0
$\begingroup$

We can represent $\mathbb{F}_{q^m}$ as $\mathbb{F}_q[\alpha]$ where $\alpha$ is root of an irreducible $m$-degree polynomial on $\mathbb{F}_q$. By sending $\sum_{i=0}^{m-1} c_i\alpha^i \mapsto (c_{m-1},\cdots,c_1,c_0)$ we have a map $\mathcal{J}:\mathbb{F}_{q^m}\mapsto(\mathbb{F}_q)^m$.

Let $\mathcal{P}_q$ be polynomial ring of polynomials having degree less than $q$ on $\mathbb{F}_q$. (we have $x^q=x$)

Let $\mathcal{P}_{q^m}$ be polynomial ring of polynomials having degree less than $q^m$ on $\mathbb{F}_q$. (we have $x^{q^m}=x$)

Then, we have a map $f:\mathcal{P}_{q^m} \rightarrow (\mathcal{P}_q)^m$. The map works as

  1. For any constant $c \in \mathbb{F}_{q^m}$, $f(c)=\mathcal{J}(c)$
  2. For any constant $c \in \mathbb{F}_{q^m}$ we have $\mathcal{J}(cx)=C\mathcal{J}(x)$ where $C$ is an $m\times m$ matrix
  3. $f$ sends $x^2\in \mathcal{P}_{q^m}$ to a quadratic $m$-variable $m$ polynomials.

Now, we can represent any polynomial in $\mathcal{P}_{q^m}$ as a polynomial in $(\mathcal{P}_q)^m$. I need the inverse of $f$. Especially, how can we calculate the $f^{-1}(Q)$ when $Q$ is a quadratic polynomial in $(\mathcal{P}_q)^m$.

We can prove that $f$ is invertible because it is one-to-one and the polynomial rings have the same order.

Let $m=50$ and $q=7$, then I feel that an algorithm calculating the map $f^{-1}$ has impractical time complexity. However, I hope that for a quadratic $Q$ there may be a practical way to find $f^{-1}(Q)$.

$\endgroup$
4
  • 1
    $\begingroup$ You switched between $F_q$ and $\mathbb F$ with no subscript for the field that I think, for consistency with $\mathbb F_{q^m}$, you would mean to denote by $\mathbb F_q$. I edited for consistency, I hope correctly. But I do not understand what you mean when you say that $(\mathbb F[x_1, \dotsc, x_m])^m$ (which I take to mean $(\mathbb F_q[x_1, \dotsc, x_m])^m$) has $(q^{q^m - 1})^m$ elements, or that $\mathbb F_{q^m}[x]$ has $(q^m)^{q^m - 1}$ elements; like all polynomial rings over non-trivial rings, these are infinite. Also, what is your map $\mathbb F_{q^m} \to \mathbb F_q^m$? $\endgroup$
    – LSpice
    Commented Oct 4, 2022 at 18:41
  • 1
    $\begingroup$ In addition there is no "natural map" as you say since although there is a $\mathbf{F}_q$-linear isomorphism $\mathbf{F}_{q^n}\to(\mathbf{F}_q)^n$, there is no "natural" one. Fixing such an isomorphism is equivalent to choosing a basis. $\endgroup$
    – YCor
    Commented Oct 4, 2022 at 22:27
  • $\begingroup$ The OP presumably means the rings of polynomial functions, which are finite. @LSpice $\endgroup$ Commented Oct 5, 2022 at 8:46
  • $\begingroup$ @LSpice Now, I edited. I hope it make sense. $\endgroup$
    – user
    Commented Oct 5, 2022 at 10:00

0

You must log in to answer this question.

Browse other questions tagged .