Unknownpgr

Finding the Number of Zeros Using Rouché's Theorem

2021-10-10 05:13:19 | English, Korean

This post was translated from Korean into English by AI.

Yesterday, while watching YouTube, I came across the fascinating video below. It explains how to use Rouché's Theorem to find the number of zeros of a complex function within a given region.

Definition

Rouche-thm

In simple terms, Rouché's Theorem as explained in this video says the following.

"If f(z)>g(z)|f(z)|>|g(z)| holds everywhere on the boundary D\partial D of a region DD, then f(z)f(z) and f(z)+g(z)f(z)+g(z) have the same number of zeros inside DD."

When f>g|f|>|g| holds everywhere, we say that ff is dominant over gg.

Using this theorem, if we want to find the number of zeros of a complicated polynomial h(z)h(z) within some region DD, we can make the problem easier by splitting h(z)h(z) into a simple, dominant term ff and the remaining terms gg.

Example

Let us follow the example from the video. We want to find the number of zeros of h(z)=z5+3z2+1h(z)=z^5+3z^2+1 in D={z1<z<2}D = \set{z|1<|z|<2}.

To do so, we divide the region into D1={zz<1},D2={zz<2}D_1 = \{z | |z| <1\}, D_2 = \{z ||z|<2\}, then subtract the number of zeros in D1D_1 and on D1\partial D_1 from the number of zeros in D2D_2.

First, to determine the number of zeros inside each region, note that on D1\partial D_1, z=1|z|=1, so z5=1|z^5|=1 and 3z2=3|3z^2|=3. Therefore, z5+1z5+1=2<3z2|z^5+1|\leq|z^5|+1=2<|3z^2|.

Thus, we can split hh into f(z)=3z2f(z)=3z^2 and g(z)=z5+1g(z)=z^5+1.

It follows that the number of zeros of h(z)h(z) inside D1D_1 is 2, the same as the number of zeros of f(z)f(z). (A double zero at z=0z=0)

Note that only the number of zeros is the same; the zeros themselves are not necessarily identical. Therefore, (in general) hh does not have a zero at z=0z=0.

Similarly, on D2={zz=2}\partial D_2 = \{ z| |z|=2\}, we have z5=32>3z2+1=4>3z2+1|z^5|=32>|3z^2|+1=4>|3z^2+1|, so we can split hh into f(z)=z5f(z)=z^5 and g(z)=3z2+1g(z)=3z^2+1.

Therefore, the number of zeros of h(z)h(z) inside D2D_2 is 5, the same as the number of zeros of f(z)f(z).

Finally, we need to show that hh has no zeros on D1={zz=1}\partial D_1 = \{z||z|=1\}.

The video also demonstrates this in a very simple way, as follows.

If a zero with z=1|z|=1 existed, then

z5+3z2+1=03z2=z513z2=z5+13=z5+1<z5+1=23<2z^5+3z^2+1=0\\ \therefore 3z^2=-z^5-1\\ \therefore |3z^2| = |z^5+1|\\ \therefore 3 = |z^5+1| < |z^5| + 1 = 2\\ \therefore 3<2

would have to hold. This is clearly false, so there is no z=1|z|=1 satisfying z5+3z2+1=0z^5+3z^2+1=0.

This is actually inevitable because f(z)f(z) is dominant over g(z)g(z) on D1\partial D_1. If a zero existed on the boundary, then h=f+g=0h=f+g = 0, so f=gf=-g and hence f=g|f|=|g|, which would mean that ff is not dominant.

Therefore, the number of zeros inside D2D_2 - the number of zeros on D1\partial D_1 - the number of zeros inside D1D_1 = 502=35-0-2= 3.


- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -