Quartic with no rational solutions

Algebra Level 5

x 4 2 x 2 8 x 3 = 0 x^4 - 2x^2 - 8x - 3 = 0

Let a , b , c a,b,c and d d denote the roots of the equation above. And suppose we denote 2 n + 2 m = a + b + c + d 2\sqrt n + 2\sqrt m = |a| + |b| + |c| + |d| for positive integers m m and n n , find the value of m 2 + n 2 m^2 + n^2 .


The answer is 13.

This section requires Javascript.
You are seeing this because something didn't load right. We suggest you, (a) try refreshing the page, (b) enabling javascript if it is disabled on your browser and, finally, (c) loading the non-javascript version of this page . We're sorry about the hassle.

3 solutions

Chew-Seong Cheong
Oct 23, 2015

Let f ( x ) = x 4 2 x 2 8 x 3 f(x) = x^4-2x^2-8x-3 . If f ( x ) f(x) can be factorized into product of polynomials with rational coefficients, it could either be one linear factor with a cubic factor or two quadratic factors. We note that f ( ± 1 ) 0 f(\pm1) \ne 0 , and f ( ± 3 ) 0 f(\pm3) \ne 0 . By rational root theorem, this means that f ( x ) f(x) has no linear factor and therefore it is of the form ( x 2 + A x + C ) ( x 2 + B x + D ) (x^2 + Ax + C) (x^2+Bx+D)

Assume x 4 2 x 2 8 x 3 = ( x 2 + A x ± 1 ) ( x 2 + B x 3 ) = x 4 + ( A + B ) x 3 + ( A B ± 1 3 ) x 2 + ( 3 A + ± B ) x 3 \begin{aligned} \text{Assume } x^4-2x^2-8x-3 & = (x^2 + Ax \pm 1) (x^2+Bx \mp3) \\ & = x^4 + (A+B)x^3 + (AB \pm 1 \mp 3)x^2 + (\mp 3A + \pm B)x -3 \end{aligned}

Equating coefficients on both sides, we have:

\(\begin{array} {} A+B = 0 & \Rightarrow B = -A \\ AB = -A^2 & = \begin{cases} -1+3-2 = 0 & \Rightarrow A = 0 & \color{red}{\text{rejected}} \\ 1-3-2 = -4 & \Rightarrow A = \pm 2 & \color{blue}{\text{accepted}} \end{cases} \end{array} \)

\(\begin{array} {} \Rightarrow A = \pm 2 \quad B = \mp 2 \quad C = -1 \quad D = 3 \end{array} \)

And note that x 4 2 x 2 8 x 3 = ( x 2 + 2 x 1 ) ( x 2 2 x + 3 ) = 0 \begin{aligned} \text{And note that } x^4-2x^2-8x-3 & = (x^2 + 2x -1) (x^2-2x+3) = 0 \end{aligned}

And the roots are: x = { 2 ± 4 + 4 2 = { a = 1 + 2 a = 2 1 b = 1 2 b = 1 + 2 2 ± 4 12 2 = { c = 1 + i 2 c = 3 d = 1 i 2 d = 3 x = \begin{cases} \dfrac{-2 \pm \sqrt{4+4}}{2} & = \begin{cases} a = -1 + \sqrt{2} & \Rightarrow |a| = \sqrt{2} - 1 \\ b = -1 - \sqrt{2} & \Rightarrow |b| = 1 + \sqrt{2} \end{cases} \\ \dfrac{2 \pm \sqrt{4-12}}{2} & = \begin{cases} c = 1 + i\sqrt{2} & \Rightarrow |c| = \sqrt{3} \\ d = 1 - i\sqrt{2} & \Rightarrow |d| = \sqrt{3} \end{cases} \end{cases}

Therefore, a + b + c + d = 2 2 + 2 3 m 2 + n 2 = 13 \quad |a|+|b|+|c|+|d| = 2\sqrt{2} + 2\sqrt{3}\quad \Rightarrow m^2 + n^2 = \boxed{13}

nice solution sir. upvoted. would you perhaps mind to check this note: https://brilliant.org/discussions/thread/an-easier-way-to-solve-quartics/?ref_id=964870

Aareyan Manzoor - 5 years, 7 months ago

Log in to reply

You can key in [link title] (link address), but without the space between [] and (), to convert it to be a actual link like here .

