1. Limited time only! Sign up for a free 30min personal 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 approximations

  1. Jul 10, 2009 #1
    1. The problem statement, all variables and given/known data

    1. Use Taylor's Theorem to determine the accuracy of the approximation.

    [tex] arcsin(0.4) = 0.4 + \frac{(0.4)^{3}}{2*3}}[/tex]

    2. Determine the degree of the Maclaurin polynomial required for the error in the approximation of the function at the indicated value fo x to be less than 0.0001. Use a symbolic differentiation utility to obtain and evaluate the required derivatives.

    [tex]f(x) = ln(x+1)[/tex] approximate f(1.5)

    2. Relevant equations



    3. The attempt at a solution

    1.
    [tex]f(x) = arcsin (0.4)[/tex]
    [tex] x = 0.4[/tex]
    [tex]a = 0.5[/tex]
    [tex]N = 3[/tex]

    [tex]R_{n}(x) = \frac{f^{N+1}(c)}{(N+1)!}(x-a)^{N+1}[/tex]
    [tex]R_{3}(0.4) = \frac{f^{4}(c)}{4!}(0.4-0.5)^{4}[/tex]
    [tex]= \frac{\frac{-3c(2c^{2}+3)}{(c^{2}-1)^{3}\sqrt{1-c^{2}}}}{4!}(0.4-0.5)^{4}[/tex] [tex]\leq[/tex] [tex]\frac{(0.4-0.5)^{4}}{4!} = 4.166666667 * 10^{-6}[/tex]

    Book Answer: [tex]R_{3} \leq 7.82 * 10^{-3}[/tex]


    No idea what I did wrong here.

    2.
    [tex]f(x) = ln(x+1)[/tex]
    [tex]a = 0[/tex]
    [tex]x = 1.5[/tex]
    [tex]N = ?[/tex]

    [tex]R_{n}(x) = \frac{f^{N+1}(c)}{(N+1)!}(x-a)^{N+1}[/tex]
    [tex]R_{n}(x) = \frac{f^{N+1}(c)}{(N+1)!}(1.5)^{N+1}[/tex]
    [tex]|R_{n}(1.5)| \leq 0.0001[/tex]

    To make a long sheet of work short, I got all the way up to N=12 before I finally got to 0.0001. Here is the work for N=12:

    N=12: [tex]|R_{12}(1.5)| = |\frac{f^{13}(c)}{13!}(1.5)^{13}| = |\frac{(479001600)}{(c+1)^{13}} * \frac{(1.5)^{13}}{13!}| = |\frac{(479001600)(3.1254*10^{-8})}{(c+1)^{13}}| = \frac{14.9707}{(c+1)^{13}} [/tex]

    [tex]\frac{14.9707}{(1.5+1)^{13}} = 0.0001[/tex]

    The book answer says N=9, for which when I tested I got this:

    N=9: [tex]|R_{9}(1.5)| = |\frac{f^{10}(c)}{10!}(1.5)^{10}| = |\frac{(-362880)}{(c+1)^{10}} * \frac{(1.5)^{10}}{10!}| = |\frac{(-362880)(0.000106)}{(c+1)^{10}}| = \frac{5.7665}{(c+1)^{10}} [/tex]

    [tex]\frac{5.77665}{(1.5+1)^{10}} = 0.000605[/tex] which isn't exactly less than 0.001.


    When I went back through my notes to see if I did something wrong, I realized that since the function was decreasing, (c+1), on the bottom of the fraction, maybe I should have plugged in 0, since c should be greater than or equal to 0 and less than or equal to 1.5, thus the biggest R could be would be whatever I get when I plug 0 in, not 1.5.

    However, if I plug in 0 for these instead of 1.5, it seems to make the problem worse, since as you can see, N=9, with 0 instead of 1.5 for C would get 1 on the bottom and thus R_9 would be 5.77665, which is very very far from 0.0001, much farther than when I used 1.5.

    I'm thoroughly frustrated by this one. Please help!
     
  2. jcsd
  3. Jul 11, 2009 #2
    You chose a=0.5. Probably that is a good choice. But then what would be the Taylor series? Not the one shown. So the writers of the problem did not choose a=0.5. What did they choose? The point is that the Taylor series for their choice is much simpler to write down than for your choice a=0.5.
     
  4. Jul 12, 2009 #3
    I think they chose a=0? Is that right?
     
  5. Jul 12, 2009 #4
    a = 0 is the way to get that Taylor series, yes. So do the rest of the problem on that basis.
     
Know someone interested in this topic? Share this thread via Reddit, Google+, Twitter, or Facebook




Similar Discussions: Taylor approximations
  1. Taylor approximation (Replies: 2)

  2. Taylor approximation (Replies: 3)

  3. Taylor Approximation (Replies: 5)

Loading...