0
$\begingroup$

I want to know about the convergence or divergence of the following series:

$$\sum \sin (a_n) $$

where

$$a_n=\frac{(-1)^n}{n}$$

The tests that I tried were inconclusive. Is it possible to know?

Or, more generally, does this series converge for all conditionally convergent series given by the sequence $a_n$? I know this result holds for any absolutely convergent series (here), but I want to know if it works for any convergent series or if there's an explicit counterexample taking conditionally convergent series.

EDIT: as pointed out at the answer section, my example converges, but the general case is still unanswered.

$\endgroup$
3
$\begingroup$

For the first part

Does $\sum\limits_{n=1}^\infty \sin\frac{(-1)^n}{n}$ converges?

The answer is YES by Alternating series test.

For the second part

Does $\sum\limits_{n=1}^\infty \sin a_n$ converges for all conditional converging series $\sum\limits_{n=1}^\infty a_n$ ?

The answer is NO. Consider the series

$$a_n = \frac{\epsilon_n}{\lceil n/3 \rceil ^{1/3}} \quad\text{ where }\quad \epsilon_n = \begin{cases} +1, & n \not\equiv 0 \pmod 3\\ -2, & n \equiv 0 \pmod 3 \end{cases}$$

If we group the terms in units of three, one find

$$a_{3k-2} + a_{3k-1} + a_{3k} = 0$$ and the series $\sum\limits_{n=1}^\infty a_n$ converges conditionally. However, $$\sin a_{3k-2} + \sin a_{3k-1} + \sin a_{3k} = 2\sin\frac{1}{\sqrt[3]{k}} - \sin\frac{2}{\sqrt[3]{k}} = \frac{1}{k} + O\left(\frac{1}{k^{5/3}}\right)$$

This implies the partial sums $\sum\limits_{n=1}^N a_{n}$ behaves roughly as $\sum\limits_{k=1}^{\lfloor N/3 \rfloor} \frac{1}{k} \sim \log\frac{N}{3}$ for large $N$ and hence the corresponding series diverges.

$\endgroup$
1
$\begingroup$

$\sin(a_n)=(-1)^n\sin(\frac{1}{n})$ so you are dealing with an alternating series. Then the fact that $\sin(\frac{1}{n})$ is monotonically decreasing and converges to $0$ justifies the conclusion that the series is converging. The so-called Leibniz-test is applied.

$\endgroup$
0
$\begingroup$

For $x$ close to $0$, we have $\frac{|x|}{2}\leq |\sin{x}|\leq |x|$. Does that give you any ideas?

Edit: can be done way simpler, just noticed. $\sin(-x)=-\sin(x)$, and the sequence $\frac{1}{n}$ has the limit $0$. Now, if I were to mention a guy called Leibniz, we are almost done.

$\endgroup$

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.