1
$\begingroup$

Consider a sum of exponentials function of the integer $x$: $f(x)=A(x)+B(x)$, where $A(x)=\sum_{i=1}^n c_i \theta_i^x$ with $\theta_i$ roots of unity, and $B(x)=\sum_{j=1}^m d_j\lambda_j^x$ with $|\lambda_j|=1$ but not roots of unity. Assume the $\theta_i$'s and $\lambda_j$'s are all distinct, and that the coefficients $c_i,d_j$'s are all nonzero. It is known that $A(x)$ is a periodic function of period say $T_A$.

Now let us assume that $g(x):=|f(x)|$ is periodic of period $T_g$.

Question: what can be said about the relation between $T_g$ and $T_A$? In particular, is it true that $T_g\geq T_A$? Any reference would be appreciated.

$\endgroup$
2
  • 2
    $\begingroup$ Something here seems to be misstated. How is A(x) supposed to be periodic unless the $\theta_i$ are purely imaginary? Same question for $|f(x)|$. $\endgroup$ Commented Sep 17 at 3:03
  • $\begingroup$ @MichaelRenardy. Thanks, edited, sorry for the confusion. $\endgroup$
    – Michele
    Commented Sep 17 at 8:14

0

You must log in to answer this question.

Browse other questions tagged .