1
$\begingroup$

$$\sum_{k=1}^\infty \frac{(-1)^k}{\sqrt{k-1}+(-1)^k}$$ I put it on wolfram alpha and it shows that converges, but I think the analysis has to be different, because I need to know if the series converge absolutely or uniform or conditional and wolfram doesn't show it. Any help you can give will be greatly appreciated

$\endgroup$
  • 1
    $\begingroup$ We cannot speak about uniform convergence here. $\endgroup$ – hamam_Abdallah May 31 '17 at 1:57
1
$\begingroup$

The series fails to converge even conditionally since

$$\frac{(-1)^k}{\sqrt{k-1} + (-1)^k} = \frac{(-1)^k(\sqrt{k-1} - (-1)^k)}{(\sqrt{k-1} + (-1)^k)(\sqrt{k-1} - (-1)^k)} = \frac{(-1)^k \sqrt{k-1}}{k-2} - \frac{1}{k-2}$$

and we have divergence of

$$\sum_{k=3}^\infty \frac{1}{k-2}$$

and convergence by the Dirichlet test of

$$\sum_{k=3}^\infty \frac{(-1)^k\sqrt{k-1}}{k-2}$$

$\endgroup$
  • $\begingroup$ This presents the quintessential example of a divergent alternating series for which the absolute value of general terms approach zero, thereby showing the importance of the monotonicity condition in Leibniz's test. (+1) $\endgroup$ – Mark Viola May 31 '17 at 2:44
  • $\begingroup$ @MarkViola: Glad somebody reads this stuff. Congratulations BTW $\endgroup$ – RRL May 31 '17 at 2:48
  • $\begingroup$ Ah, thank you! I have been told all of the secrets of the universe ... and I get a t-shirt! What's better than that? $\endgroup$ – Mark Viola May 31 '17 at 2:51
  • $\begingroup$ Sorry but I don't understand how to get this step $$\frac{(-1)^k(\sqrt{k-1} - (-1)^k)}{(\sqrt{k-1} + (-1)^k)(\sqrt{k-1} - (-1)^k)} = \frac{(-1)^k \sqrt{k-1}}{k-2} - \frac{1}{k-2}$$ can you explain it please? @RRL $\endgroup$ – C. Ballez Sep 21 '17 at 19:53
  • 1
    $\begingroup$ Sure. We know $(a-b)(a+b) = a^2 - b^2$. Applying this to the denominator we get $(\sqrt{k-1} + (-1)^k)(\sqrt{k-1} - (-1)^k) = (\sqrt{k-1})^2 - (-1)^{2k} = k-1 - ((-1)^2)^k = k - 1 -1^k = k-2$.// In the numerator we get $(-1)^k(\sqrt{k-1} - (-1)^k) = (-1)^k\sqrt{k-1} - (-1)^{2k} = (-1)^k\sqrt{k-1} - 1$. This rationalization is applicable to all terms with index $k \geqslant 3$ where the denominator is not $0$ and it is enough to consider the sum starting with $k=3$ to examine convergence / divergence. $\endgroup$ – RRL Sep 21 '17 at 20:24
0
$\begingroup$

hint for absolute

$$|u_k|\ge \frac {1}{\sqrt {k-1}+1} $$

$$\frac {1}{\sqrt {k-1}+1}\sim \frac {1}{\sqrt {k}} $$

$\sum \frac {1}{\sqrt {k}} $ is divergent thus

$\sum u_k $ is not absolutely conv.

$\endgroup$

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.