3
$\begingroup$

Recall that two cyclotomic integers are called associated if their ratio is a unit in the corresponding ring of cyclotomic integers.

We will view cyclotomic integers as polynomials (of degree $<\varphi(n)$) with integer coefficients modulo the cyclotomic polynomial $\Phi_n(x)$.

Q: Given a cyclotomic integer $p(x)$ modulo $\Phi_n(x)$, how to find all its associated cyclotomic integers $q(x)$ that have bounded coefficients?

Particularly interesting cases:

  • $q(x)$ has coefficients from $\{0,1\}$;

  • $q(x)$ has nonnegative coefficients summing to a given integer $s$.


Remark. A naive approach would be generating all polynomials $r(x)$ with bounded integer coefficients of degree at most $\varphi(n)-1$ (there is a finite number of them) and checking if $r(x)/p(x)$ is a unit, however it's hardly feasible for large $n$. I hope it's possible to somehow exploit the structure of units to recover the required polynomials $q(x)$ more efficiently.

$\endgroup$
12
  • $\begingroup$ As a side comment, the units are not easy to find/check: see this question. Also, you could try a simplified search by multiplying for the cyclotomic units which are quite explicit: planetmath.org/…. One could hope to find a bound on which cyclotomic units could possibly contribute (maybe some coefficient diverges when you multiply many times?). But then there is the problem of non-cyclotomic units, which implies a finite search from any candidate polynomial. $\endgroup$ Aug 27, 2022 at 22:07
  • 1
    $\begingroup$ Wouldn't it be better to bound the unit, say u, and compute all the u(x)p(x)? Using that the units form a f.g. group under multiplication will give a smaller search space. $\endgroup$ Aug 27, 2022 at 22:07
  • $\begingroup$ @AndreaMarino: As for checking for a unit, isn't it enough to check that the norm equals 1? $\endgroup$ Aug 27, 2022 at 22:19
  • $\begingroup$ @FelipeVoloch: The question here - how exactly to bound suitable units $u(x)$? $\endgroup$ Aug 27, 2022 at 22:20
  • 1
    $\begingroup$ If $q=up$, then $u=q/p$ in the cyclotomic field. The height satisfies $H(ab)\le H(a)H(b), H(1/a)=H(a)$ (Lang, Fund. Dioph Geom 3.1), so $H(u) \le H(q)H(1/p)=H(q)H(p)$. $\endgroup$ Sep 5, 2022 at 3:45

0

Your Answer

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