5

If $q\colon\mathbb{C}\to\mathbb{C}$ is a polynomial, $f\colon\mathbb{C}\to\mathbb{C}$ is analytic on all of $\mathbb{C}$, and if there exists $a\gt 0$ such that $|f(z)| \lt a|q(z)|$ for every $z\in \mathbb{C}$, then $f = bq$ for some $b\in \mathbb{C}$.

Can an arbitrary analytic function (on all of $\mathbb{C}$) replace $q$?

Jonas Meyer
  • 53,602
lively
  • 150

1 Answers1

5

If you really mean strict inequality, then this follows from Liouville's theorem applied to $f/q$. Note that $q$ must be constant too, as $q$ can have no zeros from the condition.

It is actually true if $<$ is replaced by $\leq$, though. (Maybe you meant this?) First, the Cauchy estimates show that $f$ is a polynomial; indeed, since $f$ grows polynomially, we can just take averages of $f/(z-\alpha)^N$ over larger and larger circles. From this one can see that $f$ is a polynomial.

Now the question reduces to showing that if a polynomial $p$ is bounded by a constant multiple of another polynomial $q$, then $p$ and $q$ differ by a constant. This is a straightforward consequence of factorization.

Akhil Mathew
  • 31,310
  • 1
    Good answer. You could also apply Liouville's theorem without strict inequality, in conjunction with the fact that an analytic function has isolated zeros and a singularity of a bounded function is removable. – Jonas Meyer Jan 30 '11 at 06:40
  • Thank you so much Akhil and Jonas! I meant strict inequality, sorry about that. – lively Jan 30 '11 at 06:40