Find the largest possible value of

$$\sin(a_1)\cos(a_2) + \sin(a_2)\cos(a_3) + \cdots + \sin(a_{2014})\cos(a_1)$$

Since the range of the $\sin$ and $\cos$ function is between $1$ and $-1$, shouldn't the answer be $2014$?

  • 1
    $\begingroup$ You cannot have $\sin (a_1)=\cos (a_1)=1$ so the maximum is not 2014. $\endgroup$ – Kabo Murphy Aug 8 '18 at 8:24
  • 1
    $\begingroup$ Fairly obviously it won't exceed $2014$ but it does not necessarily achieve it either. $\endgroup$ – badjohn Aug 8 '18 at 8:24

Let $a_{2015}=a_1.$

Thus, by AM-GM $$\sum_{k=1}^{2014}\sin{a_k}\cos{a_{k+1}} \leq\sum_{k=1}^{2014}|\sin{a_k}||\cos{a_{k+1}}| \le$$ $$\leq\sum_{k=1}^{2014}\frac{\sin^2a_k+\cos^2a_{k+1}}{2}=\sum_{k=1}^{2014}\frac{\sin^2a_k+\cos^2a_k}{2}=\frac{2014}{2}=1007.$$

The equality occurs for $a_i=45^{\circ},$ which says that $1007$ is a maximal value.

  • $\begingroup$ You mention AM-GM, which makes me expect to see a big sum turn into a big product at some point. But it never happens. Something clever is clearly going on in the third inequality -- but what? Could you expand this just a bit more? $\endgroup$ – Daniel Wagner Aug 8 '18 at 13:19
  • $\begingroup$ @Daniel Wagner I used $ab\leq\frac{a^2+b^2}{2}.$ $\endgroup$ – Michael Rozenberg Aug 8 '18 at 13:32
  • $\begingroup$ Ah! Very simple -- I kept trying to figure out where the calculation of a 2014th root went. Thanks for the clarification. $\endgroup$ – Daniel Wagner Aug 8 '18 at 13:35
  • $\begingroup$ You are welcome! $\endgroup$ – Michael Rozenberg Aug 8 '18 at 13:37


$$ f_n(a) = \sum_{k=1}^n \sin a_k \cos a_{k+1} $$

with $a_{n+1} = a_1$

the stationary points are located at the solutions for

$$ \frac{\partial }{\partial a_k}f_n(a) = -\sin a_{k-1}\sin a_k + \cos a_k \cos a_{k+1} = 0 $$

and then

$$ \tan a_n\tan a_{n-1}\cdots\tan a_{2} = \cot a_1 $$


$$ \tan a_n\tan a_{n-1}\cdots\tan a_{2}\tan a_1 = 1 $$


$$ \prod_k\sin a_k = \prod_k\cos a_k $$

which is obtained for $a_k = \frac{\pi}{4}$ when

$$ f_n(a) = \frac n2 $$

  • $\begingroup$ Nice answer. The calculus is easier than it looks at first. Is it obvious that $a_k=\frac{\pi}{4}$ is the absolute maximum? The $\tan a_n\tan a_{n-1}\cdots\tan a_{2}\tan a_1 = 1$ may have many solutions but I guess they are not all absolute maximum... $\endgroup$ – Taladris Aug 8 '18 at 13:49
  • $\begingroup$ @Taladris There are many relative extrema. For instance for $n = 3$ we have: $\left\{-\frac{3}{2},\frac{3}{2},-\sqrt{2},\sqrt{2}\right\}$ but with some additional considerations like symmetry... $\endgroup$ – Cesareo Aug 8 '18 at 14:44

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.