2
$\begingroup$

Question: Suppose $f(x)$ is a bounded polynomial, in other words, there is an $M$ such that $|f(x)|\le M$ for all $x\in R$. Prove that $f$ must be a constant.

I think the question assumes $M\in R$. I can't think of any polynomial that is bounded by some real number unless the polynomial degree is zero, so $f$ should be constant. But, I am not sure about how to prove this.

My attempt: Prove the statement by contrapositive.

If $f$ is not constant, then $f$ should be unbounded. Let $f(x)=b_dx^d+b_{d-1}x^{d-1}+...+b_1x+b_0$ for $d>0$. Since polynomial is infinitely differentiable (explanation), we can differentiate $f$ for $d$ times. Then, we get $b_d\in R$. That is, $f$ is increasing or decreasing continuously. Therefore, if $f$ is bounded, $f$ should be constant.

Is it okay? If not, could you give some hint?

Thank you in advance!

$\endgroup$
1
  • 1
    $\begingroup$ Contraposition is really the most natural approach. Suppose $f$ is non-constant, hence $n := \text{deg}(f) \geq 1$. Then $f(x) \sim a_nx^n$ and we know $|a_n x^n| \to + \infty$ for $n \geq 1$. $\endgroup$ Commented Apr 8, 2018 at 2:36

2 Answers 2

4
$\begingroup$

Assume that $f(x)=a_{n}x^{n}+\cdots+a_{1}x+a_{0}$ for $a_{n}\ne 0$ and $n\geq 1$, then \begin{align*} \lim_{x\rightarrow\infty}\dfrac{f(x)}{a_{n}x^{n}}&=\lim_{x\rightarrow\infty}\left(1+\cdots+\dfrac{a_{1}}{a_{n}x^{n-1}}+\dfrac{a_{0}}{a_{n}x^{n}}\right)\\ &=1. \end{align*} On the other hand, we have \begin{align*} \left|\dfrac{f(x)}{a_{n}x^{n}}\right|\leq\dfrac{M}{|a_{n}||x|^{n}}\rightarrow 0, \end{align*} so \begin{align*} \lim_{x\rightarrow\infty}\dfrac{f(x)}{a_{n}x^{n}}=0, \end{align*} a contradiction, so we must have $a_{n}=0$. Proceed in the similar fashion we get all $a_{n-1}=\cdots=a_{1}=0$, and hence $f$ is constant.

$\endgroup$
5
  • $\begingroup$ I don't understand your last line. I agree that it is a contradiction, but why does it leads to conclude that $a_n=0$?? $\endgroup$
    – shk910
    Commented Apr 8, 2018 at 2:55
  • $\begingroup$ Because it has been assumed that the leading coefficient of the polynomial is not the constant one, while it is of $x^{n}$. $\endgroup$
    – user284331
    Commented Apr 8, 2018 at 2:56
  • $\begingroup$ Then, why does $a_n=0$ imply $f$ is constant? We don't know yet if $a_{n-1},a_{n-2},...$ are 0. $\endgroup$
    – shk910
    Commented Apr 8, 2018 at 2:59
  • $\begingroup$ Good question, then you simply proceed in a similar way to eliminate all the remaining leading coefficients. $\endgroup$
    – user284331
    Commented Apr 8, 2018 at 3:00
  • 1
    $\begingroup$ We don't need to proceed downward from $n.$ If the largest $n$ such that $a_n\ne 0$ is not $n=0$,(that is, if $f$ is not constant), the first part of this A shows that assuming $f$ is bounded yields a contradiction, and we are done. $\endgroup$ Commented Apr 8, 2018 at 6:16
2
$\begingroup$

If you differentiate $d$ times, you get $d! \cdot b_d$. This is a real number, but remember, it's the only the first derivative that determines increasing or decreasing, not the $d$th derivative! So no, unfortunately, your proof doesn't work.

Here's what I would try: consider the polynomial divided by $x^d$. If $d > 0$ and the polynomial were bounded, this function should go to $0$ as $x \to \infty$ (you need to prove this). However...

$\endgroup$

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .