3
$\begingroup$

Suppose we have a sequence of entire functions $f_n$ such that $$\text{$f_n(z)\to0$ for each natural $z$}\tag{1}$$ (as $n\to\infty$).

Is it possible to give general additional conditions on the sequence $(f_n)$ ensuring that (1) implies
$$\text{$f_n(z)\to0$ for each complex $z$?}\tag{2}$$

As a minimum, I would like such general additional conditions to hold and be easily verifiable when $$f_n(z)=-1+\frac1{n^z}\sum_{k=0}^\infty k^z\,\frac{n^k}{k!}\,e^{-n};$$ cf. this answer.

$\endgroup$
2
  • $\begingroup$ Did you try interpolational results like Ramanujan Master theorem? $\endgroup$ Commented Sep 15, 2021 at 18:49
  • $\begingroup$ @FedorPetrov : Thank you for your comment. I did not know of the Ramanujan Master theorem -- will try to learn about it. $\endgroup$ Commented Sep 15, 2021 at 19:03

1 Answer 1

3
$\begingroup$

Edited.

  1. For your general question, one sufficient condition is that your functions are of exponential type $<\pi$. This is best possible since anything convergent to $\sin\pi z$ would be a counterexample, and $\sin\pi z$ has exponential type exactly $\pi$.

One can slightly improve this. For example when all $L^2(R)$ norms of your functions are bounded, independently of $n$. Then one can apply the Sampling theorem (which is usually credited to Nyquist in the West and Kotelnikov in the former Soviet Union, but in fact goes back to Cauchy), which estimates the $L^2(R)$ norm in terms of $\ell^2(Z)$ norm, provided that the exponential type is $\leq\pi$.

  1. However your specific function does not have finite exponential type, so these general results will not work, and you need to use some specific properties of your function. You say that yout functions satisfy $f_n(z)\to 0$ for every integer $z$. Why is this so, and why your proof does not work for non-integer $z$?
$\endgroup$
18
  • $\begingroup$ Good point about the maximum modulus argument. Could you please provide details or references on the exponential type $<\pi$ being sufficient? $\endgroup$ Commented Sep 15, 2021 at 19:32
  • 1
    $\begingroup$ For the first paragraph, the keyword is "Jensen's formula", and a reference is Boris Ya. Levin (any of his two books). For the second paragraph, for example, G. Holland, Fourier Analysis and its applications, sect. 7.3, Sampling theorem. $\endgroup$ Commented Sep 15, 2021 at 19:40
  • $\begingroup$ Thank you for the references. Concerning the maximum modulus argument, now I am having a second thought: in view of of the term $-1$, how to show that the maximum modulus in $|z|\le r$ is achieved at $z=r$? $\endgroup$ Commented Sep 15, 2021 at 19:51
  • $\begingroup$ Also, I have found "2.3 The Jensen formula" in Levin's book, but it is about counts of zeroes, and I don't see how to use it here. $\endgroup$ Commented Sep 15, 2021 at 20:21
  • $\begingroup$ Also, "The Sampling Theorem" in G. Folland, Fourier Analysis and its applications, sect. 7.3, provides an expression for $f(t)$ (for real $t$?) as an infinite linear combination of the values of $f$ on a lattice in $\mathbb R$. But I am not sure how to use this to prove the convergence, even on $\mathbb R$. $\endgroup$ Commented Sep 15, 2021 at 20:26

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .