I know that $$\cos(z^4)-1=-\frac{z^8}{2!}+\frac{z^{16}}{4!}+...$$ but how do I take the reciprocal of this series (please do not use little-o notation)? Or are there better methods to obtain the required series?
-
1A follow-up? Still working on this function? Is this homework? – Julien Mar 26 '13 at 04:46
-
@julien Hi Julien! I decided to create a new question because I feel it deserves its own answer. No, it's not homework. If you don't mind to elaborate on the little-O notation usage in the original question, I'd love to read it. – ryang Mar 26 '13 at 05:14
-
Hi Ryan. Definitely, the full Laurent series is a different question which deserves its own answer. O and o are useful in many situations. To begin with limit determination. It allows you to keep the relevant part in a Taylor expansion and to throw away the remainder. Unlike equivalents, O's and o's can be safely added, subtracted, etc... – Julien Mar 26 '13 at 05:24
-
1The reciprocal doesn't help you a lot and I don't see a nice system here – Dominic Michaelis Mar 26 '13 at 05:40
-
@DominicMichaelis How do you mean by not seeing a nice system? And do you mean that the way to obtain the Laurent series isn't by taking the reciprocal as I suggested? If so, then how can I produce the Laurent series? – ryang Mar 27 '13 at 01:18
-
Long division... – GEdgar Mar 27 '13 at 01:47
-
@Ryan I think the reciprocal doesn't help you to find the full taylor series. My first idea would be leipnizformula to get all derivatives but to be honest I guess you won't find an explicit formula (to be more honest I didn't even try after calculating the first 10 derivatives with mathematica the terms got about 100 times to long for me) – Dominic Michaelis Mar 27 '13 at 07:07
-
@DominicMichaelis Hey, did you mean for finding the residue at 0 instead? If so, yes, it's crazy to take derivatives. For this particular problem, to find the resiude at 0 (which is an order-8 pole), long division (though I'm still not comfortable with long division involving infinite series) to first obtain the Laurent series seems to be the best method. – ryang Mar 27 '13 at 10:58
-
@Ryan no i mean for the Laurentseries $$\sum_{n=-8}^\infty a_n z^n $$ the long division won't work well – Dominic Michaelis Mar 27 '13 at 10:59
-
Isn't the series explicit from the formula? Why need to perform division? – ryang Mar 27 '13 at 11:02
4 Answers
Long Division ... still worth learning
To get more terms in the quotient, use more terms in the divisor.

- 111,679
-
Thank you. While I understand the mechanics of the above division, I am unsure why it works, because for polynomial long division I'm accustomed to listing the polynomials in decreasing instead of increasing powers. I also don't know how the radius of convergence of a Laurent series obtained by division is related to that of the original Taylor series (in this case, the series in the denominator). – ryang Mar 27 '13 at 03:55
-
That doesn't help you to get the full laurent series, and when you wan't another term you have to calculate all, I see no advantage to making the taylor series by taking derivates often – Dominic Michaelis Mar 27 '13 at 06:21
-
-
@Ryan, you get the coefficients of the laurent series for $n>$ with $$a_n=\frac{f^{(n)}}{n!}$$ just like the taylor series. – Dominic Michaelis Mar 27 '13 at 11:05
-
And you think computing the 8th derivative of $f(z)$ is easier than the long-division calculation above? – GEdgar Mar 27 '13 at 15:48
Of course the easy way is to ask a CAS. This is from Maple:
$$ (-2) z^{(-8)} - \frac{1}{6} - \frac{1}{120} z^{8} - \frac{1}{3024} z^{16} - \frac{1}{86400} z^{24} - \frac{1}{2661120} z^{32} - \\ \quad{}\quad{}\frac{691}{59439744000} z^{40} - \frac{1}{2874009600} z^{48} - \frac{3617}{355687428096000} z^{56} - \\ \quad{}\quad{}\frac{43867}{150267476975616000} z^{64} - \frac{174611}{21127833228902400000} z^{72} - \\ \quad{}\quad{}\frac{77683}{335740477128376320000} z^{80} - \frac{236364091}{36822263841994427596800000} z^{88} \\ \quad{}\quad{} - \frac{657931}{3722690410399436636160000} z^{96} + \operatorname{O} \bigl(z^{104}\bigr) $$

- 111,679
The function $f(z)=\frac1{\cos(z^4)-1}$ can be rearranged as \begin{align*} f(z)&=-\frac1{1-\cos(z^4)}\\ &=-\frac{1}{2\sin^2\frac{z^4}2}\\ &=-\frac12\biggl(\frac{2}{z^4}\biggr)^2 \Biggl(\frac{\frac{z^4}{2}}{\sin\frac{z^4}{2}}\Biggr)^2\\ &=-\frac{2}{z^8} \Biggl\{1+\sum_{q=1}^{\infty}(-1)^q\Biggl[\sum_{k=1}^{2q}\frac{(2)_k}{k!} \sum_{j=1}^k(-1)^j\binom{k}{j} \frac{T(2q+j,j)}{\binom{2q+j}{j}}\Biggr]\frac{z^{8q}}{(2q)!}\Biggr\},\quad z^4<2\pi, \end{align*} where $(2)_k$ is the Pochhammer symbol or the rising factorial and $T(2q+j,j)$ denotes the central factorial numbers of the second kind. The last step comes from employing Theorem 4.1 in the paper
- Feng Qi and Peter Taylor, Series expansions for powers of sinc function and closed-form expressions for specific partial Bell polynomials, Applicable Analysis and Discrete Mathematics 18 (2024), no. 1, in press; available online at https://doi.org/10.48550/arXiv.2204.05612.

- 1,691
From \begin{equation} \begin{aligned}\label{sin-square-recip-ser} \biggl(\frac{x}{\sin x}\biggr)^2&=1+2B_2x^2-4\sum_{k=2}^{\infty}(-1)^k\Biggl[\bigl(2^{2k-1}-1\bigr)B_{2k}\\ &\quad-\sum_{j=1}^{k-1}\binom{2k}{2j}\bigl(2^{2j-1}-1\bigr)\bigl(2^{2k-2j-1}-1\bigr)B_{2j}B_{2k-2j}\Biggr]\frac{x^{2k}}{(2k)!}\\ &=1+\frac{x^2}{3}+\frac{x^4}{15}+\frac{2 x^6}{189}+\frac{x^8}{675}+\frac{2 x^{10}}{10395}+\frac{1382 x^{12}}{58046625}+\dotsm \end{aligned} \end{equation} for $x\in(-\pi,\pi)$, we can arrive at \begin{align*} \frac1{\cos(z^4)-1}&=-\frac12\biggl(\frac2{z^4}\biggr)^2\Biggl(\frac{\frac{z^4}2}{\sin\frac{z^4}2}\Biggr)^2\\ &=-\frac2{z^8} \Biggl\{1+\frac{B_2}2z^8-4\sum_{k=2}^{\infty}(-1)^k\Biggl[\bigl(2^{2k-1}-1\bigr)B_{2k}\\ &\quad-\sum_{j=1}^{k-1}\binom{2k}{2j}\bigl(2^{2j-1}-1\bigr)\bigl(2^{2k-2j-1}-1\bigr)B_{2j}B_{2k-2j}\Biggr]\frac{z^{8k}}{2^{2k}(2k)!}\Biggr\} \end{align*} for $z^4\in\bigl(-2\pi,2\pi\bigr)$.
References
- Xue-Yan Chen, Lan Wu, Dongkyu Lim, and Feng Qi, Two identities and closed-form formulas for the Bernoulli numbers in terms of central factorial numbers of the second kind, Demonstratio Mathematica 55 (2022), no. 1, 822--830; available online at https://doi.org/10.1515/dema-2022-0166.

- 1,691
-
The power series of the function $\frac1{2\cos x-1}$ is at https://math.stackexchange.com/a/4672327 – qifeng618 Apr 04 '23 at 21:45
-