Radius of Convergence for a Power Series with Ratio Test

Join the discussion
Registration is free. Start your own thread to ask a follow-up.
6 replies · 2K views
Saladsamurai
Messages
3,009
Reaction score
7

Homework Statement



Determine the radius of Convergence using the ratio test of:

[tex]\sum_o^{\infty}\frac{n^6}{3^n+n}(x+4)^{8n+1}\qquad(1)[/tex]

Homework Equations



[tex]R = \frac{1}{\lim_{n\rightarrow\infty}\left|\frac{a_{n+1}}{a_n}\right|}\qquad(2)[/tex]

The Attempt at a Solution



Ok. In order to use (2), we must first put (1) into standard form: [itex]\sum a_n(x - x_o)^n[/itex].

I am following a hint that I should let m = 8n +1 however I am not sure what to do with the summation limits? If m = 8n +1, then at n = 0, m = 1. So do I just run the sum from 1 to [itex]\infty[/itex]? And replace n everywhere with n = (m - 1)/8 ?

Thanks!
 
Physics news on Phys.org
I don't understand why we can't just consider the convergence of the series:

[tex]\sum_{n=0}^{\infty} a_n[/tex]

with:

[tex]a_n=\frac{n^6}{3^n+n}(x+4)^{8n+1}[/tex]

and using the radio test with:

[tex]\lim_{n\to\infty}\left|\frac{a_{n+1}}{a_n}\right|[/tex]

If that's valid, I get:

[tex]R=3^{1/8}[/tex]
 
Saladsamurai said:

Homework Statement



Determine the radius of Convergence using the ratio test of:

[tex]\sum_o^{\infty}\frac{n^6}{3^n+n}(x+4)^{8n+1}\qquad(1)[/tex]

Homework Equations



[tex]R = \frac{1}{\lim_{n\rightarrow\infty}\left|\frac{a_{n+1}}{a_n}\right|}\qquad(2)[/tex]


The Attempt at a Solution



Ok. In order to use (2), we must first put (1) into standard form: [itex]\sum a_n(x - x_o)^n[/itex].

I am following a hint that I should let m = 8n +1 however I am not sure what to do with the summation limits? If m = 8n +1, then at n = 0, m = 1. So do I just run the sum from 1 to [itex]\infty[/itex]? And replace n everywhere with n = (m - 1)/8 ?

Thanks!

This is geting silly:

I have that

[tex] \frac{a_{m+1}}{a_m} = \left(\frac{m}{8}\right)^6\left(\frac{8}{m-1}\right)^6<br /> \left(\frac{3^{\frac{m-1}{8}}+\frac{m-1}{8}}{3^{m/8}+m/8}\right)[/tex]

[tex]= \frac{a_{m+1}}{a_m} = \left(\frac{m}{m-1}\right)^6\left(\frac{3^{\frac{m-1}{8}}+\frac{m-1}{8}}{3^{m/8}+m/8}\right)[/tex]

I cannot see how to reduce this further. Any ideas? I'm lookin' at you jackmell :wink:
 
jackmell said:
I don't understand why we can't just consider the convergence of the series:

[tex]\sum_{n=0}^{\infty} a_n[/tex]

with:

[tex]a_n=\frac{n^6}{3^n+n}(x+4)^{8n+1}[/tex]

and using the radio test with:

[tex]\lim_{n\to\infty}\left|\frac{a_{n+1}}{a_n}\right|[/tex]

If that's valid, I get:

[tex]R=3^{1/8}[/tex]

I'm not entirely sure about that. Also, numerically, looks like the radius of convergence appears to be one over that:

[tex]\frac{1}{3^{1/8}}[/tex]

but I can't seem to get to that value.
 
Saladsamurai said:
[tex]= \frac{a_{m+1}}{a_m} = \left(\frac{m}{m-1}\right)^6\left(\frac{3^{\frac{m-1}{8}}+\frac{m-1}{8}}{3^{m/8}+m/8}\right)[/tex]

I cannot see how to reduce this further. Any ideas? I'm lookin' at you jackmell :wink:

We can do that. That first term has a limit of one right (it approaches m/m)^6). Now look at the [itex](m-1)/8[/itex] in the numerator and the [itex]m/8[/itex] in the denominator. Both of those are O(m) but pale in comparison to the other terms having exponential order so for very large m, we can neglect them and just consider:

[tex]\lim_{m\to\infty} \frac{3^{1/8(m-1)}}{3^{m/8}}[/tex] but when we add the exponents we get:

[tex]\frac{m-1}{8}-\frac{m}{8}=-1/8[/tex] so that:

[tex]\lim_{m\to\infty} \frac{3^{1/8(m-1)}}{3^{m/8}}=3^{-1/8}[/tex]
 
Last edited:
jackmell said:
We can do that. That first term has a limit of one right (it approaches m/m)^6). Now look at the [itex](m-1)/8[/itex] in the numerator and the [itex]m/8[/itex] in the denominator. Both of those are asymptotic to m. But both pale in comparison to the other terms so for very large m, we can neglect them and just consider:

[tex]\lim_{m\to\infty} \frac{3^{1/8(m-1)}}{3^{m/8}}[/tex] but when we add the exponents we get:

[tex]\frac{m-1}{8}-\frac{m}{8}=-1/8[/tex] so that:

[tex]\lim_{m\to\infty} \frac{3^{1/8(m-1)}}{3^{m/8}}=3^{-1/8}[/tex]

Ok great. I follow your qualitative description just fine. Now I woud like to flesh out the details using appropriate limit rules.

We have a limit of a product which equals the product of the limits. The first factor

[tex]\left(\frac{m}{m-1}\right)^6[/tex]

is a rational function of two polynomials of the same order, so its limit is simply the ratio of the coefficients of the leading terms in numerator and denominator: 1/1 = 1.

Now as for the second factor:

[tex]\left(\frac{3^{\frac{m-1}{8}}+\frac{m-1}{8}}{3^{m/8}+m/8}\right)[/tex]

I am trying to pinpoint a nice rule that I can apply to back the statement that the exponentials are 'more important.'

Any ideas? Thanks for your help so far :smile:
 
It's all a matter of order, or big-O:

[tex]O(3^n)+O(n)\sim O(3^n)[/tex]

That is, for very big n, that sum asymptotically approaches the [itex]3^{n}[/tex] term as the smaller-ordered term becomes less and less significant.<br /> <br /> So that:<br /> <br /> [tex]\lim_{n\to\infty}\left(\frac{O(3^n)+O(n)}{O(3^{n-1})+O(n)}\right)\sim \frac{O(3^{n})}{O(3^{n-1})}[/tex]<br /> <br /> Essentially, as n grows very large, only the highest-ordered terms matter.<br /> <br /> I may not have that exactly right but I think it's close.[/itex]