Facebook Page
Twitter
RSS
+ Reply to Thread
Results 1 to 7 of 7
  1. MHB Craftsman
    ognik's Avatar
    Status
    Offline
    Join Date
    Feb 2015
    Location
    Auckland
    Posts
    471
    Thanks
    306 times
    Thanked
    85 times
    #1
    The Cauchy Ratio test says: If $ \lim_{{n}\to{\infty}}\frac{a_{n+1}}{a_n} < 1 $ then the series converges. OK.

    Now I read that for a power series (of functions of x), the same test also provides the interval of convergence, i.e. If the series converges, then $ \lim_{{n}\to{\infty}}\frac{a_{n+1}}{a_n} = {R}^{-1} $ and the interval is -R < x < R

    Could someone please explain why this works? Thanks.

  2. MHB Oldtimer
    MHB Site Helper
    MHB Math Scholar
    Opalg's Avatar
    Status
    Offline
    Join Date
    Feb 2012
    Location
    Leeds, UK
    Posts
    2,118
    Thanks
    718 times
    Thanked
    5,944 times
    Thank/Post
    2.806
    Awards
    Graduate POTW Award (2016)  

MHB Analysis Award (2016)  

Graduate POTW Award (2015)  

Graduate POTW Award (Jul-Dec 2013)  

MHB Pre-University Math Award (Jul-Dec 2013)
    #2
    Quote Originally Posted by ognik View Post
    The Cauchy Ratio test says: If $ \lim_{{n}\to{\infty}}\frac{a_{n+1}}{a_n} < 1 $ then the series converges. OK.

    Now I read that for a power series (of functions of x), the same test also provides the interval of convergence, i.e. If the series converges, then $ \lim_{{n}\to{\infty}}\frac{a_{n+1}}{a_n} = {R}^{-1} $ and the interval is -R < x < R

    Could someone please explain why this works? Thanks.
    Both those results need to be stated a bit more carefully.

    The Cauchy Ratio test says: If $\color{red}{\sum a_n}$ is a series of positive terms and $ \lim_{{n}\to{\infty}}\frac{a_{n+1}}{a_n} < 1 $ then the series converges.

    If you want to apply the test in a situation (like a power series with $x<0$) where the terms may not all be positive, then you have to take absolute values. The test then says that if $ \lim_{{n}\to{\infty}}\Bigl|\frac{a_{n+1}}{a_n}\Bigr| < 1 $ then the series converges (absolutely).

    For a power series $\sum a_nx^n$, the limit $ \lim_{{n}\to{\infty}}\Bigl|\frac{a_{n+1}}{a_n}\Bigr|$ does not necessarily exist. The correct statement of the result is that if that limit exists and is equal to $R^{-1}$, then $R$ is the radius of convergence of the series.

    The reason why this works is that you can apply the Cauchy Ratio test to see that the series converges if $ \lim_{{n}\to{\infty}}\Bigl|\frac{a_{n+1}x^{n+1}}{a_nx^n}\Bigr| < 1 $. But $\Bigl|\frac{a_{n+1}x^{n+1}}{a_nx^n}\Bigr| = \Bigl|\frac{a_{n+1}}{a_n}\Bigr|\,|\,x\,|$ and so $ \lim_{{n}\to{\infty}}\Bigl|\frac{a_{n+1}x^{n+1}}{a_nx^n}\Bigr| = R^{-1}|\,x\,|.$ Therefore the series converges if $R^{-1}|\,x\,| < 1$, in other words if $|\,x\,| < R.$

  3. MHB Craftsman
    ognik's Avatar
    Status
    Offline
    Join Date
    Feb 2015
    Location
    Auckland
    Posts
    471
    Thanks
    306 times
    Thanked
    85 times
    #3 Thread Author
    Glad to see I was correct about the approach needing the modulus of the ratio, but the point behind R still eludes me.

    For example consider a series $ \displaystyle \sum_{n=1}^{\infty}n^3(x-s)^n,$ s > 0.

