Polynomials question (or rather moving stuff about) (1 Viewer)

DVDVDVDV

Premium Member
Joined
Jul 2, 2010
Messages
114
Gender
Male
HSC
2011
x^3 + 3x + 2 = 0 has roots a, b and c. Find the equation with roots a + 1/a, b + 1/b and c + 1/c.

I've substituted ((x +/- sqrt(x^2 - 4))/2 in place of x, but the solutions for the Cambridge book skips several steps and goes straight to

+/-sqrt(x^2 - 4)(4x^2 + 8) = -4x^3 - 16.

I just can't work out how they got the above equation.

Help please?
 

K4M1N3

Member
Joined
Jun 7, 2010
Messages
177
Gender
Male
HSC
2011
Just sub in : x + 1/x and solve the polynomial....

EDIT: nvm that doesnt work....ive resorted to finding a + 1/a + b + 1/b + c + 1/c ...etc etc

btw if your doing the further questions: 6....the answer is 2x^3 + 3x^2 + 8 = 0 (from the book)
 
Last edited:

largarithmic

Member
Joined
Aug 9, 2011
Messages
202
Gender
Male
HSC
2011
I assume they just subbed back in (x +/- root(x^2 - 4))/2 into the original equation and manipulated it using index laws/general algebra. You gotta be careful of the plusminus sign though but it should be alright? no tricks involved

the other way to do it would be to start from:
a+b+c = -3, ab+bc+ac = 0, abc = -2 and use Viete's formulae (Im not sure what they're called in school? 'product and sum of roots' or something?)

Youd get (a+1/a) + (b+1/b) + (c+1/c) = (a+b+c) + (1/a + 1/b + 1/c) = -3 + (ab+bc+ac)/abc = -3, so the x^2 term is 3.

(a+1/a)(b+1/b) + (a+1/a)(c+1/c) + (b+1/b)(c+1/c) = (ab+bc+ac) + (a/b + b/a + b/c + c/b + a/c + c/a) + (1/ab + 1/bc + 1/ac),
but (ab+bc+ac) = 0, and (a/b + b/a + b/c + c/b + a/c + c/a) = (a+b+c)(1/a + 1/b + 1/c) - 3 = (-3)(ab+ac+bc)/abc - 3 = (-3)(zero) - 3 = -3 and (1/ab + 1/bc + 1/ac) = (a+b+c)/abc = 3/2
So the x term is -3 + 3/2 = -3/2

And (a+1/a)(b+1/b)(c+1/c) = abc + 1/abc + ab/c + bc/a + ac/b + a/bc + b/ac + c/ab.
abc = -2, 1/abc = -1/2. Then ab/c + bc/a + ac/b = (ab+bc+ac)(1/a+1/b+1/c) - 2(a+b+c) = (-3)(zero, we've found 1/a +1/b+1/c before) - 2(-3) = 6. And also a/bc + b/ac + c/ab = (a+b+c)(1/ab + 1/bc + 1/ac) - (a/b + b/a + b/c + c/b + a/c + c/a) = (-3)(3/2, found this before) - (-3, found this before) = 3 - 9/2 = -3/2.

So the constant term is -(-2 -1/2 + 6 - 3/2) =- 2.

This gives an answer of x^3 + 3x - 3/2 x - 2 = 0, I have no idea whether thats right though coz this method could easily mess up the arithmetic. But this method will almost always work with enough algebraic manipulation; from the three original polynomial root/coefficient equations you can derive almost any symmetric expression in a,b,c; including the root/coefficient equations for the polynomial you want to form.
 

DVDVDVDV

Premium Member
Joined
Jul 2, 2010
Messages
114
Gender
Male
HSC
2011
Yeah, they subbed it in, but I can't do it with the plusminus sign unfortunately
 

largarithmic

Member
Joined
Aug 9, 2011
Messages
202
Gender
Male
HSC
2011
Yeah, they subbed it in, but I can't do it with the plusminus sign unfortunately
its not that bad, just act as if its a normal plus or minus sign. When you multiply a plusminus sign by a plusminus sign (i.e. you square it) you get a plus, and I think thats all you really need to know. Or just split it into two cases, 1. when its a plus, 2. when its a minus etc
 