Chew-Seong Cheong - 5 years, 7 months ago

Log in to reply

sir would you like to see this note

Aareyan Manzoor - 5 years, 7 months ago

You need to explain why the quartic polynomial don't have any linear factor with rational coefficients.

Pi Han Goh - 5 years, 7 months ago

Log in to reply

The product of two quadratic equations should work for all cases. If there are linear factors, it would factorize out.

Chew-Seong Cheong - 5 years, 7 months ago

Log in to reply

No. not necessarily. You could factor a quartic polynomial to a product of a linear factor and a cubic factor with this cubic factor couldn't be factored any further. But this is not the case for the quartic polynomial in question.

The simplest way to show that is via Rational Root Theorem .

Pi Han Goh - 5 years, 7 months ago

Log in to reply

@Pi Han Goh You are right. But it has mentioned that there are four roots and that if there is only one linear factor, the assumption would have failed.

Chew-Seong Cheong - 5 years, 7 months ago

Log in to reply

@Chew-Seong Cheong It mentioned that there are four roots. But that doesn't indicate whether the roots came from 2 quad eqns or 1 linear eqn and 1 cubic eqn.

Pi Han Goh - 5 years, 7 months ago

Log in to reply

@Pi Han Goh Thanks. Yes, you are right. I was wrong about the 4 roots, but the assumption of two quadratic equations worked means there is no linear factor. Anyway, it is better to explain there is no rational root first, so that we don't have to assume. I am changing the solution.

Chew-Seong Cheong - 5 years, 7 months ago

Log in to reply

@Chew-Seong Cheong No no. You're doing it wrong. By Rational root theorem, you just need to show that f ( ± 1 ) , f ( ± 3 ) 0 f(\pm 1), f(\pm 3) \ne 0 . That's all. Do you know why I chose ± 1 , ± 3 \pm1, \pm3 only?

Pi Han Goh - 5 years, 7 months ago

Log in to reply

@Pi Han Goh Sorry, the ± 2 \pm 2 was a typo. I wasn't sure about ± 1 3 \pm \frac{1}{3} . Is it because if the root r = a b r = \frac{a}{b} , then a p 0 a|p_0 and b p n b|p_n . So a = ± 3 a = \pm 3 and b = ± 1 b = \pm 1 ?

Chew-Seong Cheong - 5 years, 7 months ago

Log in to reply

@Chew-Seong Cheong You're almost right. You can use this formula below which is less fanciful (it actually conveys the same thing).

If there exist a rational root of x x of a polynomial with integer coefficients, then x = ± factors of constant factors of leading coefficient x = \pm \dfrac{ \text{ factors of constant}}{\text{factors of leading coefficient}} .

So in this case, the constant is -3 and the leading coefficient is 1, so the factors of -3 are (consider the positive value only) 1 and 3, and the factor of leading coefficent of 1 is simply 1. This means we have x = ± 1 , 3 1 = ± 1 , ± 3 x = \pm \dfrac{1,3}{1} = \pm1,\pm3 . Hope this clears things up!

For another example, recall your old problem , we should check for f ( ± 1 ) , f ( ± 2 ) , f ( ± 3 ) , f ( ± 4 ) , f ( ± 6 ) , f ( ± 12 ) , f ( ± 24 ) = 0 f(\pm1), f(\pm2), f(\pm3), f(\pm4) , f(\pm6) , f(\pm12), f(\pm24) = 0 .

Pi Han Goh - 5 years, 7 months ago

Log in to reply

@Pi Han Goh Thanks. Got it. The previous examples that I read were unclear.

