2
$\begingroup$

If we consider $0<r<1$ and $j\in\mathbb{N}$, is this inequality true?

$$\left|\sum_{k=0}^jr^ke^{ikt}\right|\leq \left|\sum_{k=0}^je^{ikt}\right|$$


EDIT: A counterexample is in the comments. I'm wondering, however, if there is actually an inequality when you take integrals:

$$\int_0^{2\pi}\left|\sum_{k=0}^jr^ke^{ikt}\right|dt\leq \int_0^{2\pi}\left|\sum_{k=0}^je^{ikt}\right|dt$$

$\endgroup$
  • 3
    $\begingroup$ $j = 1, t = \pi, r = .5$ $\endgroup$ – mathworker21 Oct 3 '18 at 17:11
  • $\begingroup$ Thanks. Could be possible that although the inequality is not true for all $t$, the integrals of those modulus from $0$ to $2\pi$ actually held the inequality ? $\endgroup$ – Mark_Hoffman Oct 3 '18 at 17:19
  • 2
    $\begingroup$ not an answer, but just wanted to let you know that $$\int_0^{2\pi} |\sum_{k=0}^j r^ke^{ikt}|^2 dt \le \int_0^{2\pi} |\sum_{k=0}^j e^{ikt}|^2dt.$$ Indeed, the LHS is $\sum_{k=0}^j r^{2k}$ while the RHS is $j$. $\endgroup$ – mathworker21 Oct 3 '18 at 17:29
0
$\begingroup$

Going to answer myself. The inequality with the integrals is true. If you call $g(t):=\sum_{k=0}^je^{ikt}\in P(\mathbb{T})\subset L^1(\mathbb{T})$ and denote the Poisson kernel by $P_r(t)=\sum_{k\in\mathbb{Z}}r^{|k|}e^{ikt}$, using Young's inequality for $p=1$ you have:

$$\int_0^{2\pi}\left|\sum_{k=0}^jr^ke^{ikt}\right|dt=\|P_r\ast g \|_1\leq \|P_r\|_1\cdot\|g\|_1= \|g\|_1=\int_0^{2\pi}\left|\sum_{k=0}^je^{ikt}\right|dt$$

| cite | improve this answer | |
$\endgroup$
  • 1
    $\begingroup$ So this shows the inequality for any $L^p$ norm, right? $\endgroup$ – mathworker21 Oct 4 '18 at 17:45
  • $\begingroup$ Yes, that is correct, for $1\leq p \leq\infty$. $\endgroup$ – Mark_Hoffman Oct 8 '18 at 8:47

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.