Newton's Inequality
Contents
Background
For , we define the symmetric sum to be the coefficient of in the polynomial (see Viete's sums). We define the symmetric average to be .
Statement
For non-negative and ,
,
with equality exactly when all the are equal.
Proof
Lemma. For real , there exist real with the same symmetric averages .
Proof. We consider the derivative of . The roots of are . Without loss of generality, we assume that the increase as increases. Now for any , must have a root between and by Rolle's theorem if , and if , then is a root of times, so it must be a root of times. It follows that must have non-positive, real roots, i.e., for some non-negative reals ,
.
It follows that the symmetric sum for is , so the symmetric average .
Thus to prove Newton's theorem, it is sufficient to prove
for any . Since this is a homogenous inequality, we may normalize it so that . The inequality then becomes
.
Expanding the left side, we see that this is
.
But this is clearly equivalent to
,
which holds by the rearrangement inequality.
Proof: without calculus
We will proceed by induction on .
For , the inequality just reduces to AM-GM inequality. Now suppose that for some positive integer the inequality holds.
Let , , , be non-negative numbers and be the symmetric averages of them. Let be the symmetric averages of , , . Note that .
\begin{align*}
d_{k-1}d_{k+1} =& \left(\frac{n-k+1}{n} {d'}_{k-1} + \frac{k-1}{n} {d'}_{k-2} x_m \right)\left(\frac{n-k-1}{n} {d'}_{k+1} + \frac{k+1}{n} {d'}_k x_m \right) \\ =& \frac{(n-k+1)(n-k-1)}{n^2} {d'}_{k-1}{d'}_{k+1} + \frac{(k-1)(n-k-1)}{n^2} {d'}_{k-2} {d'}_{k+1} x_m\\ & + \frac{(n-k+1)(k+1)}{n^2} {d'}_{k-1}{d'}_k x_m + \frac{(k-1)(k+1)}{n^2} {d'}_{k-2}{d'}_k x_m^2\\ \le & \frac{(n-k+1)(n-k-1)}{n^2} {d'}_k^2 + \frac{(k-1)(n-k-1)}{n^2} {d'}_{k-2} {d'}_{k+1} x_m\\ & + \frac{(n-k+1)(k+1)}{n^2} {d'}_{k-1}{d'}_k x_m + \frac{(k-1)(k+1)}{n^2} {d'}_{k-1}^2 x_m^2\\ \le & \frac{(n-k+1)(n-k-1)}{n^2} {d'}_k^2 + \frac{(k-1)(n-k-1)}{n^2} {d'}_{k-1} {d'}_{k} x_m\\ & + \frac{(n-k+1)(k+1)}{n^2} {d'}_{k-1}{d'}_k x_m + \frac{(k-1)(k+1)}{n^2} {d'}_{k-1}^2 x_m^2\\ =& \frac{(n-k)^2}{n^2} {d'}_k^2 + \frac{2(n-k)k}{n^2} {d'}_k {d'}_{k-1} x_m +\frac{k^2}{n^2} {d'}_{k-1}^2 x_m^2 - \left(\frac{d_k}{n} - \frac{d_{k-1}x_m}{n}\right)^2\\ \le & \left(\frac{n-k}{n} {d'}_k + \frac{k}{n} {d'}_{k-1} x_m \right)^2\\ =& d_k^2
\end{align*}