math man

Member
Joined
Sep 19, 2009
Messages
503
Location
Sydney
Gender
Male
HSC
N/A
so you should work out that you have to sub a = (1/2)[x + sqrt(x^2 - 4)] into the original eqn...
doing so gives:
(1/8)[x + sqrt(x^2 - 4)]^3 + (3/2)[x + sqrt(x^2 - 4)] + 2 = 0 ...now i will times everything by 8
[x + sqrt(x^2 - 4)]^3 + 12[x + sqrt(x^2 - 4)] + 16 = 0 ...here expanding this cubic is nasty so what you should do is factorise the [x + sqrt(x^2 - 4)] from the first two terms
[x + sqrt(x^2 - 4)]{[x + sqrt(x^2 - 4)]^2 + 12]} + 16 = 0 ...now you expand the squared
[x + sqrt(x^2 - 4)]{x^2 + 2xsqrt(x^2 - 4) + x^2 - 4 + 12} + 16 = 0 ...now simply this
[x + sqrt(x^2 - 4)]{2x^2 + 2xsqrt(x^2 -4) +8} + 16 = 0 ...now im afraid we expand this..but this is easier than expanding the cubic
2x^3 + 2x^2sqrt(x^2 -4) + 8x + 2x^2sqrt(x^2 -4) + 2x(x^2 -4) + 8sqrt(x^2 -4) + 16 = 0 ...simply this obtains
2x^3 + 4x^2sqrt(x^2 -4) + 8x + 2x^3 -8x + 8sqrt(x^2 -4) + 16 = 0 ....simply this again
4x^3 + 4x^2sqrt(x^2 -4) + 8sqrt(x^2 -4) + 16 = 0 .....divide everything by 4 and factorise the sqrt(x^2 -4) from the 2 terms...
x^3 + 4 + sqrt(x^2 -4)[x^2 + 2] = 0 ...take the sqrt(x^2 -4) part to the other side and square both sides to finally eliminate the sqrt
(x^3 + 4)^2 = [-sqrt(x^2 -4)(x^2 + 2)]^2 ....expand this gives
x^6 + 8x^3 + 16 = (x^2 - 4)(x^4 + 4x^2 + 4) ...dont be scared with these x^6 and x^4 they will cancel as follows...
x^6 + 8x^3 + 16 = x^6 + 4x^4 +4x^2 - 4x^4 - 16x^2 - 16 ....simplyfying this gives...
8x^3 +12x^2 + 32 = 0 ...dividing through by 4 gives
2x^3 + 3x^2 + 8 = 0, which is the answer....this was very long working so dont be afraid if it takes long..this was a messy question
 

Alkanes

Active Member
Joined
May 20, 2010
Messages
1,417
Gender
Male
HSC
2012
lol you guys should learn how to use latex :p
 

math man

Member
Joined
Sep 19, 2009
Messages
503
Location
Sydney
Gender
Male
HSC
N/A
so you should work out that you have to sub
into the original eqn...
doing so gives:

...now i will times everything by 8

...here expanding this cubic is nasty so what you should do is factorise the
from the first two terms

...now you expand the squared part

...now simply this

...now im afraid we expand this..but this is easier than expanding the cubic

...simply this obtains

....simply this again

.....divide everything by 4 and factorise the
from the 2 terms...

...take the sqrt(x^2 -4) part to the other side and square both sides to finally eliminate the sqrt part

....expand this gives


...dont be scared with these x^6 and x^4 they will cancel as follows...

....simplyfying this gives...

...dividing through by 4 gives

, which is the answer....this was very long working so dont be afraid if it takes long..this was a messy question
 

seanieg89

Well-Known Member
Joined
Aug 8, 2006
Messages
2,662
Gender
Male
HSC
2007
My point is that you only end up with an equation that is satisfied at (a+1/a, b+1/b, c+1/c) after the step where you square both sides to eliminate the square root sign.
It is not obvious that this method is valid without use of the +- sign.

If you try the same method for say:
P(x)=x^3+3x-13/8 (which has root 1/2)

Then you will find that your equations before the squaring step i mentioned are not satisfied at x=2+1/2=5/2.
 

math man

Member
Joined
Sep 19, 2009
Messages
503
Location
Sydney
Gender
Male
HSC
N/A
My point is that you only end up with an equation that is satisfied at (a+1/a, b+1/b, c+1/c) after the step where you square both sides to eliminate the square root sign.
It is not obvious that this method is valid without use of the +- sign.

If you try the same method for say:
P(x)=x^3+3x-13/8 (which has root 1/2)

Then you will find that your equations before the squaring step i mentioned are not satisfied at x=2+1/2=5/2.
it doesnt matter which you use, because we have , where a is our root of the new eqn,
as a is our root P(a)= 0, so that means P()=P()=0
so you can use either cause both will give right answer
 

math man

Member
Joined
Sep 19, 2009
Messages
503
Location
Sydney
Gender
Male
HSC
N/A
i know where you are coming from lol...im just trying to point out both the + or - solution will give the same answer as they are both roots of the polynomial so you only need to sub in one of them to find the polynomial with the new desired roots
 

Users Who Are Viewing This Thread (Users: 0, Guests: 1)

Top