2
$\begingroup$

Let $f:\mathbb{C}^n\to\mathbb{C}$ be an entire holomorphic function of $n$ complex variables. Then its Nevannlinna characteristic equals $$ m_f(r)=\int_{\partial B(r)}\log^+|f(z)|d\eta(z),\quad\forall r>0, $$ where $\partial B(r)$ is the sphere of radius $r$ and $\eta$ is the normalized Euclidean measure on it. In the $n=1$ case one can deduce from Poisson-Jensen formula that the following estimate holds (e.g., Levin p. 13), $$ \log|f(z)|\le\frac{r+|z|}{r-|z|}m_f(r),\quad\forall |z|<r. $$ I am looking for an analogous estimate of the modulus in terms of the Nevanlinna characteristic for $n>1$. But all various Poisson/Jensen/Nevanlinna formulae I have seen so far work with $n$-dimensional real hypersurfaces $\partial D$ (Bergman-Shilov boundaries etc.) rather than with the natural $2n-1$-dimension sphere $\partial B(r)$, and are thus of no use for me.

Question: Does there exist a similar estimate for $n>1$, and where can I find it?

To be more precise, I am looking for an estimate of the form $$ \log|f(z)|\le F\left(\frac{\|z\|}r\right)m_f(r),\quad\forall z\in\mathbb{C}^n,\quad\forall r>0\quad\mbox{s.t.}\quad c_1<\frac{\|z\|}r<c_2, $$ $$ F:(c_1,c_2)\to\mathbb{R}_+,\quad c_1,c_2\in\mathbb{R}_+. $$ Here $F$, $c_1$ and $c_2$ may depend on $f$, if it helps. Thank you.

$\endgroup$

1 Answer 1

2
$\begingroup$

I belive the following should work: basically the only thing we know is 2-dimensional estimate that you wrote (I'll always talk about real dimension to avoid ambiguity). Then to get estimate of $\log |f(z_0)|$ in terms of $\int_{\partial B(r)} \log^{+}|f(z)|d\eta(z)$ let's take average over all 2-dimensional planes passing through $z_0$. First of all all sections by these planes are 2-dimensional disks and $z_0$ is not to close to its boundary, so that we have something like $\log |f(z_0)| \le G\left(\frac{||z_0||}{r}\right) m_{f}(T) $, where $m_f(T)$ is Nevannlinna characteristic in the section by 2-dimensional plane $T$ (actually, distance from $z_0$ to the sphere and therefore to the circle is at least $r - ||z_0||$ while radius of all circles is at most $r$). Finally it is not very hard to convince oneself that when you average all $T$'s every point on the $\partial B(r)$ will contribute to the result with coefficient bounded by some other $H\left(\frac{||z_0||}{r}\right)$ and we therefore get the desired bound with $F(t) = G(t)H(t)$.

$\endgroup$
2
  • $\begingroup$ Thank you, Aleksei, for the answer. This sounds like an excellent idea from a first glance. I will try to fill in all details before making further comments. Thanks once again. $\endgroup$
    – Bedovlat
    Jan 24, 2019 at 4:33
  • $\begingroup$ Details filled in. The idea works. Answer accepted. Thank you, Aleksei. Your contribution will be duly acknowledged. $\endgroup$
    – Bedovlat
    Jan 25, 2019 at 3:08

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.