It is well-known that $\{e^{i n t}\}_{n\in\mathbb Z}$ is an orthonormal basis for $L^2(-\pi,\pi)$. A theorem by Kadec (Kadec $1/4$ theorem) studies the perturbed exponential system:

If $\{\lambda_n\}$ is a sequence of real numbers for which $$|\lambda_n-n|\leqq L<\frac{1}{4}, \ \ n=0, \pm 1, \pm 2, \dots$$ then $\{e^{i \lambda_n t}\}_{n\in\mathbb Z}$ satisfies the Paley-Wiener criterion and so forms a Riesz basis for $L^2(-\pi,\pi)$.

It implies that a set $\{e^{i\lambda_n t} \}_{n\in \mathbb Z}$ is a Riesz basis of $L^2(-\pi,\pi)$ whenever $l=\sup_{n\in\mathbb Z} |n - \lambda_n | < \frac{1}{4}$. But it does not exclude that there are sequences $\lambda_n$ which have distance from the integers greater than $1/4$ and such that $\{e^{i\lambda_n t} \}_{n\in \mathbb Z}$ is a Riesz basis for $L^2(-\pi,\pi)$.

A recent work

De Carli, L., 2019. Concerning exponential bases on multi-rectangles in $\mathbb R^d$, in: Abell, M., Iacob, E., Stokolos, A., Taylor, S., Tikhonov, S., Zhu, J.(Eds.), Topics in Classical and Modern Analysis: In Memory of YingkangHu. Springer International Publishing. Applied and Numerical HarmonicAnalysis. Chapter 3. arXiv version here

shows (see, in particular, Corollary 5.3 therein) that $l \geq \frac{1}{4}$ but under Corollary 5.3 $\{e^{i\lambda_n t} \}_{n\in \mathbb Z}$ is a Riesz basis of $L^2\left(-\pi,\pi\right)$.

I wonder if anyone knows other works that contain this kind of result, i.e. $l \geq \frac{1}{4}$ but $\{e^{i\lambda_n t} \}_{n\in \mathbb Z}$ still a Riesz basis for $L^2\left(-\pi,\pi\right)$.


There is a complete characterization of sequences $\lambda_n$ for which which $e^{i\lambda_n t}$ is a Riesz basis: G. Semmler, Complete interpolating sequences, discrete Muckenhoupt condition, and conformal mapping, which permits to construct many examples.

  • $\begingroup$ Thank you Alexandre Eremenko. I found the paper really interesting... I didn't know it! The main result of this paper is Theorem 6 which characterizes complete interpolating sequences. I'm not an expert in this filed but it seems that Th. 6 allows to construct many examples of the type I described above. But it also seems that this contruction depends on the knowledge of the zero set of the entire exponential function F. Am I right? $\endgroup$ – Mark Oct 13 '19 at 20:38
  • $\begingroup$ Theorem 6 gives a necessary and sufficient condition in completely elementary terms, without mentioning any entire function. $\endgroup$ – Alexandre Eremenko Oct 14 '19 at 23:06

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.