Without using calculus I was told to find the maximum of $$f(x)=6x-x^2.$$

The "symmetry" approach notices that $f(x)=x(6-x)$, and that replacing $x\leftrightarrow6-x$ does not change $f(x)$ which means it doesn't change the maximum. Then the solution says that the only value of $x$ that is unchanged by $x\leftrightarrow 6-x $ is $x=3$. So that's the location of the maximum.

This is all the solution says. I am under the impression that if $f(x)$ is unchanged then no number should be affected unchanged. How did they figure out that $x=3$ is the only thing unchanged?

  • $\begingroup$ Essentially you used the average of the $x$ intercepts of the parabola to find $x=3$.You can then check and see that $f(3+h)=f(3-h)$ for all $h$ $\endgroup$ – imranfat May 8 '18 at 21:53
  • $\begingroup$ Check this: a parabola of this form has two zeros, i.e. $x = 0$ and $x = 6$. Since it is symmetric, the maximal value corresponds to $x = 3$ - the middle point of the zeros. $\endgroup$ – Gibbs May 8 '18 at 21:55
  • $\begingroup$ The only solution of $x = 6 - x$ is? $\endgroup$ – NickD May 8 '18 at 21:58
  • $\begingroup$ This can be made to work once you add/prove the necessary assumptions about concave parabolas. But you have to be careful in general, try for example to apply the same argument to the very similarly looking $\,f(x)=x^2-6x\,$, or $\,f(x)=x^2(6-x)^2\,$, and see why/where it fails. $\endgroup$ – dxiv May 8 '18 at 22:09

Since $f(x)=x(6-x)$, it is $f(6-x)=(6-x)x=f(x)$. Now let $x_0$ be a maximum of $f$. We have $f(x_0)=f(6-x_0)$, therefore since this maximum is attained only one time (e.g. the graph is a parabola), it must be $6-x_0=x_0$.

  • $\begingroup$ This is a good answer that actually addresses some of the symmetry. (+1) $\endgroup$ – T. Bongers May 8 '18 at 21:59
  • $\begingroup$ Thank you, this is a clear answer. Do you know if the same argument works a function that has higher orders? For example $f(x)=50x^2-x^4$ $\endgroup$ – matryoshka May 8 '18 at 22:11
  • 1
    $\begingroup$ @Grace I'm not sure about this. With a first look, i think this one has 2 maxima. It is $f(x)=x^2(50-x^2)=f(50-x^2)$, so I believe the maxima are located at the solutions of $x=50-x^2$. I'm not sure about this though. $\endgroup$ – JustDroppedIn May 8 '18 at 22:14
  • $\begingroup$ Let consider $f(x)=50x^2-x^4=625-625+50x^2-x^4=625-(25-x^2)^2$ $\endgroup$ – gimusi May 9 '18 at 5:41


Note that


thus $f(x)$ is the sum of

  • a constant term $9$
  • a function $-(x-3)^2$ symmetric with respect to $x=3$

and then $f(x)$ itself is symmetric with respect to $x=3$.

  • 1
    $\begingroup$ Squared, not cubed. $\endgroup$ – NickD May 8 '18 at 21:57
  • $\begingroup$ opssss...typo! I fix. Thanks $\endgroup$ – gimusi May 8 '18 at 21:57
  • $\begingroup$ I've added some trivial details on symmetry for the less perceptive downvoters. $\endgroup$ – gimusi May 8 '18 at 22:05

HINT: Because it is a polynomial of second degree, its extremum lies on the symmetry line of the interval with ends at its zeroes.


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.