Compute the residue of a function

  • Thread starter Thread starter docnet
  • Start date Start date
  • Tags Tags
    Function Residue
docnet
Messages
796
Reaction score
488
Homework Statement
Find ##Res(h,z_0)##
Relevant Equations
##h=\frac{f}{g}##
final. problem 6.png

There is a typo. It should say ##h=\frac{f}{g}##.

Attempt: ##f## and ##g## are holomorphic on ##\Omega##. Homomorphic functions form a ##\mathcal{C}^*## algebra, so ##h## is holomorphic on ##\Omega## where ##g\neq 0##.

If ##z_0## is a removal singularity of ##h##, then ##Res(h,z_0)=0## by Morera's theorem.

Assume ##f(z_0)\neq 0##, then ##z_0## is a second order pole of h since ##g''(x_0)\neq 0 ##.
$$Res(h,z_0)=lim_{z\rightarrow z_0}\frac{d}{dz}\Big[\frac{f}{g}(z-z_0)^2\Big]$$
This method produces indeterminate forms, even after applying the L'Hopital's rule.

Let ##\mathcal{C}=\{z:|z-z_0|=1/2\}##. By Residue theorem,
$$Res(h,z_0)=\frac{1}{2\pi i}\int_\mathcal{C}\frac{f}{g}dz$$
This integral cannot be computed since ##f## and ##g## are not given.

Can anyone think of a clever method that could solve this problem?
 
Last edited:
Physics news on Phys.org
You can just start writing out series. The Taylor series for f is known. For ##1/g##,, you know if looks like ##1/((z-z_0)^2(a+b(z-z_0)+c(z-z_0)^2+...) ## with a nonzero.

You can basically just bash this with the Taylor series of ##1/(1-x)##. Let ##x=\frac{1}{a}(-b(z-z_0)-c(z-z_0)^2+...)## and write
$$\frac{1}{g}= \frac{1}{(z-z_0)^2 a (1-x)}$$

Now apply the normal Taylor expansion to ##1/(1-x)##, and after the first couple terms you should get stuff that is high order in ##(z-z_0)## and hence can be ignored.
 
Office_Shredder said:
You can just start writing out series. The Taylor series for f is known. For ##1/g##,, you know if looks like ##1/((z-z_0)^2(a+b(z-z_0)+c(z-z_0)^2+...) ## with a nonzero.

You can basically just bash this with the Taylor series of ##1/(1-x)##. Let ##x=\frac{1}{a}(-b(z-z_0)-c(z-z_0)^2+...)## and write
$$\frac{1}{g}= \frac{1}{(z-z_0)^2 a (1-x)}$$

Now apply the normal Taylor expansion to ##1/(1-x)##, and after the first couple terms you should get stuff that is high order in ##(z-z_0)## and hence can be ignored.
Thank you so much for your reply. I will try this method.
 
There are two things I don't understand about this problem. First, when finding the nth root of a number, there should in theory be n solutions. However, the formula produces n+1 roots. Here is how. The first root is simply ##\left(r\right)^{\left(\frac{1}{n}\right)}##. Then you multiply this first root by n additional expressions given by the formula, as you go through k=0,1,...n-1. So you end up with n+1 roots, which cannot be correct. Let me illustrate what I mean. For this...

Similar threads

Back
Top