0
$\begingroup$

Is there efficient methods to find primitive-$n$-th roots of unities over $\mathbb{F}_a$??

In other word, find $\zeta$ such that,

$\zeta^n \equiv 1 $

where $\zeta \in \mathbb{F}_a$

Also, is there efficient methods to find primitive-$n$-th roots of unities over $\mathbb{Z}_a$??

In other word, find $\zeta$ such that,

$\zeta^n \equiv 1 \mod a$

where $\zeta \in \mathbb{Z}_a$

  • 1
    This depends on the size of the parameters. How is the field given to you? Often enough $\Bbb{F}_a$ is given as $\Bbb{F}_p[x]/(p(x))$, where $p(x)$ is a primitive polynomial. In that case an $n$th primitive root is given by the recipe $x^{(a-1)/n}+(p(x))$. Even with some black box description of $\Bbb{F}_a$ you can take the non-deterministic approach of raising random elements of the field to power $(a-1)/n$ (which is fast by square-and-multiply). Of course, quick testing of whether such a candidate actually is of order $n$ requires you to know a factorization of $n$. – 2017-02-12

1 Answers 1

0

I'm afraid that the answer is negative. If you would have such an algorithm then you'd have an efficient algorithm for factorizing large Mersenne's numbers...

  • 0
    By "efficient" I mean **computationally efficient**. – 2017-02-12
  • 0
    AKS is an efficient algorithm for testing primality. – 2017-02-12
  • 0
    Yes but what about factorisation of large Mersenne numbers? – 2017-02-12
  • 0
    Thanks all. Basically shouldn't we have to do brute force search? – 2017-02-12
  • 0
    As far I know yes – 2017-02-12
  • 0
    I edited my answer to take into account the remark of @mathma – 2017-02-12