2
$\begingroup$

Factorize in $\mathbb{R}[x]$ or $\mathbb{Q}[x]$ :

$$P(x)=x^6+x^2+1$$

I know all roots of $P$ are complex and I don't know if there a factorization

But My try as following :

$$P(x)=x^6+x^2+1=(x^3)^2+1+2x^3-2x^3+x^2$$

$$=(x^3+1)^2+x^2-2x^3$$

I don't know how I complete?

$\endgroup$
1
  • 1
    $\begingroup$ @AderinsolaJoshua Thank you sir , very difficult root here $\endgroup$ Apr 3, 2020 at 13:15

4 Answers 4

8
$\begingroup$

For factorisation in $\mathbb{Q}[X]$:

As you note $X^6+X^2+1$ has no real roots.

Now $X^6+X^2+1=(X^3+X+1)^2$ modulo $2$, and this cubic is irreducible modulo $2$.

Hence, if the polynomial factorises over $\mathbb{Q}$ - and so over $\mathbb{Z}$ - it factorises as the product of two cubics. But every cubic has a real root, so we can't have such a factorisation.

$\endgroup$
2
  • $\begingroup$ Can you explain more for the final factorization? $\endgroup$ Apr 3, 2020 at 13:16
  • $\begingroup$ I don't understand: there is no rational factorisation... $\endgroup$ Apr 3, 2020 at 15:34
2
$\begingroup$

Hint

Since your polynomial has real coefficient, then $$p(X)=X^6+X^2+1$$ has roots $$\{z_1,\bar z_1, z_2,\bar z_2, z_3,\bar z_3\},$$ where $\bar z$ denotes the complex conjugate of $z$. Therefore you can write $p(X)$ as $$(X^2+a_1X+b_1)(X^2+a_2X+b_2)(X^2+a_3X+b_3)$$

where $$X^2+a_iX+b_i=(X-z_i)(X-\bar z_i).$$

$\endgroup$
1
  • $\begingroup$ Thank you sir , but roots here very difficult $\endgroup$ Apr 3, 2020 at 13:17
-1
$\begingroup$

Note that we can try if we have a purely imaginary root, and indeed we have this particular factorisation: $$(x^2+a)(x^4-ax^2+a^2+1)=x^6+x^2+a^3+a$$

Which fits the orignal polynomial for $\ a^3+a=1$, and this cubic has at least one real root that we can calculate by Cardan's method (I skip to the result though...)

Le set $\alpha=\sqrt[3]{108+12\sqrt{93}}\ $ then $\ a=\dfrac{\alpha}6-\dfrac 2{\alpha}$

Now we get to factorize the remaining quartic, there are formulas for this that should lead you to a factorization in $\mathbb R[x]$ in product of quadratics, but considering there are no odd powers we can try to factorize under the form:

$$(x^2+bx+c)(x^2-bx+c)=x^4+(2c-b^2)x^2+c^2$$

This gives us by identification of the coefficients (note that we are interested in one possible factorization, so I just take the positive square root each time) $\begin{cases}c=\sqrt{a^2+1}\\b=\sqrt{2c+a}\end{cases}$

And there you have it:

$$x^6+x^2+1=(x^2+a)(x^2+bx+c)(x^2-bx+c)$$

$\endgroup$
-3
$\begingroup$

It's root are not so difficult, because it is a reducible sextic, a bi-cubic to be précised and It's roots are

$$x_1 = -(\sqrt{2^{2/3}-(\sqrt{31}-3^{3/2})^{2/3}}i)/(2^{1/6}3^{1/4}(\sqrt{31}-3^{3/2})^{1/6})$$ $$x_2 = (\sqrt{2^{2/3}-(\sqrt{31}-3^{3/2})^{2/3}}i)/(2^{1/6}3^{1/4}(\sqrt{31}-3^{3/2})^{1/6})$$ $$x_3 = -\sqrt{\sqrt{3}(\sqrt{31}-3^{3/2})^{2/3}i+2^{2/3}\sqrt{3}i-(\sqrt{31}-3^{3/2})^{2/3}+2^{2/3}}/(2^{2/3}3^{1/4}(\sqrt{31}-3^{3/2})^{1/6})$$ $$x_4 = \sqrt{\sqrt{3}(\sqrt{31}-3^{3/2})^{2/3}i+2^{2/3}\sqrt{3}i-(\sqrt{31}-3^{3/2})^{2/3}+2^{2/3}}/(2^{2/3}3^{1/4}(\sqrt{31}-3^{3/2})^{1/6})$$ $$x_5 = -\sqrt{(-\sqrt{3}(\sqrt{31}-3^{3/2})^{2/3}i)-2^{2/3}\sqrt{3}i-(\sqrt{31}-3^{3/2})^{2/3}+2^{2/3}}/(2^{2/3}3^{1/4}(\sqrt{31}-3^{3/2})^{1/6})$$ $$x_6 = \sqrt{(-\sqrt{3}(\sqrt{31}-3^{3/2})^{2/3}i)-2^{2/3}\sqrt{3}i-(\sqrt{31}-3^{3/2})^{2/3}+2^{2/3}}/(2^{2/3}3^{1/4}(\sqrt{31}-3^{3/2})^{1/6})$$

So factoring the polynomial can be done in many ways

  1. if I factor it into two cubics, I'll need to include a square root extension from a quartic polynomial
  2. If I factor it into a quadratic and a quartic, I'll need to include a cube root extension from a sextic polynomial

Polynomial decomposition is helpful here, I factored it but the results was large and messy, so I'll post the simplest case

$x^6+x^2+1 = 0$

$x^4y-x^2y^2+1 = 0$

Where $y$ is the root of $y^3+y-1=0$

$\endgroup$

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.