Does anyone know of a convergence test for a complex series of the form

$$\sum_n a_n \cdot \exp(i \cdot b_n)$$


The particular series I need to understand has a_n going to zero as n goes to infinity, but it fails the absolute convergence test. However numerically I do find it to converge. It should have something to do with convergence of Fourier series (since $\ \exp(i\cdot t)\ =\ \cos(t)\,+\,i\cdot \sin(t)$).

  • 2
    $\begingroup$ This seems very vague to me. Why don't you describe the particular series? $\endgroup$ – Yemon Choi Sep 19 '14 at 21:55
  • $\begingroup$ I need something stronger than the Dirichlet test, since my series fails it, but still converges! Are such examples known? $\endgroup$ – André LeClair Sep 20 '14 at 17:35
  • 1
    $\begingroup$ Somehow, this got put on hold, but whoever suggested summation by parts, thank you! It did the trick. $\endgroup$ – André LeClair Oct 15 '14 at 18:14

If your $a_n$s are nonnegative, decrease monotonically, and approach zero, and the $b_n$s are such that $\sum e^{i x b_n}$ remains bounded, then Dirichlet's test would apply here and give you convergence. If the $b_n$'s are an arithmetic sequence, as in the case of Fourier series, then you have the second of these criteria (for $x\neq0$).

  • $\begingroup$ Yoav, thanks. But my series fails the Dirichlet test but still converges! Any ideas? $\endgroup$ – André LeClair Sep 20 '14 at 17:41
  • $\begingroup$ I refined the question in my next post. $\endgroup$ – André LeClair Sep 20 '14 at 21:09
  • $\begingroup$ Apparently the Dirichlet test is not an "if and only" statement. My series fails it, but still converges. Is that right? $\endgroup$ – André LeClair Sep 20 '14 at 21:21

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.