3
$\begingroup$

I want to find the Laurent expansion for $\frac{\exp\left(\frac{1}{z^{2}}\right)}{z-1}$ about $z=0$,

I've tried to apply this formula $\frac{1}{1-\omega}=\sum_{n=0}^{\infty }\omega^{n}$ and the usual Taylor series of the exponential function, but I don't know how to continue:
$$\begin{align}f(z)&=\frac{1}{z-1}\exp\left(\frac{1}{z^{2}}\right)\\ &=-\frac{1}{1-z}\exp\left(\frac{1}{z^{2}}\right)\\&=-\left (\sum_{n=0}^{\infty }z^{n} \right )\left ( \sum_{n=0}^{\infty}\frac{1}{n!z^{2n}} \right )\end{align}$$ Thanks in advance.
Ps: I tried applying a Cauchy product, but I think this is not appropriate.
Edit 1: If it is useful at the end of the text, the authors say that the Laurent expansion is:
$\sum_{k=-\infty }^{\infty }a_{k}z^{k}$ with $a_{k}=-e$ if $k\geq 0$ and $a_{k}=-e+1+\frac{1}{1!}+\frac{1}{2!}+...+\frac{1}{(j-1)!}$if $k=-2$ or $k=-2j+1$ where $j=1,2,...$

$\endgroup$
0

2 Answers 2

1
$\begingroup$

Starting with your $=-\left (\sum\limits_{m=0}^{\infty }z^{m} \right )\left ( \sum\limits_{n=0}^{\infty}\frac{1}{n!z^{2n}} \right )$ changing one of the $n$ to $m$, you can say the coefficient of $z^k$ is

  • $-\sum\limits_{n=0}^{\infty} \frac1{n!} =-e$ when $k\le 0$
  • $-\sum\limits_{n=k/2}^{\infty} \frac1{n!} =\sum\limits_{n=0}^{n=(k-2)/2} \frac1{n!}-e$ when $k\gt 0$ and even
  • $-\sum\limits_{n=(k+1)/2}^{\infty} \frac1{n!} =\sum\limits_{n=0}^{(k-1)/2} \frac1{n!}-e$ when $k\gt 0$ and even

But that looks wrong to me: I do not think $$\cdots -e z^{-5} -e z^{-4} -e z^{-3} -e z^{-2} -e z^{-1} -e z^{0}+ \\(1-e)z^1 +(1-e)z^2 +(2-e)z^3 +(2-e)z^4+\left(\frac52-e\right)z^5+\cdots$$ converges when $|z| \le 1$.

Meanwhile for the same question asked elsewhere, a suggested answer was in effect $$z^{-1}+z^{-2}+2 z^{-3}+2 z^{-4}+\frac{5 }{2}z^{-5}+\frac{5}{2}z^{-6}+\cdots$$ but I do not think that converges either when $|z|\le 1$

$\endgroup$
10
  • $\begingroup$ The series you wrote for $\frac1{1-z}$ fails to converge for $|z|>1$. $\endgroup$
    – Mark Viola
    Nov 18, 2020 at 21:43
  • $\begingroup$ @MarkViola - You may be correct if you say it fails to converge for $|z|\le 1$ $\endgroup$
    – Henry
    Nov 18, 2020 at 21:49
  • 1
    $\begingroup$ Henry, $\frac1{1-z}=\sum_{n=0}^\infty z^n$ for $|z|<1$, and fails to converge for $|z|\ge 1$. $\endgroup$
    – Mark Viola
    Nov 18, 2020 at 21:51
  • $\begingroup$ @MarkViola I am more worried that $\cdots -e z^{-5} -e z^{-4} -e z^{-3} -e z^{-2} -e z^{-1} -e z^{0}$ does not converge than I am about $(1-e)z^1 +(1-e)z^2 +(2-e)z^3 +(2-e)z^4+(\frac52-e)z^5+\cdots$ $\endgroup$
    – Henry
    Nov 18, 2020 at 21:54
  • $\begingroup$ I've posted a complete solution that includes the Laurent series for $|z|>1$ and the Laurent series for $0<|z|<1$. $\endgroup$
    – Mark Viola
    Nov 18, 2020 at 23:00
0
$\begingroup$

First, we can write two series for $\frac1{z-1}$ in the two regions $|z|<1$ and $|z|>1$ as

$$\frac1{z-1}=\begin{cases} -\sum_{n=0}^\infty z^n&,|z|<1\\\\ \sum_{n=1}^\infty z^{-n}&,|z|>1\tag1 \end{cases}$$


