0

I want to calculate Laurent expansion of $\frac{1}{z-1}$ thtat are valid in the annuli

$\begin{align} (a) & \;\;1<|z|<3\\ (b) & \;\;0<|z-3|<2 \end{align}$

For part $(a)$ since $|z|>1$ we use the trick $$\frac{1}{z-1} = \frac{1}{z(1-\frac{1}{z})}$$ which can be turned into a geometric series.

For part $(b)$ why did the book used $$\frac{1}{z-1} = \frac{1}{2(1+\frac{z-3}{2})}.$$ I know $|\frac{z-3}{2}| < 1$ but why can't the the method in part $(a)$ be applied here, since we also have $|z| > 1$.

The original problem was $\frac{1}{(z-1)(z-3)}$ I think this might effect the two different calculations because $z$ can not equal to $3$ in the expansion? And suppose we do not have the factor $(z-3)$ on the bottom, could the method in $(a)$ applied to $(b)$?

Xiao
  • 9,764

2 Answers2

2

Note: The regions of convergence are somewhat different than you might think. Here I provide a complete Laurent expansion of $f$ around $z=0$. From this you should be able to check your results.

The function

\begin{align*} f(z)&=\frac{1}{(z-1)(z-3)}\\ &=-\frac{1}{2}\frac{1}{z-1}+\frac{1}{2}\frac{1}{z-3}\\ \end{align*} has two simple poles at $1$ and at $3$.

Since we want to find a Laurent expansion with center $0$, we look at the poles $1$ and $3$ and see they determine three regions.

\begin{align*} |z|<1,\qquad\quad 1<|z|<3,\qquad\quad 3<|z| \end{align*}

  • The first region $ |z|<1$ is a disc with center $0$, radius $1$ and the pole $1$ at the boundary of the disc. In the interior of this disc all two fractions with poles $1$ and $3$ admit a representation as power series at $z=0$.

  • The second region $1<|z|<3$ is the annulus with center $0$, inner radius $1$ and outer radius $3$. Here we have a representation of the fraction with pole $1$ as principal part of a Laurent series at $z=0$, while the fraction with pole at $3 admits a representation as power series.

  • The third region $|z|>3$ containing all points outside the disc with center $0$ and radius $3$ admits for all fractions a representation as principal part of a Laurent series at $z=0$.

A power series expansion of $\frac{1}{z+a}$ at $z=0$ is \begin{align*} \frac{1}{z+a}&=\frac{1}{a}\cdot\frac{1}{1+\frac{z}{a}} =\frac{1}{a}\sum_{n=0}^\infty\left(-\frac{z}{a}\right)^n\\ &=\sum_{n=0}^{\infty}\frac{(-1)^n}{a^{n+1}}z^n \end{align*} The principal part of $\frac{1}{z+a}$ at $z=0$ is \begin{align*} \frac{1}{z+a}&=\frac{1}{z}\cdot\frac{1}{1+\frac{a}{z}}=\frac{1}{z}\sum_{n=0}^{\infty}\left(-\frac{a}{z}\right)^n\\ &=\sum_{n=1}^{\infty}(-a)^{n-1}\frac{1}{z^n} \end{align*}

We can now obtain the Laurent expansion of $f(x)$ at $z=0$ for all three regions

  • Region 1: $\ |z|<1$

\begin{align*} f(z)&=\frac{1}{2}\sum_{n=0}^{\infty}z^n -\frac{1}{2}\sum_{n=0}^{\infty}\frac{1}{3^n}z^n\\ &=\frac{1}{2}\sum_{n=0}^{\infty}\left(1-\frac{1}{3^n}\right)z^n \end{align*}

  • Region 2: $\ 1<|z|<3$

\begin{align*} f(z)&=-\frac{1}{2}\sum_{n=1}^{\infty}\frac{1}{z^n} -\frac{1}{2}\sum_{n=0}^{\infty}\frac{1}{3^n}z^n \end{align*}