Chew-Seong Cheong - 5 years, 7 months ago

Log in to reply

@Chew-Seong Cheong Ahah no problem! If you're don't understand what people wrote, you can't always ask. That's what the solution discussion is for.

By the way, you should rephrase your solution slightly more by mentioning that if it could be factored into products of polynomials with rational coefficients then it must either have a linear factor or a quadratic factor. Because there's a possibility that that the quartic polynomial can't be further factored.

Pi Han Goh - 5 years, 7 months ago
Arjen Vreugdenhil
Oct 31, 2015

I started writing the polynomial as a product ( x 2 b x + c 1 ) ( x 2 + b x + c 2 ) , (x^2 - bx+c_1)(x^2+bx+c_2), realizing that the linear terms should be opposites because the cubic term vanishes. It would be really nice if c 1 c_1 and c 2 c_2 are integers; so I factored the constant term as ( 1 ) 3 (-1)\cdot 3 or ( 3 ) 1 (-3)\cdot 1 . But we don't want c 1 + c 2 = 2 c_1+c_2 = -2 , because then the contribution b x 2 -bx^2 to the quadratic term would be zero, and there would be no linear term. Therefore I chose c 1 + c 2 = 2 c_1 + c_2 = 2 , i.e. c 1 = 1 , c 2 = 3 c_1 = -1, c_2 = 3 . It was now easy to decide on b = 2 b = 2 , so that the equation factors as ( x 2 2 x 1 ) ( x 2 + 2 x + 3 ) = 0. (x^2 - 2x - 1)(x^2 + 2x + 3) = 0. The solutions are now easily found to be 1 ± 2 1\pm\sqrt 2 and 1 ± 2 i -1\pm\sqrt 2 i , with absolute values 2 ± 1 \sqrt 2\pm 1 and 3 \sqrt 3 (twice). Therefore a + b + c + d = 2 2 + 2 3 , |a|+|b|+|c|+|d| = 2\sqrt 2 + 2\sqrt 3, and the solution is 2 2 + 3 2 = 13 2^2 + 3^2 = \boxed{13} .

Aareyan Manzoor
Oct 22, 2015

by the note ,we are to find that the cubic is q ( z ) = z 3 2 2 z 2 + 2 2 + 4 3 16 z 8 2 64 = z 3 z 2 + z 1 q(z)=z^3-\frac{2}{2}z^2+\frac{2^2+4*3}{16}z-\frac{8^2}{64}=z^3-z^2+z-1 we see that a 2 = 8 a_2=-8 which is negative telling us that we are looking for positive square root sum of q's roots. z 3 z 2 + z 1 = z 2 ( z 1 ) + 1 ( z 1 ) = ( z 1 ) ( z 2 + 1 ) = ( z 1 ) ( z + i ) ( z i ) z^3-z^2+z-1=z^2(z-1)+1(z-1)=(z-1)(z^2+1)=(z-1)(z+i)(z-i) so the roots are 1,-i,i. we find that x = 1 + i + i = 1 + 2 x=\sqrt{1}+\sqrt{i}+\sqrt{-i}=1+\sqrt{2} we put this value and get zero.we know that in most polynomails, if k + s k+\sqrt{s} is a root, then k s k-\sqrt{s} is another root. we put in x= 1 2 1-\sqrt{2} and also get zero. next we divide x 4 2 x 2 8 x 3 ( x ( 1 + 2 ) ) ( x ( 1 2 ) ) = x 4 2 x 2 8 x 3 x 2 2 x 1 = x 2 + 2 x + 3 \frac{x^4-2x^2-8x-3}{(x-(1+\sqrt{2}))(x-(1-\sqrt{2}))}=\frac{x^4-2x^2-8x-3}{x^2-2x-1}=x^2+2x+3 we factor that part to get the othe two roots 1 i 2 , 1 + i 2 -1-i\sqrt{2},-1+i\sqrt{2} . you can do the rest

From which note?

Aakash Khandelwal - 5 years, 7 months ago

0 pending reports

×

Problem Loading...

Note Loading...

Set Loading...