Ask your own question, for FREE!
Mathematics 17 Online
OpenStudy (anonymous):

I had a quick question about series: Which tests prove convergence, and not absolute or conditional convergence?

OpenStudy (anonymous):

@rational

OpenStudy (rational):

Except for alternating series test, almost all other tests require terms to be positive right ?

OpenStudy (anonymous):

Umm yeah

OpenStudy (anonymous):

I was just thinking that I want to prove convergence, and show that it doesn't converge absolutely

OpenStudy (rational):

then i think you may use alternating series test to prove "convergence of actual series" and use any other test to prove "convergence/divergence of absolute value series."

OpenStudy (anonymous):

I thought that the alternating series test proved conditional convergence though

OpenStudy (rational):

may be lets work an example, are you working on anything specific ?

OpenStudy (anonymous):

Yeah. Determine whether the series converges absolutely, conditionally, or diverges. I already proved that the ratio test diverged. \[\sum_{n=1}^{\infty}(-1)^{n+1}((1+n)/n^2)\]

OpenStudy (anonymous):

And I started using the alternating series test, but I didn't know if that showed that the series converged conditionally, or converged in general

OpenStudy (rational):

again, alternating series test has nothing to do with "conditional convergence" To say the series is "conditionally convergent", you must establish two things : 1) the actual series converges. 2) the absolute value of series diverges

OpenStudy (anonymous):

Oooh ok, right. I think that I understand what my teacher meant now when she said that it proved conditional convergence, because we made a flowchart. You could only use the alternating series test if absolute convergence didn't work out

OpenStudy (anonymous):

But in itself, it doesn't actually prove conditional convergence, just that the series converged

OpenStudy (anonymous):

So because the series that I posted diverged when tested for absolute convergence, and converged when tested with the AST, it converges conditionally

OpenStudy (rational):

Yes.. \[\sum_{n=1}^{\infty}(-1)^{n+1}\dfrac{1+n}{n^2}\] \(b_n := \frac{1+n}{n^2}\) 1) Clearly \(b_n\ge 0\) for all \(n\). 2) the sequence \(\{b_n\}\) is decreasing because : \[\color{blue}{b_{n+1}}=\frac{n+1+1}{(n+1)^2} = \frac{1}{n+1}+\frac{1}{(n+1)^2} \color{blue}{\lt} \frac{1}{n}+\frac{1}{n^2}=\frac{1+n}{n^2}=\color{blue}{b_n}\] 3) \(\lim\limits_{n\to\infty}b_n = \lim\limits_{n\to\infty}\frac{1+n}{n^2}=0\) Therefore this series meets the hypotheiss of alternating series test, so the series \(\sum_{n=1}^{\infty}(-1)^{n+1}\dfrac{1+n}{n^2}\) converges by alternating series test.

OpenStudy (rational):

Alternating series test story ends there.

OpenStudy (rational):

Notice that you need to still test the convergence of absolute value of series to say whether it covnerges absolutely or conditionally

OpenStudy (anonymous):

Yes, and that diverged, so it converges conditionally?

OpenStudy (rational):

Yes!

OpenStudy (anonymous):

Ok great! I also had a quick question about the ratio test if you dont mind

OpenStudy (rational):

sure ask

OpenStudy (anonymous):

Ok. Do I always need the absolute value bars for the ratio test?

OpenStudy (rational):

Yes, thats trur for ratio test also

OpenStudy (anonymous):

Ok, so even if the series could not possibly have any negative terms (for example, 1/n), you still need the abs vals?

OpenStudy (rational):

absolute value of a positive term is positive, so it wont change in anything. but it is good to put absolute bars always when u apply ratio test

OpenStudy (rational):

|1/n| = 1/n for n>0

OpenStudy (anonymous):

So the ratio test always proves absolute convergence regardless of the sign of the terms in the series

OpenStudy (rational):

that looks like a correct statement

OpenStudy (anonymous):

Ok. Thank you so much!! I might have to get a tattoo that says "the alternating series test only proves regular convergence"

OpenStudy (anonymous):

Wait, then why don't we start with the alternating series test if that proves convergence? Absolute convergence tests dont matter if the series doesn't converge at all

OpenStudy (rational):

Alternating series test is the first thing to try if you see (-1)^n in your series

OpenStudy (anonymous):

If it doesn't have (-1)^n, how do I prove that it just generally converges?

OpenStudy (rational):

use any other tests

OpenStudy (anonymous):

I just found a series that does not follow the AST, but converges absolutely as well

OpenStudy (rational):

possible, what is it ?

OpenStudy (anonymous):

(-1)^n+1 (sin n/n^2)

OpenStudy (anonymous):

Doesn't that mean that it doesn't converge, but it converges absolutely? That doesn't make any sense

OpenStudy (rational):

did you use comparison with p-series to determine it converges absolutely ?

OpenStudy (anonymous):

Yes

OpenStudy (anonymous):

the LCT

OpenStudy (rational):

good, since the absolute value of series converges, the actual series also converges.

OpenStudy (anonymous):

But it doesn't pass the AST, so that means that it doesn't converge

OpenStudy (rational):

absolute convergence =====> the actual series converge

OpenStudy (rational):

who cares about alternating series test

OpenStudy (anonymous):

I thought that you said the AST proves convergence

OpenStudy (rational):

you said we cannot apply alternating series test here because the terms are not positive

OpenStudy (rational):

so we're done with alternating series test. it is silent for this series.

OpenStudy (anonymous):

The terms aren't all positive. Theres a (-1)^n+1

OpenStudy (anonymous):

and a sin(n)

OpenStudy (rational):

yes bn = sin(n)/n^2 the terms in bn are not all positive, so we cannot apply alternating series test

OpenStudy (rational):

that doesn't mean the series diverges. that just means alternating series test is silent and we need to try some other test

OpenStudy (anonymous):

But I think that I understand. You can only use the AST if absolute convergence doesn't work. The fact that absolute convergence didn't work means that if the AST works, it has to be conditional convergence! I think that I get it now!

OpenStudy (anonymous):

Does that sound legit?

OpenStudy (rational):

that looks okay for now, but after working a few more problems you will see why i said "absolute convergence/conditional convergence" has nothing to do with AST.

OpenStudy (anonymous):

I think that I understand that too. AST proves that the series has to converge, and the only other way that it can converge is conditional

Can't find your answer? Make a FREE account and ask your own questions, OR help others and earn volunteer hours!

Join our real-time social learning platform and learn together with your friends!
Can't find your answer? Make a FREE account and ask your own questions, OR help others and earn volunteer hours!

Join our real-time social learning platform and learn together with your friends!