0
$\begingroup$

I am trying to understand the proof from this book of the following theorem:
$\textbf{Theorem}:$ If $f(z)= z+ \sum_{n\geq 2} a_n z^n$ is such that $\sum_{n\geq 2} n|a_n| \leq 1$, then $f$ is univalent in the unit disk $\mathbb{D}$.

In the proof, the authors first proved that the series converges and hence $f$ is an analytic function in $\mathbb{D}$. Till now I have no problem in understanding. The remaining part of the proof given in the book is as follows:

Let $|z_0|<1$, then \begin{equation} (f(z)-f(z_0))-(z-z_0) = \sum_{n\geq 2}a_n (z^n -z_{0}^{n}) = (z-z_0)\sum_{n\geq 2} a_n (z^{n-1} +z^{n-2}z_0 + \cdots + z_{0}^{n-1}). \end{equation} As $|z^{n-1} +z^{n-2}z_0 + \cdots + z_{0}^{n-1}|<n$ for $|z|<1$, we have $$|(f(z)-f(z_0))-(z-z_0)| < |z-z_0|\sum_{n\geq 2} n|a_n| \leq |z-z_0|.$$ According the Rouche's theorem, $f(z)-f(z_0)$ and $z-z_0$ have the same number of zeros in $\mathbb{D}$, that is $f(z)=f(z_0)$ has exactly one solution.

$\textbf{My doubt}:$ In Rouche's theorem, the inequality must hold on the boundary of the region, but in this proof $|z|<1$. So how can we apply Rouche's theorem here?

$\textbf{My explanation}: $ From the proof we can say that the inequality holds on any circle inside $\mathbb{D}$, as we can fix $|z|=\delta <1$, and $|z_0|<|z|$. We get the conclusion by allowing $\delta \rightarrow 1^-$.

$\textbf{My question}:$ Is my explpanation correct? Or I am missing out some basic fact?

$\endgroup$
  • $\begingroup$ $|z^{n-1} +z^{n-2}z_0 + \cdots + z_{0}^{n-1}|<n$ is also true when $|z|=1$ , so when $|z|=1$, you can get $|(f(z)-f(z_0))-(z-z_0)| < |z-z_0|$. $\endgroup$ – Riemann Jun 21 at 4:34
  • $\begingroup$ OK. But in that case don't we need $f$ to be analytic on a region containing the closed unit disk? $\endgroup$ – skylark Jun 21 at 4:53
  • 1
    $\begingroup$ From $\sum_{n\geq 2} n|a_n| \leq 1$, you can prove the convergent radius $R$ of $f(z)= z+ \sum_{n\geq 2} a_n z^n$ is not less than than $1$, that is to say $R\geq1$. $\endgroup$ – Riemann Jun 21 at 5:15
  • $\begingroup$ You can see the post math.stackexchange.com/questions/65298/… $\endgroup$ – Riemann Jun 21 at 5:16
  • 2
    $\begingroup$ Your explanation with $\delta$ (but with $\delta\to 1^-$) is fine $\endgroup$ – Hagen von Eitzen Jun 21 at 6:02
6
$\begingroup$

This is not an explanation of the steps in the book but I would like to point out that the proof in the book looks a bit silly to me. You don't require Rouche's Theorem. There is a very elementary 'High School' proof: suppose $f(z)=f(w)$. Then $z-w=a+a_2(w^{2}-z^{2})+a_3(w^{3}-z^{3})+\cdots$. Note that $|w^{n}-z^{n}| =|z-w|(|w^{n-1}+w^{n-2}z+\cdots+wz^{n-2}+z^{n-2}|< n|z-w|$ if $z \neq w$. Using the hypothesis we arrive at the contradiction $|z-w| <|z-w|$ if $z \neq w$.

$\endgroup$

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.