7
$\begingroup$

The problem is, as the title suggests, to find the Power Series Expansion of $\frac{1}{1- \cos x}$ around $x=c$.

What I've tried:

  • Direct Computation: Derivatives get very ugly quickly, and don't yield a nice formula that I can recognize as a "series."
  • Tried finding the integral of $\frac{1}{1- \cos x}$, finding it's series and then differentiating it to get the new series.
  • Tried reverse of the above, differentiating and finding it's series, then integrating (very messy).
  • Then I tried some substitution "tricks", like using the series of $\frac{1}{1-x}$ and then plugging in the series expansion for $\cos x$, but that's a double sum that I struggled to produce anything useful from :$\displaystyle \sum_{k=0}^\infty\left(\sum_{n=0}^\infty\frac{(-1)^nx^{2n}}{(2n)!}\right)^k$

I am literally at my witts end with this problem. I have spent perhaps a day or two trying to figure it out, because I feel that I am so close - but just barely missing something. I do not want the solution posted - now it's personal and I have to figure it out, but I would greatly appreciate a hint in the right direction, or to point out a mistake that I may be overlooking.

$\endgroup$
11
  • $\begingroup$ Is it 1/cos, as in the title, or 1/(1-cos) as in the first sentence? $\endgroup$
    – coffeemath
    Mar 7, 2014 at 20:55
  • $\begingroup$ Sorry, corrected it - but you beat me to it. It's 1/(1-cos). $\endgroup$
    – user133846
    Mar 7, 2014 at 20:56
  • $\begingroup$ Also, if you want the expansion around the (arbitrary) point $x=c$ it will be very complicated. Are you sure you don't mean to expand around $x=0$? [Looks by your attempts you do mean expand around $0$.] $\endgroup$
    – coffeemath
    Mar 7, 2014 at 20:59
  • 2
    $\begingroup$ $ 1 - \cos x = 2 \sin^2 \frac{x}{2} $ $\endgroup$
    – hjpotter92
    Mar 7, 2014 at 20:59
  • 1
    $\begingroup$ @coffeemath I wish it was indeed expand around $x=0$. But the intent is for the point of expansion to be arbitrary, which understandably makes the problem harder. My attempts may have been flawed in that sense, but I wanted something to start with, and finding it around $x=0$ was one of those "give it a shot." moments. $\endgroup$
    – user133846
    Mar 7, 2014 at 21:01

3 Answers 3

2
$\begingroup$

As @hjpotter92 suggest, you have $$\frac{1}{1-\cos(x)} = \frac{1}{2\sin^2(x/2)} = -\frac{\text{d}}{\text{d}x}(\cot(x/2)).$$ Now you can exploit the series expansion of $$\cot(x)=\sum_{n=0}^\infty\frac{(-1)^n 2^{2n}B_{2n}}{(2n)!}x^{2n-1}, \quad \forall 0<\left|x\right|<\pi.$$ Now, by evaluating in $x/2$, differentiating each coefficient and changing the sign, you get $$\frac{1}{1-\cos(x)} =\sum_{n=0}^\infty\frac{(-1)^{n+1} 2(2n-1)B_{2n}}{(2n)!}x^{2n-2}, \quad \forall 0<\left|x\right|<2\pi,$$ where $B_n$ are the Bernoulli numbers.

$\endgroup$
6
  • $\begingroup$ This works for $c=0$. I don't get anything that doesn't look messy for general $c$. Do you get anything that looks decent? $\endgroup$
    – robjohn
    Mar 7, 2014 at 21:53
  • $\begingroup$ actually this series expansion holds $\forall x \in (0,2 \pi)$, because the expansion of $\cot(x)$ holds $\forall x \in(0,\pi)$, and when you substitute $x$ with $x/2$ the $\pi$ becomes $2\pi$ $\endgroup$
    – Paglia
    Mar 7, 2014 at 21:58
  • 1
    $\begingroup$ What the OP means by "around $x=c$" is expanding in a power series in $x-c$; i.e. $$\sum_{k=0}^\infty a_k(x-c)^k$$ $\endgroup$
    – robjohn
    Mar 7, 2014 at 22:01
  • 1
    $\begingroup$ Was an initial term $1/x^2$ dropped, since the expansion for $\cot(x)$ begins with $x^{-1}$ which after sign change and derivative would give $1/x^2$? $\endgroup$
    – coffeemath
    Mar 7, 2014 at 22:03
  • $\begingroup$ I think that in this sense the question is probably not answerable. Anyway, way do you need an expansion of this form, if you already have a series that holds $\forall x$ inside your domain? $\endgroup$
    – Paglia
    Mar 7, 2014 at 22:06
0
$\begingroup$

$\frac{1}{1-\cos(x)}=\frac{1+\cos(x)}{1-\cos^2(x)}=\frac{1+\cos(x)}{\sin^2(x)}=\csc^2(x)+\csc(x)\cot(x)$

That last one can be easily noticed to be the derivative of $-(\cot(x)+\csc(x))$. Now the question is a matter of finding the power series expansions of $\cot(x)$ and $\csc(x)$, which are more easily found, and differentiating.

$\endgroup$
-2
$\begingroup$

integration of 1/1-cos(x)=
= 1/1-cos(x)(1+cos(x)/1+cos(x) = 1+cos(x)/[1-cos(x)][1+cos(x)] = 1+cos(x)/[sin(x)][sin(x) = 1/sin(x)[sin(x)]+cos(x)/sin(x)[sin(x)] = cosec(x)[cosec(x)+cot(x)cosec(x) = -cot(x)+cosec(x)+c

$\endgroup$
3
  • 2
    $\begingroup$ Welcome to MSE. Please use MathJax. $\endgroup$ Aug 4, 2017 at 6:53
  • $\begingroup$ mst ho to frnds me share kro $\endgroup$ Aug 4, 2017 at 6:53
  • $\begingroup$ very good answer $\endgroup$ Aug 4, 2017 at 7:34

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.