4
$\begingroup$

I feel like this should be well-known, but haven't been able to find any reference so far. Consider the set of all smooth functions on $\mathbb{R}$ such that $$\sup_{x\in \mathbb{R}} |e^{\alpha x} f^{(n)}(x)|<\infty \qquad \mbox{for all $n\in \mathbb{N},$ $\alpha \in \mathbb{C}$. }$$ Then the two-sided laplace transform of such a functions is well defined, and gives rise to a holomorphic function on $\mathbb{C}$. Is there a Paley-Wiener kind of theorem that gives a characterization of the space of functions that can be generated in this way? (That is, as Laplace transforms of smooth functions with exponential decay for all its derivatives.)

$\endgroup$
0

1 Answer 1

4
$\begingroup$

We can certainly characterize these functions in terms of their Fourier transforms, but what I'm about to write down just rephrases the well known connection between holomorphic (on a strip) Fourier transforms and exponential decay, and it's not very exciting.

Let me write $F$ for the FT of $f$. Then I claim that $f$ satisfies your conditions if and only if $F$ is an entire function such that for any $n\ge 0$, $L\ge 1$, we have that $(x+iy)^nF(x+iy)$ is bounded on $-L\le y\le L$ and $$ \sup_{-L\le y\le L} \int_{-\infty}^{\infty} |x+iy|^n |F(x+iy)|\, dx < \infty . $$

It is obvious that $F(z)=\int f(t) e^{itz}\, dt$ has these properties if $f$ is as above (note that the smoothness of $f$ gives $F$ enough decay to make it integrable). Conversely, if $F$ satisfies these conditions, then we obtain the desired estimates on $f^{(n)}(x)=\int (-iz)^n F(z) e^{-itz}\, dz$ by moving the path of integration into the complex plane, to pick up the required exponentially small factors.

$\endgroup$
2
  • $\begingroup$ I was not aware of this possible characterization, and it actually fits very nicely with the context I'm working on. Thank you very much! $\endgroup$
    – Raul Gomez
    Commented Jun 14, 2017 at 17:11
  • $\begingroup$ @RaulGomez: Actually, I think it becomes a characterization only because you impose exponential decay at arbitrarily fast rates, I think it's not as clean when you consider the connection FT holomorphic on a strip <-> function has exponential decay. $\endgroup$ Commented Jun 14, 2017 at 23:23

You must log in to answer this question.

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