1
$\begingroup$

Decide whether the series are absolutely convergent, conditionally convergent, or divergent.

$$a) \sum _{n=1}^{\infty} (-1)^n \frac{(\log n)^{\log n}}{n^a} , a>0$$

$$b) \sum_{n=1}^{\infty} \frac {(-1)^{[\log n]}}{n}$$ where [ ] is greatest integer symbol.

My attempt : the easiest way to think about this problem is the Leibniz test both $a)$ and $b)$ will be conditionally convergents.

Is it correct ?

any hints/solutions

thanks u

$\endgroup$

2 Answers 2

1
$\begingroup$

For the second part note that for $n$ between $e^{k}$ and $e ^{k+1}$, $(-1)^{[\log n]}$ has a constant sign. Compare $\sum_{[e^{k}]}^{[e^{k}+1]} \frac 1 n$ with the integral of $\frac 1 x $ from $[e^{k}]$ to $[e^{k+1}]$ to see that the sum of the terms from $[e^{k}]$ to $[e^{k+1}]$ does not tend to $0$. Hence the series is divergent.

$\endgroup$
3
  • $\begingroup$ sir how $e^k $ and$ e^{k+1}$ come ? $\endgroup$
    – jasmine
    Aug 10, 2018 at 23:43
  • 1
    $\begingroup$ @stupid If $[\ln n]=k$ then $k \leq \ln n < k+1$. This gives $e^{k} <n <e^{k+1}$. $\endgroup$ Aug 11, 2018 at 0:36
  • $\begingroup$ sir thanks u... $\endgroup$
    – jasmine
    Aug 11, 2018 at 0:43
0
$\begingroup$

For the first one we have that

$$\frac{(\log n)^{\log n}}{n^a}\to \infty$$

indeed by $x=e^y \to \infty$ as $y\ge e^{2a}$

$$\frac{(\log x)^{\log x}}{x^a}=\frac{y^y}{e^{ay}}\ge\frac{e^{2ay}}{e^{ay}}=e^{ay}\to \infty$$

therefore the given series diverges.

For the second one we can write

$$\sum_{n=1}^{\infty} \frac {(-1)^{[\log n]}}{n}=\sum_{n=1}^{1} \frac {1}{n}-\sum_{n=2}^{2} \frac {1}{n}+\sum_{n=3}^{7} \frac {1}{n}-\sum_{n=8}^{20} \frac {1}{n}+\ldots+(-1)^{k+1}\sum_{n=[e^k]}^{[e^{k+1}-1]} \frac {1}{n}+\ldots =\sum_{k=1}^\infty (-1)^{k+1} a_k$$

that is an alternating sum of terms not convergent to zero and thus it diverges too, indeed as $k\to \infty$ by harmonic series

$$a_k=\sum_{n=[e^k]}^{[e^{k+1}-1]} \frac {1}{n}\sim \log (e^{k+1})-\log(e^k)=k+1-k=1$$

$\endgroup$
7
  • $\begingroup$ @RushabhMehta Once $a_n \not \to 0$ the series diverges anyway, the$ (-1)^n$ factor is completely useless to conclude here. $\endgroup$
    – user
    Aug 10, 2018 at 22:17
  • $\begingroup$ @RushabhMehta For the second one I'm thinking about it, it is not so trivial as the first one. $\endgroup$
    – user
    Aug 10, 2018 at 22:18
  • $\begingroup$ @RushabhMehta Now it should be complete. And please revise also your wrong first coment about the $(-1)^n$ term. $\endgroup$
    – user
    Aug 10, 2018 at 22:44
  • $\begingroup$ thanks u @gimusi..one last confusion how $y \ge e^{2a}$ ? $\endgroup$
    – jasmine
    Aug 10, 2018 at 23:51
  • 1
    $\begingroup$ @stupid As $y \ge e^{2a}$ we have $(y)^y\ge (e^{2a})^y=e^{2ay}$. Why are you considering $\log(\log y)>2$? $\endgroup$
    – user
    Aug 11, 2018 at 6:41

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .