6
$\begingroup$

I am trying to find the taylor series expansion about $0$ (maclaurin series) of $$x \rightarrow \frac{1}{\sqrt{1-\beta x(x+1)}} \text{ with } \beta \in \mathbb{R}^{+*}$$

I've tried using the taylor series expansion of $$\frac{1}{\sqrt{1-X}} = \sum_{n=0}^{\infty}4^{-n}{2n \choose n}X^n \text{ }\text{ }\text{ }\text{ with } \text{ } X=\beta x(x+1)$$

But I can't turn it into a power series because of the $(x+1)^n$...

I've also tried to derive $$\frac{1}{n!}\cdot\frac{\text{d}^n}{\text{d}x^n}\left(\frac{1}{\sqrt{1-\beta x(x+1)}}\right)_{x=0}$$ But no results so far...

Edit : with a more powerful method, I found that, if we call $\left(a_n\right)_{n\in\mathbb{N}}$ the coefficients of the taylor series expansion $\left(\frac{1}{\sqrt{1-\beta x(x+1)}}=\sum_{n=0}^{\infty}a_nx^n\right)$, the sequence $\left(a_n\right)_{n\in\mathbb{N}}$ is then defined by : $$\forall n\geq3, \text{} na_n=\beta\left(n-\frac{1}{2}\right)a_{n-1}+\beta\left(n-1\right)a_{n-2}$$ $$\text{with }\text{ }a_1 = \frac{\beta}{2} \text{ , } a_2 = \frac{3}{8}\beta^2+\frac{1}{2}\beta$$

It's definitely a step forward, but I don't know how to proceed from there. Is there a way to handle sequences that are defined by such a way ?

$\endgroup$
1
$\begingroup$

I would set $y = \beta x(x+1)$. With that: \begin{align} \frac1{\sqrt{1 - y}} &= \sum_{n=0}^{\infty} 4^{-n} \binom{2n}{n} y^n\\ &= \sum_{n=0}^{\infty} 4^{-n} \binom{2n}{n} \beta^n x^n (x+1)^n\\ &= \sum_{n=0}^{\infty}\sum_{k=0}^{n} 4^{-n} \binom{2n}{n} \binom{n}{k} \beta^n x^n x^k \\ &= \sum_{m=0}^{\infty} a_m x^m \end{align} where $m = n + k$, so $a_m = \sum_{n=0}^{m} 4^{-n}\binom{2n}{n}\binom{n}{m-n} \beta^n$.

$a_m$ is simply an number which you can calculate doing some sums and products. For $m = 0, \dots, 3$ the $a_m$ are $1,\frac{\beta }{2},\frac{1}{8} \beta (3 \beta +4),\frac{1}{16} \beta ^2 (5 \beta +12)$.

Mathematica "simplifies" $a_m$ to $$ a_m = 4^{-m} \binom{2 m}{m} \beta ^m \, _2F_1\left(\frac{1}{2}-\frac{m}{2},-\frac{m}{2};\frac{1}{2}-m;-\frac{4}{\beta }\right) $$ where $_2F_1$ denotes the Hypergeometric function.

$\endgroup$
  • $\begingroup$ Well actually, after numerical simulation, your numerical values don't seem to work... $\endgroup$ – Harmonic Sun Dec 14 '18 at 0:17
  • $\begingroup$ @HarmonicSun Yes, I had forgotten to include $\beta$. The values are not as pretty now but you can still easily calculate them. $\endgroup$ – 0x539 Dec 14 '18 at 0:20
  • $\begingroup$ Ok thanks a lot ! Additionaly, do you think that based on my edit on my initial question, one can derive a prettier closed form for the $a_m$ ? $\endgroup$ – Harmonic Sun Dec 14 '18 at 0:26
  • $\begingroup$ I might be picky but, in your formula for $a_m$, ${n \choose m-n}$ is not defined for, say, n=0... Or do we take it a 0 conventionaly ? $\endgroup$ – Harmonic Sun Dec 14 '18 at 0:33
  • $\begingroup$ @HarmonicSun Yes, the Binomial coefficient $\binom{a}{b}$ is understood to be zero when $a < b$ (and $a, b \in \mathbb{Z}$). $\endgroup$ – 0x539 Dec 14 '18 at 0:39

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.