliday & Resnick. Clearly one of ours. $\endgroup$ – Ron Gordon Nov 21 '13 at 17:39
  • 1
    $\begingroup$ @RonGordon: forgive me if you already knew this, but Bill Amend was actually a physics major at Amherst. One of ours indeed! :) $\endgroup$ – Alex Wertheim Dec 11 '14 at 14:05
  • 1
    $\begingroup$ @AWertheim: I did not know that, thanks! In fact, I was a physics (and math) major just down the street at UMass Amherst. $\endgroup$ – Ron Gordon Dec 11 '14 at 14:06
15
$\begingroup$

I can derive the sum quite easily, using partial fractions and the residue theorem.

Note that $$\frac{k^2}{k^3+1} = \frac13 \left [\frac{1}{k+1} + \frac{2 k-1}{k^2-k+1}\right ]$$

Now, note that

$$\sum_{k=1}^{\infty} \frac{(-1)^{k+1}}{k+1} = 1-\log{2}$$

For the other piece, note the fortuitous coincidence that

$$\sum_{k=1}^{\infty} \frac{(-1)^{k+1} (2 k-1)}{k^2-k+1} = -\frac12 \sum_{k=-\infty}^{\infty} \frac{(-1)^{k} (2 k-1)}{k^2-k+1}$$

(One may see this by showing that the map $k \mapsto -k$ produces the $k+1$th term of the summand.)

Now, the sum on the RHS may be evaluated using residue theory. In general, one may show that

$$\sum_{k=-\infty}^{\infty} (-1)^k f(k) = -\pi \sum_n \operatorname*{Res}_{z=z_n} [\csc{\pi z} \, f(z)]$$

where the $z_n$ are the non-integral poles of $f$. In this case, the poles are at $z=e^{\pm i \pi/3}$, and we are also lucky to have an $f$ of the form $g'/g$, so that the residue at each pole is simply $1$. Therefore, we have

$$\sum_{k=1}^{\infty} \frac{(-1)^{k+1} (2 k-1)}{k^2-k+1} = \frac{\pi}{2} \left ( \csc{\left (\pi e^{i \pi/3}\right )}+ \csc{\left (\pi e^{-i \pi/3}\right )} \right ) = \pi \, \text{sech}{\left ( \frac{\sqrt{3} \pi}{2}\right )}$$

Therefore,

$$\sum_{k=1}^{\infty} \frac{(-1)^{k+1} k^2}{k^3+1} = \frac13 \cdot \left [ 1-\log{2} + \pi \, \text{sech}{\left ( \frac{\sqrt{3} \pi}{2}\right )}\right ]$$

$\endgroup$
  • 1
    $\begingroup$ I knew this had to be one of your answers by the time I was 4 lines in... :) $\endgroup$ – apnorton Dec 30 '13 at 17:31
  • 2
    $\begingroup$ @anorton: good, I am pleased. That means I am teaching you by rote. $\endgroup$ – Ron Gordon Dec 30 '13 at 18:40
  • 1
    $\begingroup$ @anorton Same! $\mathrm{}\\$ $\endgroup$ – Ahaan S. Rungta Dec 31 '13 at 5:54
5
$\begingroup$

For 1, divide the numerator and denominator by $x^{\frac 32}$ This gets you $\frac {\sqrt{1+\text{small stuff}}}{1+\text{small stuff}}$ which goes to $1$

For 2, Alpha gets a mess: $\frac 13\left(1-\log 2 + \frac{\pi}{ \cosh \left(\frac {\sqrt 3 \pi}2\right)}\right)$

For 3, I agree

For 4, the integrals are independent as the limits don't depend on the variables, so it is $\left(\int_0^{2\pi} d\theta \right)\left( \int_0^{\frac \pi 4} \sin \phi \cos \phi d\phi \right)\left( \int_0^4 \rho^3 d\rho \right)$, all of which are not too tough.

$\endgroup$
4
$\begingroup$

My take on $\#3$: $$ \begin {align*} \dfrac {\mathrm{d}}{\mathrm{d}u} \left[ \dfrac {u^{n+1}}{(n+1)^2} \cdot \left[ (n+1) \ln u - 1 \right] \right] &= \dfrac {\mathrm{d}}{\mathrm{d}u} \left[ \dfrac {u^{n+1}}{(n+1)^2} \cdot (n+1) \ln u - \dfrac {u^{n+1}}{(n+1)^2} \right] \\&= \dfrac {\mathrm{d}}{\mathrm{d}u} \left[ \dfrac {u^{n+1} \ln u}{n+1} - \dfrac {1}{(n+1)^2} \cdot u^{n+1} \right] \\&= \dfrac {1}{n+1} \cdot \dfrac {\mathrm{d}}{\mathrm{d}u} \left[ u^{n+1} \ln u - \dfrac {u^{n+1}}{n+1} \right] \\&= \dfrac {1}{n+1} \cdot \left[ \dfrac {\mathrm{d}}{\mathrm{d}u} \left( u^{n+1} \ln u \right) - \dfrac {\mathrm{d}}{\mathrm{d}u} \left( \dfrac {u^{n+1}}{n+1} \right) \right] \\&= \dfrac {1}{n+1} \cdot \left[ \dfrac {u^{n+1}}{u} + (n+1) \cdot u^n \ln u - u^n \right] \\&= \dfrac {1}{n+1} \cdot \left[ {u^n} + (n+1) \cdot u^n \ln u {-u^n} \right] \\&= \boxed {u^n \ln u}. \end {align*} $$Beautifully done intentionally?

$\endgroup$
  • 1
    $\begingroup$ $\color{red}{\boxed{+1}}$ Great work @Ahaan ! :-) $\endgroup$ – user93957 Jan 2 '14 at 22:02

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.