0
$\begingroup$

If a function $f$ is analytic in a disk $\Delta(z_0,r)$ and if $m$ is a constant such that $|f(z)-f(z_0)|\leq m$ for all $z \in \Delta(z_0,r)$, then $|f'(z_0)|\leq m/r$ and $|f(z)-f(z_0)|\leq(m/r)|z-z_0|$ for all $z \in \Delta(z_0,r)$.

This is my revised attempt at a proof:

Let $g(z)=\frac{f(z)-f(z_0)}{z-z_0}$ for $z \in \Delta(z_0,r)\setminus\{z_0\}.$ Let $0< \rho < r$. Then we have the following:

$$\left| g(z) \right|=\max\limits_{|z|=\rho}\left|g(z)\right|=\max\limits_{|z|=\rho}\left|\frac{f(z)-f(z_0)}{z-z_0}\right|\leq \frac{m}{r}.$$

I believe the above more clearly establishes the inequality.

To establish the estimate since $\Delta(z_0,r)$ is a disk that means any two points $z_1, z_2$ have a line entirely in $\Delta(z_0,r).$ Let $\gamma(t)=z_1-t(z_2-z_1)$. Then $|f(z_2)-f(z_1)|=\left|\int_{0}^{1}f(\gamma(t))\gamma'(t)\,dt\right|\leq \int_{0}^{1}\left|f(\gamma(t)) \right|\left| \gamma'(t)\right|\,dt=\frac{m}{r}\int_{0}^{1}\left|f(\gamma(t))\right|\,dt$

What I was attempting to do was follow the reasoning from this question I had asked earlier. I am still unable to establish the inequality.

$\endgroup$
1
$\begingroup$
  1. You haven't defined $g$. I assume that $$g(\zeta) = \frac{f(z_0+\zeta)-f(z_0)}{\zeta}$$ extended by $g(0)=f'(z_0)$ to become a holomorphic function in $\Delta(0,r)$.
  2. I do not understand your manipulations with $g$, especially with the mysterious "$z\in \Delta$" at the end. What is $z$ here? What is $\Delta$? Neither is defined.
  3. If (as is likely) $\Delta(0,r)$ means an open disk, then you can't talk about the values of $g(\zeta)$ when $|\zeta|=r$. Instead, having fixed $\zeta$, consider some circle of radius $|\zeta|<\rho<r$, and apply the maximum principle there: $$ |g(\zeta)|\leq \max_{|\zeta|=\rho} |g(\zeta)| \le \frac{m}{\rho}$$ Then let $\rho\to r$ to obtain the desired inequality. With $\zeta=0$, it gives $|f'(z_0)|\le m/r$. For general $\zeta\in \Delta(0,r)$ it gives $|f(z)-f(z_0)|\le (m/r)|z-z_0|$
  4. I do not see the logic of "follows from the fact that the disk is a convex set." It's not like you can integrate the estimate for $f'(z_0)$ along a segment: the point $z_0$ is fixed all along, it's the center of the disk.
$\endgroup$
  • $\begingroup$ I cleaned up my question a lot and hopefully cleared up some stuff. I am still having problems establishing the estimate at the end of the problem. $\endgroup$ – emka Feb 17 '14 at 4:04
  • $\begingroup$ @Mark You should give up the idea of obtaining the estimate on $|f(z)-f(z_0)|$ by integrating the derivative. This is just not how Schwarz lemma works: you don't have a suitable derivative estimate to integrate. Instead use the maximum principle for $g$. the bound $|g|\le m/r$ is exactly the estimate on $|f(z)-f(z_0)|$ that you are trying to prove. Also, $|g(z_0)|\le m/r$ is the estimate on $f'(z_0)$ that you need. $\endgroup$ – user127096 Feb 17 '14 at 4:10
  • $\begingroup$ Are you saying I just get it by multiplying both sides by $|z-z_0|$. $\endgroup$ – emka Feb 17 '14 at 4:31
  • $\begingroup$ @Mark Yes, that's how you should get it. $\endgroup$ – user127096 Feb 17 '14 at 4:36

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.