0
$\begingroup$

I am computing the 4th order talyor approximation of this function $$\big(\frac{b}{g-x}\big)^{0.25}$$

The analytic textbook result is:

$$b^{0.25} \big(\frac{1}{g^{0.25}}+\frac{0.25 x}{g^{1.25}}+\frac{0.15625x^2}{g^{2.25}}+\frac{0.117188 x^3}{g^{3.25}}+\frac{0.0952148 x^4}{g^{4.25}}\big)$$

When I do it in Mathematica:

nonLin = (b/(g - x))^0.25
taylorLin = Normal[Series[nonLin , {x, 0, 4}] // Simplify] 

I get this:

$$x^4 \left(\frac{0.126465 b^4}{g^8 \left(\frac{b}{g}\right)^{3.75}}-\frac{0.03125 \left(\frac{b}{g}\right)^{0.25}}{g^4}\right)+\frac{0.117188 b^3 x^3}{g^6 \left(\frac{b}{g}\right)^{2.75}}+\frac{b x^2 \left(\frac{0.25 g}{\left(\frac{b}{g}\right)^{0.75}}-\frac{0.09375 b}{\left(\frac{b}{g}\right)^{1.75}}\right)}{g^4}+\frac{0.25 x \left(\frac{b}{g}\right)^{0.25}}{g}+\left(\frac{b}{g}\right)^{0.25}$$

Is there a method to obtain the same nice representation as the analytic textbook solution ? ... to simplify it somehow?

EDIT: If I use 1/4 in the exponent, the reuslt looks much nicer. However, it could still be simplified:

nonLin = (b/(g - x))^(1/4)
taylorLin = Normal[Series[nonLin , {x, 0, 4}] // Simplify] 

$$\frac{195 x^4 \sqrt[4]{\frac{b}{g}}}{2048 g^4}+\frac{15 x^3 \sqrt[4]{\frac{b}{g}}}{128 g^3}+\frac{5 x^2 \sqrt[4]{\frac{b}{g}}}{32 g^2}+\frac{x \sqrt[4]{\frac{b}{g}}}{4 g}+\sqrt[4]{\frac{b}{g}}$$

$\endgroup$
  • 1
    $\begingroup$ Use 1/4 instead of 0.25 in the exponent. $\endgroup$ – J. M. is away Oct 3 '18 at 10:01
  • $\begingroup$ The question is not clear: the solution you have shown is already analytic. Can you show, what are you after? $\endgroup$ – Alexei Boulbitch Oct 3 '18 at 10:08
  • $\begingroup$ @J.M.issomewhatokay. Thanks a lot for your comment. Please have a look at my edit. $\endgroup$ – james Oct 3 '18 at 10:09
  • $\begingroup$ @AlexeiBoulbitch I want the output to look like the analytic equation from the text book.( I edited my question) $\endgroup$ – james Oct 3 '18 at 10:10
  • $\begingroup$ Aha, then use please the advice of @J. M. is somewhat okay given above. $\endgroup$ – Alexei Boulbitch Oct 3 '18 at 10:19
2
$\begingroup$

Try this:

nonLin = (b/(g - x))^0.25;
taylorLin1 = 
  Simplify[Normal[Series[nonLin, {x, 0, 4}]], {b > 0, g > 0}];
taylorLin2 = b^0.25*Expand[taylorLin1/b^0.25]

The results looks as follows:

enter image description here

Done. Have fun!

$\endgroup$
  • $\begingroup$ Thanks a lot !! $\endgroup$ – james Oct 3 '18 at 11:34
1
$\begingroup$

You can take a constant factor (b/g)^.25 from the sum, then the sum is similar to what's in the book

Normal[Series[(1/(1 - x/g))^.25, {x, 0, 4}]]

Out[]= 1. + (0.25 x)/g + (0.15625 x^2)/g^2 + (0.117188 x^3)/g^3 + (
 0.0952148 x^4)/g^4
$\endgroup$
1
$\begingroup$

This is using the reply found Factorization here.

It is not giving the EXACT same result, but it is a matter of just looking the output and perform some minimal algebra in the end or just simplifying terms by hand in the end.

nonLin = (b/(g - x))^(1/4)
taylorLin = Series[nonLin, {x, 0, 4}]
sfactor[k_, p_, func_] := 
 HoldForm[StandardForm[k]]*StandardForm[func@(p*1/k)]
sfactor[(b)^(1/4), taylorLin, Apart] // Normal

Hopefully helpful.

Cheers!!!

$\endgroup$
1
$\begingroup$

Instead of using Simplify afterwards, you can provide an assumption to Series:

Series[(b/(g - x))^0.25, {x, 0, 4}, Assumptions->b>0&&g>0] //TeXForm

$b^{0.25`} \left(\frac{1}{g}\right)^{0.25`}+0.25` b^{0.25`} \left(\frac{1}{g}\right)^{1.25`} x+0.15625` b^{0.25`} \left(\frac{1}{g}\right)^{2.25`} x^2+0.1171875` b^{0.25`} \left(\frac{1}{g}\right)^{3.25`} x^3+0.09521484375` b^{0.25`} \left(\frac{1}{g}\right)^{4.25`} x^4+O\left(x^5\right)$

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.