1
$\begingroup$

Let $P(z)=\sum_{k=0}^na_kz^k $ be a polynomial of degree $n$ and $z_k (1\leq k\leq n)$'s be $n$th roots of $-1$. Then when $\theta=0$ the inequality $$|P'(e^{i\theta})|\leq \frac{4}{n}\left|\sum_{k=1}^nP(e^{i\theta}z_k)\frac{z_k}{(z_k-1)^2}\right|$$ does not hold. Suppose we modify it as $$|P'(e^{i\theta})|\leq \frac{4}{n}\sum_{k=1}^n\left|\frac{z_k}{(z_k-1)^2}\right| \max_{|z|=1}|P(z)|.$$ Is this modification true?

$\endgroup$
4
  • $\begingroup$ What is the summands in RHS corresponding to $z_k=1$? $\endgroup$ Sep 3, 2018 at 6:02
  • 3
    $\begingroup$ This is wrong. Choose $P(z)=z^n+1$ and $\theta=0$. Then the RHS vanishes by construction. The LHS does not vanish since $P'(z)=nz^{n-1}$. $\endgroup$ Sep 3, 2018 at 7:47
  • 4
    $\begingroup$ Numerical experiments suggest $$\sum_{k=1}^n\left\lvert\frac{z_k}{(z_k-1)^2}\right\rvert=\sum_{k=1}^n\left\lvert\frac{1}{(z_k-1)^2}\right\rvert=\frac{n^2}{4}.$$ This substitution transforms the statement into a Theorem of Bernstein. $\endgroup$ Sep 3, 2018 at 14:55
  • 3
    $\begingroup$ @PhilippLampe: I suggest that you turn your comment into an answer, so that this question can be closed. $\endgroup$
    – GH from MO
    Sep 3, 2018 at 19:36

1 Answer 1

4
$\begingroup$

Terms that are very similar to the terms given here can be found in an article by Aziz and Mohammad, see [1].

Suppose that $a\neq -1$ is a complex number of modulus $1$. Let us denote the roots of the polynomial $z^n+a$ by $z_1,\ldots,z_n$. The authors show that \begin{align*} \left\lvert\frac{(1+a)^2}{na}\right\rvert\sum_{k=1}^n\left\lvert\frac{z_k}{(z_k-1)^2}\right\rvert=n. \end{align*} We put $a=1$. Then the right hand side in the considered inequality becomes $n\operatorname{max}_{\lvert z\rvert=1}\lvert P(z)\rvert$.

The desired inequality follows from a Theorem of Bernstein, see [2], which states $$\operatorname{max}_{\lvert z\rvert=1}\vert P'(z)\rvert \leq n\operatorname{max}_{\lvert z\rvert=1}\lvert P(z)\rvert.$$

  1. A. Aziz and Q. G. Mohammad: Simple Proof of a Theorem of Erdős and Lax, Proc. Amer. Math. Soc. 80, no. 1 (1980), 119--122
  2. S. Bernstein: Sur la limitation des dérivées des polynomes, C. R. Math. Acad. Sci. Paris 190 (1930), 338--340
$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.