Register to reply

Do Alternating Series Have Limits?

by I'm Awesome
Tags: alternating, limits, series
Share this thread:
I'm Awesome
#1
Apr28-12, 04:41 PM
P: 12
I would imagine that an alternating series that goes of to infinity doesn't have a limit because it keeps switching back and forth, but I can't find anything in my textbook about it. I just want to make sure that this is right.
Phys.Org News Partner Science news on Phys.org
FIXD tells car drivers via smartphone what is wrong
Team pioneers strategy for creating new materials
Team defines new biodiversity metric
micromass
#2
Apr28-12, 04:48 PM
Mentor
micromass's Avatar
P: 18,299
Some alternating series do have limits. For example:

[tex]\sum_{n=1}^{+\infty} \frac{(-1)^n}{n}[/tex]

has a limit (and it equals log(2)).

Other series, like

[tex]\sum_{n=1}^{+\infty} (-1)^n[/tex]

have a too large oscillation to have a limit.
Dickfore
#3
Apr28-12, 04:51 PM
P: 3,014
Quote Quote by micromass View Post
Some alternating series do have limits. For example:

[tex]\sum_{n=1}^{+\infty} \frac{(-1)^n}{n}[/tex]

has a limit (and it equals log(2)).
Actually [itex]-\ln 2[/itex].

micromass
#4
Apr28-12, 04:53 PM
Mentor
micromass's Avatar
P: 18,299
Do Alternating Series Have Limits?

Quote Quote by Dickfore View Post
Actually [itex]-\ln 2[/itex].
Ah yes, thank you!!
I'm Awesome
#5
Apr28-12, 05:13 PM
P: 12
I'm still kinda confussed, so for example if I have the series:
[tex]\sum_{n=1}^{+\infty} (-1)^n \frac{n}{n+1}[/tex]


Is the limit nonexistant or does it equal 1?
micromass
#6
Apr28-12, 05:14 PM
Mentor
micromass's Avatar
P: 18,299
Quote Quote by I'm Awesome View Post
I'm still kinda confussed, so for example if I have the series:
[tex]\sum_{n=1}^{+\infty} (-1)^n \frac{n}{n+1}[/tex]


Is the limit nonexistant or does it equal 1?
That limit doesn't exist since

[tex]\lim_{n\rightarrow +\infty}{ \frac{(-1)^n n}{n+1}}\neq 0[/tex]
I'm Awesome
#7
Apr28-12, 05:19 PM
P: 12
Okay. So in the previos example where the limit was equal -ln2 , do I have to solve for and indeterminate power to get that answer?
micromass
#8
Apr28-12, 05:21 PM
Mentor
micromass's Avatar
P: 18,299
Quote Quote by I'm Awesome View Post
Okay. So in the previos example where the limit was equal -ln2 , do I have to solve for and indeterminate power to get that answer?
No. You get that answer if you find the Taylor series for the logarithm.
Useful nucleus
#9
Apr30-12, 11:07 AM
Useful nucleus's Avatar
P: 261
In general , an alternating series of the form [itex]\sum[/itex] (-1)k ak will converge if ak [itex]\rightarrow[/itex]0 as k[itex]\rightarrow[/itex]∞ and 0<ak+1≤ ak
HallsofIvy
#10
Apr30-12, 09:19 PM
Math
Emeritus
Sci Advisor
Thanks
PF Gold
P: 39,553
If [itex]a_n[/itex] goes to 0 then the series [itex]\sum (-1)^na_n[/itex] converges.
If it does not, then we can determine two subseries of [itex]\sum (-1)^n a_n[itex], one with n even, the other with n odd, that converge to two different limits. And so the series itself does not converge.
DonAntonio
#11
Apr30-12, 09:53 PM
P: 606
Quote Quote by HallsofIvy View Post
If [itex]a_n[/itex] goes to 0 then the series [itex]\sum (-1)^na_n[/itex] converges.



*** The convergence to zero of [itex]a_n[/itex] must be monotone, otherwise the Leibnitz test may fail.

DonAntonio ***


If it does not, then we can determine two [b]subseries[b] of [itex]\sum (-1)^n a_n[itex], one with n even, the other with n odd, that converge to two different limits. And so the series itself does not converge.
....
Useful nucleus
#12
Apr30-12, 09:57 PM
Useful nucleus's Avatar
P: 261
Quote Quote by HallsofIvy View Post
If [itex]a_n[/itex] goes to 0 then the series [itex]\sum (-1)^na_n[/itex] converges.
Not necessary for the following:
Consider [itex]\sum[/itex]k=1 (-1)k+1 ak where ak= 1/k if k is odd and 1/k2 if k is even. It is possible to show that this alternating series diverge to +∞ although ak goes to zero.
This counter example indicates the necessity of the condition 0<ak+1 ≤ ak for convergence to happens.
Mark44
#13
Apr30-12, 10:36 PM
Mentor
P: 21,295
Quote Quote by I'm Awesome View Post
I'm still kinda confussed, so for example if I have the series:
[tex]\sum_{n=1}^{+\infty} (-1)^n \frac{n}{n+1}[/tex]


Is the limit nonexistant or does it equal 1?
First off, it's not clear to me whether you're asking about the limit of the terms in the series or the limit of the series itself.

If you're asking about the limit of the terms, then the limit does not exist.
$$\lim_{n \to \infty} (-1)^n \frac{n}{n + 1}\text{ does not exist}$$

The reason is that for large n, successive terms oscillate between values close to 1 and -1, depending on whether n is even or odd, which affects the sign of (-1)n.

If you're asking about the sum of the series, then there too the limit does not exist. The Nth Term Test for Divergence says that if the limit of the terms of the series is different from zero or doesn't exist, then the series diverges. Since I established that the limit of the terms of the series doesn't exist, this theorem says that the series diverges (does not converge).


Register to reply

Related Discussions
Alternating Power Series - Limits Calculus & Beyond Homework 8
How to tell if this series converges by alternating series test Introductory Physics Homework 1
Help with alternating series sum Calculus & Beyond Homework 2
Convergence/Divergence of Given Series Using Alternating Series test. Calculus & Beyond Homework 2
Alternating series Calculus & Beyond Homework 3