Second, the Laurent series for $e^{1/z^2}$ for $0<|z|$ is given by

$$e^{1/z^2}=\sum_{n=0}^\infty \frac{a_n}{(n/2)!}\,z^{-n}\tag2$$

where $a_n$ the sequence such hat

$$a_n=\begin{cases} 1&,n\,\text{even}\\\\ 0&,n\,\text{odd} \end{cases}$$


Putting $(1)$ and $(2)$ together reveals

$$\frac{e^{1/z^2}}{z-1}= \begin{cases} -\sum_{m=0}^\infty z^m \sum_{n=0}^\infty \frac{a_n}{(n/2)!}\,z^{-n}&,0<|z|<1\tag3\\\\ \sum_{m=1}^\infty z^{-m}\sum_{n=0}^\infty \frac{a_n}{(n/2)!}\,z^{-n}&,1<|z| \end{cases} $$



For $|z|>1$, the Laurent series of $\frac{e^{1/z^2}}{z-1}$ can be written

$$\begin{align} \frac{e^{1/z^2}}{z-1}&=\sum_{m=1}^\infty z^{-m}\sum_{n=0}^\infty \frac{a_n}{(n/2)!}\,z^{-n}\\\\ &=\sum_{n=0}^\infty \frac{a_n}{(n/2)!}\,\sum_{m=1}^\infty z^{-(n+m)}\\\\ &\overbrace{=}^{p=n+m}\sum_{n=0}^\infty \frac{a_n}{(n/2)!}\sum_{p=n+1}^\infty\,z^{-p}\\\\ &=\sum_{p=1}^\infty\left(\sum_{n=0}^{p-1} \frac{a_n}{(n/2)!}\right)\,z^{-p} \end{align}$$



For $0<|z|<1$, the Laurent series of $\frac{e^{1/z^2}}{z-1}$ can be written

$$\begin{align} \frac{e^{1/z^2}}{z-1}&=-\sum_{m=0}^\infty z^{m}\sum_{n=0}^\infty \frac{a_n}{(n/2)!}\,z^{-n}\\\\ &=-\sum_{n=0}^\infty \frac{a_n}{(n/2)!}\sum_{m=0}^\infty z^{m-n}\\\\ &\overbrace{=}^{p=m-n}-\sum_{n=0}^\infty \frac{a_n}{(n/2)!}\sum_{p=-n}^\infty z^{p}\\\\ &=-\sum_{n=0}^\infty \frac{a_n}{(n/2)!}\left(\sum_{p=-n}^{0} z^{p}+\sum_{p=1}^\infty z^{p}\right)\\\\ &=-e \sum_{p=1}^\infty z^{p}-\sum_{n=0}^\infty \frac{a_n}{(n/2)!}\sum_{p=0}^{n} z^{-p}\\\\ &=-e \sum_{p=1}^\infty z^{p}-\sum_{p=0}^{\infty}\left(\sum_{n=p}^\infty \frac{a_n}{(n/2)!} \right)z^{-p}\\\\ &=-e \sum_{p=0}^\infty z^{p}-\sum_{p=1}^{\infty}\left(\sum_{n=p}^\infty \frac{a_n}{(n/2)!} \right)z^{-p} \end{align}$$


$\endgroup$
7
  • $\begingroup$ thanks, but I don't understand why took that series for $exp(\frac{1}{z^{2}})$. $\endgroup$ Nov 18, 2020 at 23:45
  • $\begingroup$ @BrigitteEliana What is it that you don't understand? $\endgroup$
    – Mark Viola
    Nov 19, 2020 at 2:51
  • $\begingroup$ Yes, I don't understand how did you determine that the Laurent series for $exp(\frac{1}{z^{2}})$ is the given in equation (2)? $\endgroup$ Nov 19, 2020 at 3:32
  • $\begingroup$ The Laurent expansion for $|z|>0$ of $e^{1/z^2}$ is given by $$e^{1/z^2}=\sum_{n=0}^\infty \frac1{n!}\,\frac1{z^{2n}}$$Now, note that all of the terms are of even inverse powers of $z$. That is, all of the coefficients of odd inverse powers are equal to $0$. So, we define $a_n$ to be $1$ when $n$ is even and $0$ when $n$ is odd. Then, $\sum_{n=0}^\infty \frac1{n!}\,\frac1{z^{2n}}=\sum_{n=0}^\infty \frac{a_n}{(n/2)!}\,\frac1{z^n}$. Is that clear now? $\endgroup$
    – Mark Viola
    Nov 19, 2020 at 3:40
  • $\begingroup$ Yes, thanks for your help. $\endgroup$ Nov 19, 2020 at 4:18

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.