I found this in a practice GRE problem. I thought I would have a crack at it after being spoiled by the answer

At how many points in the xy-plane do the graphs of $y = x^{12}$ and $y = 2^x$ intersect?

So I thought about doing something what most people would have done, solving for the intersection, $x^{12} = 2^x$, it became utterly hopeless.

Then I thought about using the Intermediate Value Theorem, that is

$f=x^{12} - 2^x = 0$

I suspect for $x<0$, $x^{12} > 2^x$, so $f>0$

For $x=0$, $f < 0$. So by IVT, there is a root somewhere between $(-\infty,0)$

For $x>0$, $x^{12} > 2^x$, so $f>0$. So by IVT, there is another root at $(0,\infty)$

So counting, I should get 2 roots, another therefore 2 points. But the actual answer was 3. So I am inclined to believe I overlooked something very important

Note: The GRE forbids technology assistance.

  • $\begingroup$ Set them equal and take the log of both sides. Then you probably can pick out a few by inspection. This is supposed to be a precalc problem and they don't know much. If this is on the GRE, I wouldn't be surprised if the people who cooked up the problem made a mistake. $\endgroup$ – toypajme Dec 21 '12 at 19:55
  • $\begingroup$ @toypajme, that was my first attempt. It didn't seem like it could lead anywhere. So I tried something else $\endgroup$ – Hawk Dec 21 '12 at 19:58
  • $\begingroup$ Exactly 3 points. $\endgroup$ – copper.hat Dec 21 '12 at 19:59
  • $\begingroup$ Hint: Look at the plot of the curves on top of each other regarding your first suspicion. $\endgroup$ – Amzoti Dec 21 '12 at 20:02
  • 3
    $\begingroup$ The argument for $x\gt 0$ was not right. In the long run $e^x\gt x^{12}$. But $e^2\lt 2^{12}$. So there is a root between $0$ and $2$, and a root somewhere beyond $2$. $\endgroup$ – André Nicolas Dec 21 '12 at 20:02

$x^{12} = 2^x$ (for $x$ real) is equivalent to: either $x = 2^{x/12}$ or $-x = 2^{x/12}$. Since $2^{x/12}$ is convex, its graph intersects a straight line in $0$, $1$ or $2$ points.

$-x$ is decreasing while $2^{x/12}$ is increasing, and $-x > 1 > 2^{x/12}$ for $x < -1$ while $-x < 2^{x/12}$ for $x > 0$, so there is exactly one real solution of $-x = 2^{x/12}$ and it is in the interval $-1 \le x \le 0$.

$x < 0 < 2^{x/12}$ for $x < 0$, $x < 2^{x/12}$ for sufficiently large $x$, while $x > 2^{x/12}$ at $x=2$, so there are two real solutions of $x = 2^{x/12}$, one with $0 < x < 2$ and one in $2 < x < \infty$.

  • $\begingroup$ Oh this is interesting take I like $\endgroup$ – Hawk Dec 21 '12 at 20:34

I suspect for $x<0$, $x^{12} > 2^x$, so $f>0$

Obviously false for negative $x$ close to 0.

For $x>0$, $x^{12} > 2^x$

Obviously false for large positive $x$.

But really, all you have to do is draw a graph. Go away and draw one, and then come back and tell us about it :-)

  • 1
    $\begingroup$ You can't bring graphing devices into a GRE. We are suppose to do this without the assistance of technology I believe $\endgroup$ – Hawk Dec 21 '12 at 20:08
  • $\begingroup$ Even if I were to draw a sketch, I could only hit two obvious points. The third one is too subtle for me considering it lands on $x \approx 74...$ $\endgroup$ – Hawk Dec 21 '12 at 20:10
  • 2
    $\begingroup$ @sizz, by "draw" he means sketch. Also, the third point is known because $2^x$ grows quicker than $x^{12}$. $\endgroup$ – picakhu Dec 21 '12 at 20:11
  • 1
    $\begingroup$ @sizz, you simply ignored my first two points. $\endgroup$ – TonyK Dec 21 '12 at 20:28

I thought of it this way; near the origin, $2^x$ is relatively flat, while $x^{12}$ points up sharply. This gives two points of intersection. But, eventually, exponentials outgrow any polynomial, so there must be another point of intersection where $2^x$ outgrows $x^{12}$.

Actually, I didn't think of the third root at the time either and put down 2 as my answer on the practice exam. But hey, only 12 percent of test takers get it right!

  • $\begingroup$ This is how I did it. There seem to be a few problems like this on each GRE, where if you try do things rigorously it's impossible, but intuitively it's obvious once you see the point of the problem. $\endgroup$ – Potato Sep 25 '14 at 1:04

First thing to observe is that both the functions are continuous and the graphs are "smooth" curves.

There is one and only one point of intersection in $(- \infty, 0)$ as $x^{12}$ is increasing & $2^x$ is decreasing as $x \rightarrow- \infty$ and $2^0 > 0^{12}$.

Now observe that when $x \rightarrow \infty, $ both the functions increase and $2^x $ dominates $x^{12}$ eventually. Since $ 2^{12} > 2^2 $, it must be the case that the graphs intersect at some point in $(2, \infty)$. This is the only point of intersection in that interval.

Also, $2^0 > 0^{12}$ and $2^{12} > 2^2 $ tells us that there is a point of intersection in $(0,2)$ and there is exactly one such intersection.

So altogether there are exactly 3 points of intersection.


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.