Take the 2-minute tour ×
Mathematica Stack Exchange is a question and answer site for users of Mathematica. It's 100% free, no registration required.

I need the series expansion of a fairly nasty function and its derivative:

f = (x - a) (1 + b (1/(x - a))^((5 + Sqrt[73])/12) *
   Hypergeometric2F1[(-7 + Sqrt[73])/12, (1 + Sqrt[73])/12, (
   6 + Sqrt[73])/6, -(a/4)/(x - a)]) + a/2

$f\left(x\right) = (x-a) \left(1+b \left(\frac{1}{x-a}\right)^{\frac{1}{12} \left(5+\sqrt{73}\right)} \, _2F_1\left(\frac{-7+\sqrt{73}}{12} ,\frac{1+\sqrt{73}}{12} ;\frac{6+\sqrt{73}}{6} ;-\frac{a}{4 (x-a)}\right)\right)+\frac{a}{2}$

where $x$, $a$ and $b$ are positive real numbers and I know $x \gg a$. I'm interested in the series expansion of $f$ and $f'$ in inverse powers of $x$:

Series[f, {x, \[Infinity], 1}] // Normal

$\dfrac{b \left(\dfrac{1}{x-a}\right)^{\dfrac{\sqrt{73}}{12}}}{\left(\dfrac{1}{x}\right)^{7/12}}-\dfrac{a}{2}+x$

Series[D[f, x], {x, \[Infinity], 1}] // Normal

$b \left(\dfrac{1}{x}\right)^{5/12} \left(\dfrac{1}{x-a}\right)^{\dfrac{\sqrt{73}}{12}}+1$

The problem is I can't figure out how to get this to simplify any further. I need to get these in the form

$ c_1 x + c_0 + \dfrac{c_{-1}}{x^{p}} $

for some constants $c_i$ and $p$, but Series doesn't expand the term with $x-a$ in the denominator. I guess because the power is irrational? How can I get the answer in the desired form? I could do it by pen and paper of course, but I would love to have it in code. Using Assumptions -> {x > a > 0, b > 0} doesn't seem to make any difference.

Any help is appreciated.

share|improve this question
add comment

2 Answers

up vote 10 down vote accepted

After you define your function execute this line:

Series[f, {x, \[Infinity], 1}] // Normal // PowerExpand

enter image description here

Following your $x \gg a$ we can rewrite it as

% /. (-a + x) -> x // FullSimplify

enter image description here

which gives:

enter image description here

Now you have to be very careful dealing with derivative. It is not enough to keep 0th and 1st terms (as you did) - you will loose information. It is very interesting that to get correct result you need to keep as much as 4 terms:

Series[D[f, x], {x, \[Infinity], 3}] // Normal // PowerExpand // FullSimplify

enter image description here

Carefully examining the above expression with respect to $x \gg a$ we rewrite it as

% /. a -> 0 // FullSimplify

enter image description here

which gives:

enter image description here

And now it is correct because derivative of the function series is equal to the series of the derivative:

enter image description here

share|improve this answer
    
Thanks. PowerExpand helps. Is there a way to set a->0 in the (-a+x) term only? Sorry if it's dumb, I'm fairly amateur at this. –  Michael Aug 30 '12 at 4:33
    
@Michael I now should it in the answer. See update. –  Vitaliy Kaurov Aug 30 '12 at 4:47
    
Ahh... Seems obvious now. :) Thanks for the help. –  Michael Aug 30 '12 at 4:54
    
@Michael I corrected an error. Please read again. –  Vitaliy Kaurov Aug 30 '12 at 5:42
    
You're right about testing the derivative of the series against the series of the derivative of course. What I neglected to mention was that my a and b are rather complicated expressions themselves, so showing four terms in the series takes up more than a screenful for me. :) But the truncation error seems to be under control. Thanks for all the help. –  Michael Aug 30 '12 at 7:10
add comment

You can do a series expansion about $a=0$.

Series[f, {a, 0, 1}] // Normal // PowerExpand // Collect[#, x, Simplify] &

$$ -\frac{a}{2}+x+b x^{\frac{1}{12} \left(7-\sqrt{73}\right)}-\frac{\left(-91+\sqrt{73}\right) a b x^{-1+\frac{1}{12} \left(7-\sqrt{73}\right)}}{48 \left(6+\sqrt{73}\right)} $$

Series[D[f, x], {a, 0, 1}] // Normal // PowerExpand // Collect[#, x, Simplify] &

$$ 1-\frac{\left(191+43 \sqrt{73}\right) a b x^{-2+\frac{1}{12} \left(7-\sqrt{73}\right)}}{288 \left(6+\sqrt{73}\right)}-\frac{1}{12} \left(-7+\sqrt{73}\right) b x^{-1+\frac{1}{12} \left(7-\sqrt{73}\right)} $$

Then you can manually discard the $O(\frac{a}{x})$ correction to the $x^{-p}$ term.

I did a little bit more exploration, and found that if you series expand to higher order in $a$ then the extra terms are all $O((\frac{a}{x})^n)$ corrections to the $x^{-p}$ term.

share|improve this answer
add comment

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

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