  • Region 3: $\ 3<|z|$

\begin{align*} f(z)&=-\frac{1}{2}\sum_{n=1}^{\infty}\frac{1}{z^n} +\frac{1}{2}\sum_{n=1}^{\infty}3^{n-1}\frac{1}{z^n}\\ &=\frac{1}{2}\sum_{n=1}^{\infty}\left(3^{n-1}-1\right)\frac{1}{z^n}\\ \end{align*}


Note: According to OPs comment let's assume a slightly different situation:

Challenge: Find the Laurent series expansion of $$f(z)=\frac{1}{(z-1)(z-3)}$$ at $z_0=3$ in the region $$0<|z-3|<2$$

We can immediately conclude from the representation of $f$ that $1$ and $3$ are simple poles.

Note, that in case of isolated singularities the radius of convergence is the distance from the center $z_0$ to the nearest singularity ($\neq z_0$).

Since $z_1=1$ is the nearest singularity to $z_0=3$ we have a radius of convergence $R=2$ and since $z_0$ is already a singularity we have a punctured disc as region of convergence \begin{align*} 0<|z-3|<2 \end{align*}

When looking at \begin{align*} f(z)=-\frac{1}{2}\frac{1}{z-1}+\frac{1}{2}\frac{1}{z-3} \end{align*} we observe the term $\frac{1}{2}\frac{1}{z-3}$ is already the principal part of the Laurent expansion at $z_0=3$ and since $\frac{1}{2}\frac{1}{z-1}$ is analytic at $z_0=3$ we can expand it as power series.

\begin{align*} f(z)&=-\frac{1}{2}\frac{1}{z-1}+\frac{1}{2}\frac{1}{z-3}\\ &=\frac{1}{2}\frac{1}{z-3}-\frac{1}{2}\frac{1}{(z-3)+2}\\ &=\frac{1}{2}\frac{1}{z-3}-\frac{1}{4}\frac{1}{1+\frac{z-3}{2}}\\ &=\frac{1}{2}\frac{1}{z-3}-\frac{1}{4}\sum_{n=0}^\infty \left(-\frac{1}{2}\right)^n(z-3)^n\\ &=\sum_{n=-1}^\infty \frac{(-1)^{n+1}}{2^{n+2}}(z-3)^n \end{align*}

Markus Scheuer
  • 112,413
  • Thank you very much. Another question: since we can do expansion in $1<|z|<3$ at $z_0=0$, why in the new region $0<|z-3|<2$, the book choose to do the expansion at $z_0=3$ (as the terms in the series is of the form $(z-3)^n$)? Is this the most nature way of doing it or it is required. – Xiao Jul 25 '16 at 15:19
  • @Xiao: The choice of the center simply depends on our needs resp. on the task we have to solve. So, it's crucial to check the precise formulation of the text for it. I've added a solution which might fit your needs. – Markus Scheuer Jul 25 '16 at 16:06
1

For rational functions, the idea is essentially always the same: transform the given annulus boundary conditions $r < \lvert z - a \rvert < R$ into the two equivalent conditions $\lvert R^{-1} (z - a) \rvert < 1$ and $\lvert r(z - a)^{-1} \rvert < 1$, and then typically use partial fraction decomposition to separate your function into terms involving $1/(1-R^{-1}(z - a))$ and $1/(1-r(z - a)^{-1})$ to express them using convergent geometric series. A typical example of application can be found here, for instance.

Alex Provost
  • 21,651
  • But there is no real requirement that the terms in the series has to be $\frac{z-a}{R}$ or $\frac{r}{z-a}$ right? It is a standard method. – Xiao Jul 24 '16 at 16:46
  • @Xiao I'm not sure what you're asking. – Alex Provost Jul 24 '16 at 19:10
  • Can I apply the formula from $(a)$ to $(b)$ in my problem even they are in different annulus. In another word, will the series expansion we get for $(a)$ also work as a solution for $(b)$. – Xiao Jul 24 '16 at 19:42