1. Not finding help here? Sign up for a free 30min tutor trial with Chegg Tutors
    Dismiss Notice
Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

Taylor Series

  1. Apr 5, 2007 #1
    1. The problem statement, all variables and given/known data

    Give the Taylor Series for exp(x^3) around x = 2.

    2. Relevant equations

    f(x) = Sum[f(nth derivative)(x-2)^n]/n!

    3. The attempt at a solution

    I know the solution for e^x but can't seem to find a formula for the nth derivative of exp(x^3) around x = 2.

    Thanks for any help.
     
  2. jcsd
  3. Apr 5, 2007 #2
    You know the formula for the taylor series expansion of e^y right?

    Just substitute y=x^3 into that expansion and voila you have the formula for e^(x^3)
     
  4. Apr 5, 2007 #3
    I do now that the trick you mentioned works around x = 0; however, I am highly suspicious that this works for x around other than 0. Please show me explicitely why if you still think as you said.
     
  5. Apr 5, 2007 #4
    It works just fine, but you get a Taylor series expansion in (x^3-2)^n, not (x-2)^n. It's still a Taylor series though.

    e^(x^3)=e^2 sum(n=0..infinity) (1/n!)(x^3-2)^n

    Edit- the correct expression about x=2 is
    e^(x^3)=e^8 sum(n=0..infinity) (1/n!)(x^3-8)^n
     
    Last edited: Apr 5, 2007
  6. Apr 5, 2007 #5
    I think your series would actually give the expansion for e^[(x-2)3] and not for e^[x3-2].
     
  7. Apr 5, 2007 #6
    Neither, it gives the T series expansion for e^(x^3)

    Edit- sorry I should have replaced 2 with 8 in the above equation. Then it works.
     
    Last edited: Apr 5, 2007
  8. Apr 5, 2007 #7
    christian how did you get the e^2?
     
  9. Apr 5, 2007 #8
    No it still does not (x-2)^3 is not the same as x3-8.
     
  10. Apr 5, 2007 #9
    d/dn exp(x) = exp(x)

    so when x=2, exp(x)=e^2

    However, when you substitute x=(x')^3, you have to change to e^8
     
  11. Apr 5, 2007 #10
    I never said it was!
     
  12. Apr 5, 2007 #11
    oh, I see. But then are saying that exp(2^3) is the nth derivative of f around x = 2? Seems strange since for instance, first f derivative = 3x^2exp(x^3). When you substitute x = 2, the above equation is unequal to exp(2^3) as you imply. Not to mention the other derivatives: all unequal to exp(2^3).
     
  13. Apr 5, 2007 #12
    Here- I'll go through the steps

    e^x = e^a sum (x-a)^n / n!

    so

    e^(x^3)=e^a sum(x^3-a)^n / n!

    = e^(b^3) sum (x^3-b^3)^n / n!

    when x=b the RHS is equal to e^(b^3), thus the T series is evaluated about x=b. So for b=2 we have

    e^(x^3)= e^8 sum (x^3-8)^n / n!
     
  14. Apr 5, 2007 #13
    "The derivative of e^(x^3) is not e^(x^3) it is (3x^2)(e^(x^3)"

    Where did I say otherwise?
     
  15. Apr 5, 2007 #14
    I see what you are doing but doesn't the last equation you give has the (x^3-8)^n, when it should be somthing like (x^3 - 2)^n, for T be around x = 2?
     
  16. Apr 5, 2007 #15
    No, for the reasons I gave above. Look at the value of the expansion when x=2.

    The reason for all the confusion is because I'm doing a T series expansion in x^3, and not in x. To do a T series expansion in x might be much more difficult, but the Q doesn't require an expansion in powers of x, though it may be implied.
     
  17. Apr 5, 2007 #16
    yeah, you might be right, but it is definitely the first time I see it as
    (x^3 - 2^3)^n for T be about x = 2. Seems unusual.
     
  18. Apr 5, 2007 #17
    First time I've seen it too! It works though, looking at some graphing software.
     
  19. Apr 5, 2007 #18
    Thanks a lot for all the help. I am convinced.
     
  20. Apr 5, 2007 #19
    Aha, I had a brainwave!

    x^3=(x-2+2)^3
    x'=x-2, and x' is small near x=2
    so x^3=(x'+2)^3
    =8(x'/2+1)^3
    =(approx)=8(3x'/2+1)
    =12x'+8
    =12(x-2)+8

    so given that x^3 = approx= 12(x-2)+8 about x=2

    and

    e^(x^3)=e^8 sum(n=0..infinity) (1/n!)(x^3-8)^n

    By substitution we have

    e^(x^3)=e^8 sum(n=0..infinity) (1/n!)(12(x-2))^n

    Which, again looking at my graphing program, works. So, this is a nicer answer, because the expansion is in (x-2). However, it would be nice to have a more concise proof.
     
    Last edited: Apr 5, 2007
  21. Apr 5, 2007 #20
    OK, here's the concise version....

    e^(x^3) about x=2
    let x'=x-2
    then e^(x^3)=e^[(x'+2)^3]
    =e^[8(x'/2+1)^3]
    =e^[12x'+8] (approximately)
    =e^8 e^12x'
    =e^8 sum{n=0..infinity} (1/n!)[12(x-2)]^n

    Edit: Aaarrgghh! it's not quite exact, because you also need to expand (x'/2+1)^3 to higher powers in x'. So you really end up with a sum over n and m, where m is a Taylor series expansion of (x'/2+1)^3 about x'=0 in powers of x'^m.

    It's messy, but doable.

    I wonder if there's a way to only sum over one index.
     
    Last edited: Apr 5, 2007
Know someone interested in this topic? Share this thread via Reddit, Google+, Twitter, or Facebook

Have something to add?



Similar Discussions: Taylor Series
  1. Taylor series (Replies: 3)

  2. Taylor series (Replies: 14)

  3. Taylor Series (Replies: 1)

  4. Taylor Series (Replies: 2)

  5. Taylor series (Replies: 15)

Loading...