Let $\zeta$ be an $n$-th root of unity and let $\chi:=\zeta+\zeta^{-1}$. Then $\zeta^k+\zeta^{-k}=P_k(\chi)$ where $P_k\in\Bbb{Z}[X]$ is a polynomial not depending on $n$. For example we have \begin{eqnarray*} \zeta^2+\zeta^{-2}&=&\chi^2-2, \qquad&\text{ so }&\qquad P_2&=&X^2-2,\\ \zeta^3+\zeta^{-3}&=&\chi^3-3\chi, \qquad&\text{ so }&\qquad P_3&=&X^3-3X,\\ \zeta^4+\zeta^{-4}&=&\chi^4-4\chi^2+2, \qquad&\text{ so }&\qquad P_4&=&X^4-4X^2+2,\\ \zeta^5+\zeta^{-5}&=&\chi^5-5\chi^3+5\chi, \qquad&\text{ so }&\qquad P_5&=&X^5-5X^3+5,\\ \zeta^6+\zeta^{-6}&=&\chi^6-6\chi^4+9\chi^2+18, \qquad&\text{ so }&\qquad P_6&=&X^6-6X^4+9X^2+18. \end{eqnarray*} It isn't hard to see that $P_{ab}=P_a\circ P_b=P_b\circ P_a$ for all positive integers $a$ and $b$, and that we have a recurrence relation $$P_a=X^a-\sum_{i=1}\binom{a}{i}P_{a-2i},$$ where we take the convention that $P_k=0$ for all $k<0$, and $P_0=1$. My question is:

Is there a simple explicit expression for $P_k$?

  • 2
    $\begingroup$ This may help you: en.wikipedia.org/wiki/Chebyshev_polynomials $\endgroup$ – Seewoo Lee Nov 15 '18 at 16:40
  • $\begingroup$ @SeewooLee Can you turn that comment into an answer? $\endgroup$ – Pedro Tamaroff Nov 15 '18 at 16:43
  • $\begingroup$ @SeewooLee Thank you for the link. So if I understand correctly we have $P_k(X)=2T_k(\frac{X}{2})$, where $T_k$ is the $k$-th Chebyshev polynomial of the first kind? $\endgroup$ – Servaes Nov 15 '18 at 16:46
  • $\begingroup$ @SeewooLee Ok all is clear, thanks again. If you care to write up an answer, I'll gladly accept. $\endgroup$ – Servaes Nov 15 '18 at 17:09

As I said, Chebyshev polynomials are exactly what you said.


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.