    This is convergent and works out to this point: $ \lim_{{n}\to{\infty}} |\frac{a_{n+1} }{a_n}| < 1$, from which I get $s-1 < x < s+1$, ie an interval of (s-1, s+1). This I'm happy with - and I didn't need an R.

    Using $R^{−1}|x-s|<1 \implies |x-s|<R, \therefore s-R < x < s+R$. To me the R seems add unnecessary complexity, I have to assume that R = 1 to get the actual values, so what am I missing?

  4. MHB Craftsman
    ognik's Avatar
    Status
    Offline
    Join Date
    Feb 2015
    Location
    Auckland
    Posts
    471
    Thanks
    306 times
    Thanked
    85 times
    #4 Thread Author
    Quote Originally Posted by Opalg View Post
    ... $= \Bigl|\frac{a_{n+1}}{a_n}\Bigr|\,|\,x\,|$ and so $ \lim_{{n}\to{\infty}}\Bigl|\frac{a_{n+1}x^{n+1}}{a_nx^n}\Bigr| = R^{-1}|\,x\,|.$
    Hi Opalg, I just don't follow this step?

    Also is what I did wrong? - $ \lim_{{n}\to{\infty}} |\frac{a_{n+1} }{a_n}| < 1$, from which I get $s-1 < x < s+1$, ie an interval of (s-1, s+1).

  5. MHB Master
    MHB Math Helper

    Status
    Online
    Join Date
    Jan 2012
    Posts
    1,282
    Thanks
    130 times
    Thanked
    2,042 times
    Thank/Post
    1.593
    #5
    Quote Originally Posted by ognik View Post
    Hi Opalg, I just don't follow this step?

    Also is what I did wrong? - $ \lim_{{n}\to{\infty}} |\frac{a_{n+1} }{a_n}| < 1$, from which I get $s-1 < x < s+1$, ie an interval of (s-1, s+1).
    You have gotten the INTERVAL of convergence (although you still need to check the endpoints) but you were asked for the RADIUS of convergence.

    Since it seems you have gotten $\displaystyle \begin{align*} \left| x - s \right| < 1 \end{align*}$ (I haven't checked myself) then that would mean that the series converges in the circle centred at $\displaystyle \begin{align*} (x, y) = (s, 0) \end{align*}$ and thus the radius of convergence is 1. (I say circle because it could be a complex value...)

  6. MHB Craftsman
    ognik's Avatar
    Status
    Offline
    Join Date
    Feb 2015
    Location
    Auckland
    Posts
    471
    Thanks
    306 times
    Thanked
    85 times
    #6 Thread Author
    OK, thanks I was mixing those up, but I still don't follow that step in Opalg's post - where R appears....?

  7. MHB Master
    MHB Math Helper

    Status
    Online
    Join Date
    Jan 2012
    Posts
    1,282
    Thanks
    130 times
    Thanked
    2,042 times
    Thank/Post
    1.593
    #7
    Quote Originally Posted by ognik View Post
    OK, thanks I was mixing those up, but I still don't follow that step in Opalg's post - where R appears....?
    Sigh... You need to get it to a point where $\displaystyle \begin{align*} \left| x - c \right| < R \end{align*}$, or equivalently $\displaystyle \begin{align*} R^{-1} \, \left| x - c \right| < 1 \end{align*}$...

Similar Threads

  1. WolframAlpha confusion...
    By Taryn in forum Mathematics Software and Calculator Discussion
    Replies: 2
    Last Post: August 27th, 2015, 23:02
  2. uniform convergence confusion
    By Fermat in forum Analysis
    Replies: 3
    Last Post: July 21st, 2014, 09:44
  3. Replies: 1
    Last Post: January 29th, 2014, 16:22
  4. Pointwise convergence implies uniform convergence
    By Siron in forum Topology and Advanced Geometry
    Replies: 1
    Last Post: January 29th, 2014, 16:22
  5. [SOLVED] Confusion about e^ipi+1=0 ??
    By I like Serena in forum Analysis
    Replies: 9
    Last Post: January 30th, 2013, 10:00

Tags for this Thread

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •  
Math Help Boards