5
$\begingroup$

I was willing to determine the sum of following $$S:=1+\frac12-\frac13-\frac14+\frac15+\frac16-\frac17-\frac18+\cdots$$

I tried the following \begin{align*} S=&\sum\limits_{n=0}^\infty (-1)^n\left(\frac{1}{2n-1}+\frac{1}{2n}\right)\\ =&\sum\limits_{n=0}^\infty (-1)^n\left(\int_0^1 x^{2n-2}dx+\int_0^1 x^{2n-1} dx\right)\\ =&\int_0^1 \sum\limits_{n=0}^\infty (-1)^n\left(x^{2n-2}+x^{2n-1}\right)\\ =&\int_0^1 [x^{-2} \sum\limits_{n=0}^\infty (-1)^nx^{2n}+x^{-1} \sum\limits_{n=0}^\infty (-1)^nx^{2n}] \end{align*} and don't know after his what to do . Can you please help me on this regard?

Thanking you in advance

$\endgroup$
2
  • 5
    $\begingroup$ you have two geometric series. sum them up and take the limit $x\rightarrow1^-$ $\endgroup$ – tired May 18 '15 at 8:36
  • 4
    $\begingroup$ I suppose that the summations start at $n=1$ instead of $n=0$ $\endgroup$ – Claude Leibovici May 18 '15 at 8:41
2
$\begingroup$

Note that $$S=A+B$$ where $\displaystyle A=\sum_{k\ge 1}\frac{(-1)^{k-1}}{2k-1},\ B=\sum_{k\ge 1}\frac{(-1)^{k-1}}{2k}$. Clearly, $B=\displaystyle \frac{\ln 2}{2}$. and, $A=\tan^{-1}1=\pi/4$ which you can find out using your technique as below $$A=\sum_{k\ge 1}(-1)^{k-1}\int_0^1 x^{2k-2}dx\\=\int_{0}^1\sum_{k\ge 1}(-1)^{k-1}x^{2k-2}dx\quad(\mbox{Use Fubini to justify the change of order})\\=\int_{0}^1 \frac{dx}{1+x^2}=\tan^{-1}1=\pi/4\\ \mbox{similarly, }\ B=\int_{0}^1 \sum_{k\ge 1}(-1)^{k-1}x^{2k-1}dx\\=\int_0^1 \frac{x}{1+x^2}dx\\=\frac{\ln 2}{2}$$

$\endgroup$
1
  • $\begingroup$ Well you can use DCT here with equal ease but Fubini's theorem also work, with the measures in mind as the counting measure on $\mathbb{N}$ and the Lebesgue measure on $[0,1]$. Please see this for more clarification. $\endgroup$ – Samrat Mukhopadhyay Dec 21 '15 at 5:51
1
$\begingroup$

Couldn't you just write this as

$$\sum_{k=0}^{\infty} \frac{(-1)^k}{2 k+1} + \frac12 \sum_{k=0}^{\infty} \frac{(-1)^{k}}{k+1} = \frac{\pi}{4} + \frac12 \log{2}$$

$\endgroup$
1
  • $\begingroup$ But that would be too easy. +1 :-)) -Mark $\endgroup$ – Mark Viola Dec 18 '15 at 20:26
0
$\begingroup$

Let $$S = \sum_{n = 1}^{\infty} (-1)^n\left(\frac{1}{2n-1}+\frac{1}{2n}\right).$$ Then

$\begin {eqnarray} S & = & \sum_{k = 1}^{\infty} (-1)^{2k}\left(\frac{1}{4k-1}+\frac{1}{4k}\right) + \sum_{k = 1}^{\infty} (-1)^{2k+1}\left(\frac{1}{4k+1}+\frac{1}{4k+2}\right) \nonumber \\ & = & \sum_{k = 1}^{\infty} \left(\frac {1} {4k - 1} - \frac {1} {4k + 1}\right) + \sum_{k =1}^{\infty} \left(\frac {1} {4k} - \frac {1} {4k + 2}\right), \end {eqnarray}$

which is easier to proceed now.

$\endgroup$
3
  • $\begingroup$ How is your final result easier?? $\endgroup$ – Mark Viola Dec 18 '15 at 20:24
  • $\begingroup$ Because it's not a strangely alternate series anymore. $\endgroup$ – user98186 Dec 18 '15 at 20:33
  • $\begingroup$ But how is that easier? How would you propose that one proceed? Use Digamma function IDs? $\endgroup$ – Mark Viola Dec 18 '15 at 21:04

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.