Difference between revisions of "Newton's Sums"
Runyangwang (talk | contribs) |
(→Statement: Edited elementary sum form of Newton Sums to be more clear) |
||
Line 24: | Line 24: | ||
<cmath>P_1 = S_1</cmath> | <cmath>P_1 = S_1</cmath> | ||
<cmath>P_2 = S_1P_1 - 2S_2</cmath> | <cmath>P_2 = S_1P_1 - 2S_2</cmath> | ||
+ | <cmath>P_3 = S_1P_2 - S_2P_1 + 3S_3</cmath> | ||
+ | <cmath>\vdots</cmath> | ||
etc., where <math>S_n</math> denotes the <math>n</math>-th [[elementary symmetric sum]]. | etc., where <math>S_n</math> denotes the <math>n</math>-th [[elementary symmetric sum]]. |
Revision as of 01:57, 6 August 2021
Newton sums give us a clever and efficient way of finding the sums of roots of a polynomial raised to a power. They can also be used to derive several factoring identities.
Contents
Statement
Consider a polynomial of degree
,
![$P(x) = a_nx^n + a_{n-1}x^{n-1} + \cdots + a_1x + a_0$](http://latex.artofproblemsolving.com/d/a/4/da42fa5fa45147429584e2685f8dc2e22247c497.png)
Let have roots
. Define the sum:
Newton's sums tell us that,
(Define for
.)
We also can write:
etc., where denotes the
-th elementary symmetric sum.
Proof
Let be the roots of a given polynomial
. Then, we have that
Thus,
Multiplying each equation by , respectively,
Sum,
Therefore,
Note: This technically only proves the statements for the cases where . For the cases where
, an argument based on analyzing individual monomials in the expansion can be used (see http://web.stanford.edu/~marykw/classes/CS250_W19/Netwons_Identities.pdf, for example.)
Example
For a more concrete example, consider the polynomial . Let the roots of
be
and
. Find
and
.
Newton's Sums tell us that:
Solving, first for , and then for the other variables, yields,
Which gives us our desired solutions, and
.