$\displaystyle f_n(x)=\sum_{i=0}^{n}a_ix^i$ is a polynomial having $n$ real roots where real numbers $a_0,a_1, a_2,a_3...a_n$ are unimodular, i.e $|a_i|=1$, and $n>0$. Then the number of possible values of n is: $$ \begin{matrix} a)\ 1 \qquad & a)\ 2 \qquad & a)\ 3 \qquad & a)\ 4 \end{matrix} $$

Since it's an MCQ, I just put $n=1,2,3$ and check if they do have all roots real or not.

For $n=1$,
$f(x)=x-1$ is a possibility.

For $n=2$,
$f(x)=x^2+x-1$ is a possibility as
$\Delta = b^2-4ac = 5 > 0$

For $n=3$,
$f(x)=x^3+x^2-x-1=(x-1)(x+1)^2$ is a possibility.

and I have no idea how to prove/disprove all the roots real in polynomials with $n>3$.


We show $n \le 3$ (assuming $n\ge 2$ as the problem is trivial for $n=1$)

We note that if $x_k$ are the roots (at least two), $x_k \ne 0$ since the product is $\pm 1$ and then $\Sigma{x_k^2} = 3$ since it's positive of the form $1-2(\pm 1)$. (note that all Newton expressions in the roots, ie sum, symmetric sum of order 2,3.., product are $\pm 1$ also)

Applying the same reasoning with the equation with roots $\frac{1}{x_k}$ which is obtained by switching the coefficients which sum to power $n$ so has same property of all coefficients $\pm 1$, we get $\Sigma{\frac{1}{x_k^2}}=3$ too, but then we get:

$9=(\Sigma{x_k^2})(\Sigma{\frac{1}{x_k^2}}) \ge n^2$ so we are done

| cite | improve this answer | |
  • $\begingroup$ Sorry, I forgot about the $a_0$. Just edited it in. $\endgroup$ – user42819 May 11 '19 at 14:45
  • $\begingroup$ assumed it as otherwise, the problem is different and makes less sense $\endgroup$ – Conrad May 11 '19 at 14:46

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.