1

I am working on the following exercise. Solve the equation $1-x+x^{2}-x^{3}+x^{4}=y^{4}$ in $\mathbb{Z}$.

I have a couple of ideas for going about this exercise.

$1)$ By moving $1$ to the other side of the equation we obtain:

$y^4-1=x^4-x^3+x^2-x \rightarrow (y^2-1)(y^2+1)=x(x-1)(x^2+1)$.

The LHS gives two consecutive integers I am able to see, but other than that I am stuck and not sure where to go from here.

$2)$ We notice that $1-x+x^2-x^3+x^4=\Phi_{10}(x)=y^4$.

We now have that $\Phi_{10}(x) \geq 0$ since $y=\pm \sqrt[4]{1-x+x^2-x^3+x^4}$.

In this case, with some computation, $x={0,1}$ are the only possibilities that will force $y \in \mathbb{Z}$. From there, I use induction to show that $x \geq 2$ and $x \leq -1$ do not yield a perfect fourth.

In either case, I keep running into some issues. If someone could please other a hint to help me continue through this exercise, that would be very helpful. Thank you.

Update:

Hello, my progress on the following problem is as follows:

Suppose that $x=y$, then we get $y^3-y^2+y-1=0 \rightarrow y=1, \pm i$. Thus we get two solutions $(x,y)={(1,1),(1,-1)}$, the $y=-1$ since $y^4$ was in the original equation.

Next, using @Batominovski hint we can write:

$(2x^2-x)^2 < (x^4-x^3+x^2-x+1) < 4(x^4-x^3+x^2-x+1) \leq (2x^2-x+2)^2 \forall x \in \mathbb{R}$ which is true for $x=0$.

Since the multiplication of a constant (in this case $k=4$) does not affect a solution from existing, we have that $x=0$ is a solution, consequently we find two more solutions are $(x,y)=(0,1),(0,-1)$.

Joe
  • 955
  • From the description of your approach (2), I think it would work (see e.g., a related problem). What is the exact difficulty you had there? Btw they are special cases of Nagell-Ljunggren equation. – Wiley Jul 27 '16 at 17:24
  • Note that $y^4-1=(y-1)(y+1)(y^2+1)$. Not sure if that is useful. – smcc Jul 27 '16 at 17:27
  • @Wiley For showing $x \geq 2$ does not yield a perfect fourth inductively I have that

    $P(k+1)=k^4+3k^3+4k^2+2k+1=(k^4-k^3+k^2-k+1)+4k^3+3k^2+3k$

    where $P(k)=k^4-k^3+k^2-k+1$ is not a perfect fourth.

    From this point, I was going to have another (mini) inductive proof for showing

    $4k^3 +3k^2+3k$ is not a perfect fourth. In the second (mini) inductive proof I would also have the base case, true for all $x \geq 2$. I don't really see a more efficient way for this.

    – Joe Jul 27 '16 at 21:40

3 Answers3

4

Hint: Note that $$\left(2x^2-x\right)^2<4\left(x^4-x^3+x^2-x+1\right)\leq\left(2x^2-x+2\right)^2$$ for all $x\in\mathbb{R}$. The right-hand side becomes an equality if and only if $x=0$.

Batominovski
  • 49,629
  • 1
    $x(x-1)(x^2+1)=y^4-1\Rightarrow (x,y)=(1,\pm 1)$ is solution (it can be seen at first sight). – Piquito Jul 27 '16 at 20:24
  • @Batominovski I understand the set of inequalities however, how can we be certain that this will be the only solutions (knowing ($1,\pm1$)) is also a solution? – Joe Aug 05 '16 at 23:53
1

Hint: note that for $x$ big enough the following holds: $$(x-1)^4 < x^4-x^3+x^2-x+1 < x^4.$$ Similarly, for $x$ small enough we have $$(x-1)^4 > x^4-x^3+x^2-x+1 > x^4.$$ Therefore you are left with a finite number of possible values of $x$.

timon92
  • 11,329
1

One has $$1-x+x^{2}-x^{3}+x^{4}=y^{4}\iff(2x^2+1)^2=4y^4+4(x^3+x)-3$$ Notice that necessarily $y$ must be odd and $x^3+x$ is even when $x$ is even or odd. Hence one has $$(2x^2+1)^2\equiv 4y^4-3 \pmod 8$$ The solutions of this congruence are $$(x,y)=(x,1),(x,3),(x,5),(x,7)\text { with } x=0,1,2,3,4,5,6,7$$ we can verify that $\color{red}{(x,y)=(0,\pm1),(1\pm 1)}$ are the only solutions because for $y=1+8m,3+8m,5+8m,7+8m$ we have for $m=0$ these only solutions (easier to verify for other candidates with $m\gt 1$making $1-x+x^{2}-x^{3}+x^{4}=\frac{x^5+1}{x+1}$)

Piquito
  • 29,594
  • In your second line $(2x+1)^2 \equiv 4y^4 -3$ (mod 8). How did you come up with this equivalence modulo 8? I still cannot see that. Can you offer a hint? – Joe Aug 08 '16 at 14:23
  • 1
    Yes, Joseph. It is because of $x^3+x$ is always even (for $x$ being both odd and even). – Piquito Aug 09 '16 at 15:38