Re: quadratics with related roots
- From: john_ramsden@xxxxxxxxxxxxxx
- Date: 14 Jun 2005 10:27:49 -0700
Adam Atkinson wrote:
>
> [...]
>
> If the roots of the quadratic equation (insert quadratic here)
> are alpha and beta, write down a quadratic whose roots are (two
> formulas involving alpha and beta), without solving for alpha
> and beta.
>
> I've seen cases like asking for roots alpha^2 and beta^2
> or alpha/beta and beta/alpha.
Usually those kinds of problems are applications of the fact
that any symmetric polynomial of n variables can be expressed
as a polynomial in solely the elementary symmetric polynomials
of those variables.
Elementary symmetric (ES) polynomials of a set of n variables
{a_i} are the coefficients of powers of x (including x^0, i.e.
the constant term) in the expansion of:
(x - a_1).(x - a_2)... (x - a_n)
So for a quadratic with roots a_1, a_2 its 2 ES polynomials are
a_1 + a_2 and a_1.a_2, which we can denote by B_1 and B_0 resp.
If you're looking for a quadratic with roots a_1^2 and a_2^2
it's coefficients up to sign are a_1^2 + a_2^2 and a_1^2.a_2^2,
denoted by say C_1 and C_0 resp, which are symmetric polynomials
in a_1, a_2 and thus, by the result mentioned in the first para
above, both polynomials in B_1 and B_0. Sure enough its easy to
see that C_1 = B_1^2 - 2.B_0 and C_0 = B_0^2
.
- Follow-Ups:
- Re: quadratics with related roots
- From: Adam Atkinson
- Re: quadratics with related roots
- References:
- quadratics with related roots
- From: Adam Atkinson
- quadratics with related roots
- Prev by Date: Re: Alexander's formula of PI
- Next by Date: Re: prime k-tuplets
- Previous by thread: Re: quadratics with related roots
- Next by thread: Re: quadratics with related roots
- Index(es):
Relevant Pages
|