4

Real values of $x$ satisfying the equation $$x^9+\frac{9}{8}x^6+\frac{27}{64}x^3-x+\frac{219}{512} =0$$

We can write it as $$512x^9+576x^6+216x^3-512x+219=0$$

I did not understand how can i factorise it.

Help me

amWhy
  • 209,954
juantheron
  • 53,015
  • 1
    One is tempted to set $u = 2x$, and then $u^9+9u^6+27u^3-256u+219 = 0$. One can quickly obtain $u = 1$ in this manner. There are, however, at least two more real roots floating around in there. – Brian Tung Jan 19 '16 at 03:20
  • @BrianTung After dividing by $u-1$ those would be among the roots of $u^8 + u^7 + u^6 + 10 u^5 + 10 u^4 + 10 u^3 + 37 u^2 + 37 u - 219 = 0$. There is an explicit factorization which gives the real roots (http://www.wolframalpha.com/input/?i=u%5E8+%2B+u%5E7+%2B+u%5E6+%2B+10+u%5E5+%2B+10+u%5E4+%2B+10+u%5E3+%2B+37+u%5E2+%2B+37+u+-+219+%3D+0) though that's not exactly obvious by inspection, to me at least. – dxiv Jan 19 '16 at 03:35
  • 1
    Where did this come from? Why do you think it can be factored? The fact that $x=2$ is a root may be a clue. – Ross Millikan Jan 19 '16 at 03:56
  • $x=1/2$ is a root, not $x=2$. – amcerbu Jan 19 '16 at 04:45
  • Before dividing by $u-1$, we can also factor the expression in $u$ as $(u^3+3)^3-4^3(4u-3)$. Not sure if that buys us anything, though. – Brian Tung Jan 19 '16 at 17:18

3 Answers3

5

$\bf{I\; have \;Solved\; Like \;This \; Way}$

$$x^9+\frac{9}{8}x^6+\frac{27}{64}x^3-x+\frac{219}{512} =0\Rightarrow 512x^9+(9\cdot 64)x^6+(27\cdot 8)x^3-512x+219=0$$

So $$\underbrace{(8x^3)^3+3(8x^3)^2\cdot 3+3(3^2)\cdot 8x^3+3^3}-512x+219-3^3=0$$

So $$(8x^3+3)^3=512x-219\Rightarrow (8x^3+3)^3=512\left(x-\frac{192}{512}\right)=8^3\left(x-\frac{3}{8}\right)^{\frac{1}{3}}.$$

So $$8^3\left(x^3+\frac{3}{8}\right) = 8\left(x-\frac{3}{8}\right)^{\frac{1}{3}}\Rightarrow x^3+\frac{3}{8} = \left(x-\frac{3}{8}\right)^{\frac{1}{3}}$$

Now Let $\displaystyle f(x)=x^3+\frac{3}{8},$ Where $f:\mathbb{R}\rightarrow \mathbb{R}\;,$ Then $\displaystyle f^{-1}(x) = \left(x-\frac{3}{8}\right)^{\frac{1}{3}}\;,$ Where $f:\mathbb{R}\rightarrow \mathbb{R}\;$

So We have To solve $$f(x) = f^{-1}(x)$$

Now We now that $f(x)$ and $f^{-1}(x)$ is Symmetrical about $y=x$ line.

So $$f(x) = f^{-1}(x) =x$$

So $$x^3+\frac{3}{8}=x\Rightarrow 8x^3-8x+3=0$$

So $$(2x-1)\left[4x^2+2x-3\right]=0\Rightarrow x=\frac{1}{2}\;\;,x=\frac{-1\pm \sqrt{13}}{4}$$

juantheron
  • 53,015
  • Nice way to solve. However, what about other possible real roots? Or, maybe, I can´t see it is implicitly answered? – Piquito Jan 27 '16 at 20:57
2

If this problem can be solved without computation it is reducible; we assume this.

$f(x)=512x^9+576x^+216x^3-512x+219=0$ has two change-sign and $f(-x)$ has three ones so $f(x)$ has at least $9-5=4$ non real roots. We try to find a quadratic factor using the fact that $219=3\cdot73$ and $512=2^9$; this factor could correspond to real or non-real roots.

Trying with $4x^2+ax\pm 3$ we find at once that $a=2$ and the sign minus fits; furthermore $4x^2+2x-3=0$ has two real roots because $\Delta=1+12>0$.

The quotient gives $$128x^7-64x^6+128x^5+32x^4+80x^3-16x^2+122x-73=0$$ and this equation has necessarily a real root because $7$ is odd; assuming this 7-degree polynomial is reducible and noticing that: $$\begin{cases}(2x)^7-(2x)^6+4(2x)5+2(2x)^4+10(2x)^3-4(2x)^2+61(2x)-73=0\\1-1+4+2+10-4+61-73=0\end{cases}$$ it follows at once that $2x=1$ gives a third real root.

Dividing again, now by $(2x)-1$ one gets $$g(x)=(2x)^6+4(2x)^4+6 (2x)^3+16(2x)^2+12(2x)+73=0$$ It is obvious that $g(x)>0$ for $x>0$ and it is easy to show that for $X<0$ $$X^6+4X^4+16X^2+73>-(6X^3+12X)$$ hence $g(x)$ is always positive so $g(x)=0$ has six non-real roots.

Thus $f(x)=0$ has only three real roots, given by $$\color{red}{4x^2+2x-3=0}\space \text {and}\space\space \color{red}{ 2x-1=0}$$

Piquito
  • 29,594
1

Using factor theorem:

If $x-a$ divides $f(x)$, then $f(a)=0$.

As $(\frac{1}{2})^{9}=\frac{1}{512}$, it's worth to try $x=\pm \frac{1}{2}$.

Substituting $x=\frac{1}{2}$, LHS vanishes, so $\frac{1}{2}$ is one possible value.

Further factorize gets, $$\left( x-\frac{1}{2} \right) \left( x^{2}-\frac{x}{2}-\frac{3}{4} \right) \left( x^{6}+x^{4}+\frac{3x^{3}}{4}+x^{2}+\frac{3x}{8}+\frac{7}{3} \right)=0$$ and the second factor can be solved by quadratic formula.

Ng Chung Tak
  • 18,990
  • OK...I get the first step that since $\frac{1}{2}$ is a factor so you adjusted terms accordingly to get $(x-1/2)(an//expression...)=0$ but how did you factorise that other term ..? – Freelancer Jan 19 '16 at 06:09
  • Actually, I'm getting help from computer and which is programmed according to some theories such as finite field. The sextic seems cannot be further reduced. – Ng Chung Tak Jan 19 '16 at 06:18
  • No...I have no problem with the polynomial of sixth degree in your factorisation ...I just wanted to ask how you reduced this expression... http://www4b.wolframalpha.com/Calculate/MSP/MSP59931g5973i6e0g950hg000027c6a0f4ed0f8a3f?MSPStoreType=image/gif&s=29&w=378.&h=41. – Freelancer Jan 19 '16 at 06:31
  • Was this also... reduced with the help of a computer ? – Freelancer Jan 19 '16 at 06:33
  • Type Factor[.], which is Mathematica language. http://www.wolframalpha.com/input/?i=Factor%5Bx%5E9%2B9x%5E6%2F8%2B27x%5E3%2F64-x%2B219%2F512%5D – Ng Chung Tak Jan 19 '16 at 07:14
  • @Freelancer Also see the link in my comment under the OP. – dxiv Jan 19 '16 at 07:37