0
$\begingroup$

I have the equation $x^2 + x \cos(x) = 1 + \sin(x)$ and I need to prove that it has exactly two solutions.
What I used to do before when I had to prove an equation has one solution was:
used the intermediate value theorem to show there exists a solution and then used contradiction with Rolle's theorem to show that there's a unique one. But here I'm not so sure what to do.

Thanks for the help!

$\endgroup$
  • $\begingroup$ Please, if you are ok, you can accept the answer and set it as solved. Thanks! $\endgroup$ – user Feb 3 '18 at 23:53
5
$\begingroup$

Let $f(x) = x^2 + x\cos(x) - \sin(x) - 1$. Then, $f'(x) = 2x - x \sin(x) = x(2 - \sin(x))$.

Now, $2 - \sin (x)$ is always positive, so the sign of $f'(x)$ is just the sign of $x$. In particular, the only root of $f'(x)$ is $0$.

Between any two roots of $f(x)$, there must be a point of zero derivative, by mean value theorem.

If there are more than two roots, then $f'(x)$ would have more than one root, contradiction.

Carefully apply intermediate value theorem and you shall be able to prove that two roots actually exist.

$\endgroup$
  • $\begingroup$ what would you say f(a) and f(b) are in this situation though? $\endgroup$ – Nicole Jan 6 '18 at 12:08
1
$\begingroup$

You can differentiate, study the variations of the function and use two times the intermediate value theorem. ( by finding four wisely chosen points ) The representation will help you

$\endgroup$
  • $\begingroup$ youre suggesting i first differentiate? i have to firstly show there exists a solution no? $\endgroup$ – Nicole Jan 6 '18 at 11:59
1
$\begingroup$

Consider $f(x)=x^2+xcosx-1-sinx$. We want to show that it has exactly two roots. First let's calculate its derivative: $$f^{'}(x)=2x-xsinx=x(2-sinx)$$ here we find out that the function is strictly increasing when $x>0$ and strictly decreasing when $x<0$ and $f(0)=-1$. Then two strict branches come up from $(0,-1)$ and intersect with x-axis at exactly two points.

$\endgroup$
1
$\begingroup$

A possible strategy is to consider the function

$$f(x)=x^2 + x \cos(x)- \sin(x)-1$$

and show that it has only two intersection point with $x$ axis.

Note that

$$x\to\pm\infty \quad f(x)\to+\infty$$

and

$$f'(x)=2x-x\sin x=x(2-\sin x)$$

thus

$$f'(x)=0\iff x=0 \quad f(0)=-1$$

$$f'(x)>0\iff x>0$$

$$f'(x)<0\iff x<0$$

Now use IVT to show that f has exactly two intersection point with x axis.

$\endgroup$
  • $\begingroup$ yeah thats my problem, how do i show it has exactly two intersection points? since we havent done investigations of functions yet.. $\endgroup$ – Nicole Jan 6 '18 at 11:57
  • $\begingroup$ @Nicole you need to investigate the function by derivative $\endgroup$ – user Jan 6 '18 at 11:58
  • $\begingroup$ i see what youre saying but then the function is equal to -1 whereas if you consider the function to be what you said then the solutions would need to make f(x) = 0 , if im not mistaken? $\endgroup$ – Nicole Jan 6 '18 at 12:04
  • $\begingroup$ @Nicole f has a negative minimum at x=0 and is positive for some x>a and x<b thus, since it is strictly increasing for x>o and strictly decreasing for x<0 it easy to show that f has necessarly only two intersection with x axis. $\endgroup$ – user Jan 6 '18 at 12:09
1
$\begingroup$

Let $f(x)=x^2+x\cos{x}-\sin{x}-1.$

Thus, for all $|x|\geq\sqrt3$ by C-S we obtain $$f(x)=x^2-1-(\sin{x}-x\cos{x})\geq$$ $$\geq x^2-1-\sqrt{(\sin^2x+\cos^2x)(1^2+(-x)^2)}=x^2-1-\sqrt{1+x^2}=$$ $$=\frac{x^2(x^2-3)}{x^2-1+\sqrt{1+x^2}}\geq0.$$ The equality does not occur, which says that the equation $f(x)=0$ has no roots for $|x|\geq\sqrt3$.

In another hand, by C-S again for $|x|\leq\sqrt3$ we obtain $$f''(x)=2-(\sin{x}+x\cos{x})\geq2-\sqrt{(\sin^2x+\cos^2x)(1+x^2)}=2-\sqrt{1+x^2}\geq0,$$ which says that $f$ is a convex function on $[-\sqrt3,\sqrt3]$, which gives that $f$ has there at most two roots.

The rest is smooth:

$f(-\pi)>0$, $f(0)<0$ and $f(\pi)>0$, which says that $f(x)=0$ has two roots exactly.

$\endgroup$
  • $\begingroup$ what is c-s? sorry for the ignorance $\endgroup$ – Nicole Jan 6 '18 at 12:55
  • $\begingroup$ I guess he means to say Cauchy Schwartz $\endgroup$ – Rohan Shinde Jan 6 '18 at 14:58

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.