I want to show that the order of a polynomial $x^4+ax^2+b$ is a divisor of $2(q^2-1)$ over a finite field with $q$ elements.
-
3What's stopping you? – Arkady Oct 23 '17 at 07:41
-
2What does "order of a polynomial" mean in this context? – Arthur Oct 23 '17 at 07:41
-
order of a polynomial f(x) in F_q[x] is defined as minimum e such that f(x)divides x^e−1 . – sekus Oct 23 '17 at 07:46
-
You need to assume that $b\neq0$ and that the quartic is not a square. Otherwise the claim may not hold. – Jyrki Lahtonen Oct 25 '17 at 06:21
1 Answers
Sorry about jotting down something in haste. In addition to the obvious missing assumption $b\neq0$ we actually need to assume that the zeros of the quadratic polynomial $p(x)=x^2+ax+b$ are distinct for the claim to hold. More details follow.
The polynomial $p(x)$ has its zeros, call them $\alpha_1$ and $\alpha_2$, in the field $K=\Bbb{F}_{q^2}$, so they have order dividing $q^2-1=|K^*|$.
If $\alpha_1\neq\alpha_2$ we thus have $$x^{q^2-1}\equiv1\pmod{x-\alpha_1}\quad\text{and}\quad x^{q^2-1}\equiv1\pmod{x-\alpha_1}. $$ By the Chinese Remainder Theorem this tells us that $$ x^{q^2-1}\equiv1\pmod{p(x)}, $$ and consequently (substitute $x^2$ for $x$ everywhere) $$ x^{2(q^2-1)}\equiv1\pmod{x^4+ax^2+b}. $$
The case $\alpha_1=\alpha_2$ is a bit different, because CRT doesn't bite as above when $x-\alpha_1$ and $x-\alpha_2$ are no longer coprime. It does follow that we then have $\alpha_1\in\Bbb{F}_q$, but the claim may actually be false in this case. The smallest counterexample is the case $p(x)=x^2+1=(x+1)^2$ when $q=2$. The original biquadratic is thus $x^4+1$, but this is obviously not a factor of $x^{2(q^2-1)}-1=x^6-1$. In this case (still assuming $b\neq0$) we have $$ x-\alpha_1\mid x^{q-1}-1, $$ so $$ p(x)\mid (x^{q-1}-1)^2\mid (x^{q-1}-1)^p=x^{p(q-1)}-1, $$ where $p$ is the characteristic. Again, plugging in $x^2$ for $x$ everywhere shows that the order of the quartic is a factor of $2p(q-1)$. Therefore the order is not necessarily a factor of $2(q^2-1)$. Another counterexample is the case $p(x)=(x+1)^2=x^2+2x+1$. It is easy to see that the order of the quartic $x^4+2x^2+1$ is $12\nmid 2(3^2-1)=16$.
- 133,153
-
Sorry about being a bit sketchy. A bridge match is about to begin, and I absolutely cannot spend more time now. – Jyrki Lahtonen Oct 23 '17 at 14:47
-
I should have respected the asker and my readers more, and not write down anything without due checking. Hopefully what I have here now will satisfy the downvoter also. – Jyrki Lahtonen Oct 25 '17 at 06:19