I know that I can easly solve this with the $4$th degree's equation, but isn't there a smarter way? It is an olympiad's problem so it shouldn't be a formula, but more find a formula... I have remarked that it is equal to $(x^2+1)\cdot(x+1)\cdot(x-1)+4x=0$ It isn't just solving the equation, I want to know if it can be solved otherwise than that... Could someone explain why this is equal to $(x^2+1)^2-(\sqrt{2}(x-1))^2$ please ?
-
It's actually equal to all that $+4x$, not $-4x$. Maybe just a typo, though. – Sep 06 '16 at 15:56
-
What is the olympiad question ? Solve the equation ? – Rene Schipperus Sep 06 '16 at 15:56
-
@tilper Yes I will edit it... – Sep 06 '16 at 15:57
-
Factorise into two quadratics – Macavity Sep 06 '16 at 15:57
-
@ReneSchipperus Yes... – Sep 06 '16 at 15:57
-
Not number theory, surely. – Macavity Sep 06 '16 at 15:58
-
@Macavity Yea you're right... What should I then put? – Sep 06 '16 at 16:00
-
@Omnomnomnom : O.k. :-) – user90369 Sep 06 '16 at 16:00
-
@DanielAlberten: See (22) in Quartic Equation of mathworld. – mathlove Sep 06 '16 at 16:17
-
Possible duplicate of Is there a general formula for solving 4th degree equations? – Claude Leibovici Sep 07 '16 at 05:51
-
No, I want to find a formula that just resolves this equation... Not even a formula but an easy way of find the solution... – Sep 07 '16 at 05:53
3 Answers
You may write it as $$ x^4+4x-1=(x^2+1)^2-2(x-1)^2 = (x^2+1)^2-(\sqrt{2}(x-1))^2$$ and factorize.
- 35,992
-
-
@DanielAlberten you have to learn to ask the right questions. Presumably, you are not asking how to verify that the solution is correct, but how one would come up with it in the first place. – Ben Grossmann Sep 06 '16 at 16:05
-
@H.H.Rugh How did you think to apply this trick to this problem? Is it a well-known approach? – Ben Grossmann Sep 06 '16 at 16:06
-
1Sort of. If a quartic equations is 'solvable' in finite time, it should be a product of 'identifiable' quadratics and the easiest ones are those that factors into $p^2-q^2$. Then look for $p$ and $q$. That works here (it helps on the guessing work that there are no $x^3$ nor $x^2$). If not, it gets more messy. – H. H. Rugh Sep 06 '16 at 16:08
If you don't want to rely on tricks (tricks are always useful though), a more systematic approach is as follows. Let $$x^4+4x-1 = (x^2+ax+b)(x^2+cx+d)$$ We then have \begin{align*} a+c & = 0\\ b+d+ac & = 0\\ bc + ad & = 4\\ bd & = -1 \end{align*} Eliminating $c$, we obtain \begin{align*} b+d & = a^2\\ d-b & = \dfrac4{a}\\ bd & = -1 \end{align*} This gives us $d = \dfrac{a^2}2 + \dfrac2a$ and $b = \dfrac{a^2}2 - \dfrac2a$. Since $bd=-1$, we obtain $$\dfrac{a^4}4 - \dfrac4{a^2} = -1 \implies a^6 - 16 = -4a^2$$ Setting $a^2=y$, we have $$y^3+4y - 16 = 0$$ By rational root theorem, we find that $y=2$ as one root. The other two roots are complex and hence we are not interested in them. This gives us $a=\pm \sqrt2$. Taking $a=\sqrt2$, we see $c=-\sqrt2$. This gives us $$d=1+\sqrt2 \text{ and }b = 1-\sqrt2$$ Hence, we obtain the factorization as $$x^4+4x-1 = (x^2+\sqrt2 x+1-\sqrt2)(x^2-\sqrt2 x+1+\sqrt2)$$
- 20,657
There is a tricky way since our polynomial can be easily written as a difference of two squares: $$ (x^2+1)^2-2(x-1)^2 = \left[x^2+x\sqrt{2}+(1-\sqrt{2})\right]\cdot \left[x^2-x\sqrt{2}+(1+\sqrt{2})\right]\tag{1}$$ hence by the quadratic formula the roots are given by $$ -\frac{1}{\sqrt{2}}\pm\sqrt{\sqrt{2}-\frac{1}{2}},\qquad \frac{1}{\sqrt{2}}\pm i\sqrt{\sqrt{2}+\frac{1}{2}}.\tag{2} $$
- 361,689