4
$\begingroup$

I was trying to solve a problem, it says like this: If $p (x)$ is a third degree polynomial and $a, b, c, d$ are real numbers with d> a since $p (a) = p (b) = p (c) = p '(b) = p' '(c) = 0 $ and $p (d)> 0$, will there exist some real number $x$ such that $(x - a) ⋅ p (x) +1 = 0$?

I thinked some time and I arrived that this problem is incorrect, because, if $p$ is supposed to be a third degree polynomial, and because $b$ is a root and a critical number, there's no exist such $p$.

I have this proof, I don't know if it is sufficient 'rigorous' or it is correct, thanks in advantage.

Let $p (a) = p (b) = p (c) = p '(b) = 0$ and $p$ a third degree polynomial. Because $b$ is a critical number and $p$ a third degree polynomial, can only exist other $k$ such that $p^{\prime}(k)=0$. If $b$ is a relative maximum, on the right of $b$, $p$ is decreasing, and on the left of $b$, $p$ is increasing. No matter that $b < a,c$ or $b> a,c$ or $a<b<c$, etc. There must exist other two $q$ and $k$ such that $p'(q)=p'(k)=0$ because $p$ have other two roots, and it would need to change of increasing/decreasing two times.. But, this leads us to a contradiction, because $p$ can only have a maximum of $2$ critical points. It is similar with $b$ a relative minimum. If $b$ is not a relative maximum or a relative minimum, $p$ has to be strictly increasing or decreasing over an interval $(b-\alpha, b+\alpha)$. No matter that $b < a,c$ or $b> a,c$ or $a<b<c$, etc. There must exist other two $q$ and $k$ such that $p'(q)=p'(k)=0$ because $p$ have other two roots, and it would need to change of increasing/decreasing two times. But, this leads us to a contradiction, because $p$ can only have a maximum of $2$ critical points. So we showed that if $p$ is a third degree polynomial with three roots, and one of these roots is a critical number, $p$ does not exist.

$\endgroup$
1
  • $\begingroup$ Do you know that if $p(b)=p'(b)=0$ for a polynomial $p(x)$ then $b$ is a double root - ie $p(x)=(x-b)^2q(x)$ for some polynomial $q(x)$? To prove this consider $p(x)=(x-b)s(x)$ so that $p'(x)=(x-b)s'(x)+s(x)$ and $p'(b)=s(b)=0$. $\endgroup$ – Mark Bennet May 8 at 12:20
3
$\begingroup$

$$p(x)=Ax^3+Bx^2+Cx+D=A(x-\alpha)(x-\beta)(x-\gamma)$$ where $\alpha,\beta,\gamma$ are the roots. Since we are given that $p(a)=p(b)=p(c)=0$ we know that there are the roots and since we only care about the sign we can get rid of $A$: $$p(x)=(x-a)(x-b)(x-c)$$ $$p(x)=x^3-(a+b+c)x^2+(ab+ac+bc)x-abc$$


we are also told that $p'(b)=0$: $$p'(x)=3x^2-2(a+b+c)x+(ab+ac+bc)$$ $$\Rightarrow 3b^2-2(a+b+c)b+(ab+ac+bc)=0\tag{1}$$ and that $p''(c)=0$: $$p''(x)=6x-2(a+b+c)$$ $$\Rightarrow 6c-2(a+b+c)=0$$

rearranging this we get that: $$c=\frac{a+b}{2}\tag{2}$$ which gives us that $c$ is also the middle root. Also if we sub and rearrange $(1)$ we get: $$3b^2-3ab-3b^2+2ab+\frac{a^2+b^2}{2}=0$$ $$\Rightarrow a^2-2ab+b^2=0$$ or: $$(a-b)^2=0$$ which gives us that: $$a=b=c\tag{3}$$


so we can rewrite our polynomial: $$p(x)=A(x-a)^3$$ now we are told that $d>a$ and $p(d)>0$ so: $$A(d-a)^3>0$$ but since $d-a>0$ that means that $A>0$.


Now finally lets look at the question: $$\exists x\in\mathbb{R}\,\text{s.t.}\,\,\,\,(x-a)p(x)+1=0$$ well we can now rewrite this as: $$A(x-a)^4+1=0$$ $$(x-a)^4=-\frac1A$$ now since $x,a$ are real and $A$ is positive, we can confirm such $x$ does not exist as it would not be real

$\endgroup$
1
  • $\begingroup$ This is quite a large answer so I apologize in advance if I have missed any mistakes in my checking $\endgroup$ – Henry Lee May 8 at 12:20
2
$\begingroup$

Here is a shorter answer than the first one posted. Note first that a cubic polynomial has three roots which need not be distinct. We are given the roots $a,b,c$ so that $p(x)=A(x-a)(x-b)(x-c)$

Now the condition $p'(b)$ is equivalent to saying that there is a double root at $b$ so we must have $b=a$ or $b=c$ so we have $p(x)=A(x-a)^2(x-c)$ or $p(x)=A(x-a)(x-c)^2$

In the second case with $p''(c)=0$ we have a triple root at $c=a$ whence $p(x)=A(x-a)^3$. The condition on $d$ ensures $A\gt 0$ and then $(x-a)p(x)+1=A(x-a)^4+1$ is the sum of two non-negative terms one of which is always positive.

In the first case $p'(x)=2A(x-a)(x-c)+A(x-a)^2$ and $p''(x)=2A(x-c)+2A(x-a)+2A(x-a)$ and $p''(c)=0$ implies $4A(c-a)=0$ from which we deduce once again that $a=c$, there is a triple root and we are back to the previous case.

$\endgroup$

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.