Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

Let $A=\begin{pmatrix} \frac{1}{2} & -\frac{\sqrt{3}}{2} \\ \frac{\sqrt{3}}{2} & \frac{1}{2} \end{pmatrix}$ and $P_0=(0,0)$. For positive arbitrary real number $a_i(i=1,2,\cdots ,6)$,

let $\begin{pmatrix} x_i\\ y_i \end{pmatrix}=A^i\begin{pmatrix} a_i\\ 0 \end{pmatrix}$ and define $\overrightarrow{P_{i-1}P_i}=(x_i,y_i)$, provide $P_6=P_0$.

and hexagon $H$ is consist of points $P_o,P_1,P_2,\cdots,P_6$ in regular sequence.

Problem 1.

Prove that $a_1-a_4=a_5-a_2=a_3-a_6$.

Problem 2.

If value provide with $\sum_{i=1}^{6}a_i=6,\ a_1-a_4=1$, find the maximum area of $H$.

Problem 3.

If value provide with $\sum_{i=1}^{6}a_i=6$, find the maximum area of $H$.

I totally can't solve these problems...

Added. I've figured out the shape of hexagon $H$, but nothing else...

enter image description here

Length of each segments are arbitrary, but angle of each segments are 120 degree and $P_0=P_6$.

Added2. Finally, I figured out problem 1. enter image description here

If you have a clear eye site, you can find the regular triangle.

So $a_1+a_2+a_3=a_3+a_4+a_5=a_5+a_6+a_1$.

We know $a_1+a_2+a_3=a_3+a_4+a_5$,

so $a_1+a_2=a_4+a_5$ and $a_1-a_4=a_5-a_2$.

and we know also $a_3+a_4+a_5=a_5+a_6+a_1$

so $a_3+a_4=a_6+a_1$ and $a_3-a_6=a_1-a_4$.

Finally we get $a_1-a_4=a_5-a_2=a_3-a_6$.

Quod Erat Demonstrandum

share|cite|improve this question
Would you happen to remember what a rotation matrix looks like? Your sines and cosines? – J. M. Sep 17 '11 at 13:48
You know your sines and cosines, right? Would you happen to recognize $A$'s entries? – J. M. Sep 17 '11 at 13:51
The question seems unclear to me. If $a_1,\ldots,a_6$ are arbitrary, you cannot prove $a_1-a_4=a_2-a_5=a_3-a_6$. – Martin Sleziak Sep 17 '11 at 13:56
@Martin stressed the fact that there seems to be no hypothesis at all on the numbers $a_i$. Say I choose $a_i=i$ for every $i\ne6$ and $a_6=23$, then these numbers DO NOT satisfy the property of your Problem 1 although my choice seems to be fully legitimate. In a word: please READ Martin's comment. – Did Sep 18 '11 at 17:20
Hmmm, I see... By the way, you do not say what you know, what you tried, and so on. – Did Sep 19 '11 at 5:37
up vote 2 down vote accepted

An easy solution to the first question uses the affixes $z_k$ of the points $P_k$, since $z_0=0$ and $z_{k+1}-z_k=a_ku^{k+1}$ with $u=\mathrm e^{\mathrm i\pi/3}$. The condition that $P_6=P_0$ reads as $\sum\limits_{k=1}^6a_ku^k=0$. Introduce the imaginary part $v$ of $u$, that is,
$$ v=\sqrt3/2. $$ Then the real and imaginary parts of the sum above are $a_1+\frac12a_2-\frac12a_3-a_4-\frac12a_5+\frac12a_6$ and $v(a_2+a_3-a_5-a_6)$ and these are both zero if and only if:

There exists $b$ such that $a_1=a_4+b$, $a_5=a_2+b$ and $a_3=a_6+b$.

Note that this implies that $a_1+a_2+a_3+a_4+a_5+a_6=2(a_2+a_4+a_6)+3b$. From now on we reduce everything to the variables $(a_2,a_4,a_6)$ and $b$.

To solve the second and third questions, first note that opposite sides of the hexagon are parallel and imagine moving points $P_3$, $P_4$ and $P_5$ in the horizontal direction, substracting some quantity $c$ to their three abscissæ. Algebraically, this amounts to adding $c$ to $a_3$ and $a_6$ and leaving the other $a_k$s and $b$ unchanged. Geometrically, this adds $c$ times the height of the hexagon to the global area.

The height of the hexagon, in other words the distance between the lines $P_5P_6$ and $P_2P_3$, is $v(a_1+a_2)=v(a_4+a_2+b)$ hence $$ A(a_2,a_4,a_6+c,b)=A(a_2,a_4,a_6,b)+cv(a_4+a_2+b), $$ where $A(x,y,z,t)$ is the area of the hexagon when $a_1=y+t$, $a_2=x$, $a_3=z+t$, $a_4=y$, $a_5=x+t$, $a_6=z$ and $b=t$. In particular, $$ A(a_2,a_4,a_6,b)=A(a_2,a_4,0,b)+va_6(a_4+a_2+b). $$ Applying the same argument to the two other directions, one gets $$ A(a_2,a_4,a_6,b)=A(0,0,0,b)+va_6(a_4+a_2+b)+va_4(a_2+b)+va_2b. $$ The hexagon parametrized by $(0,0,0,b)$ is in fact the equilateral triangle with side $b$, hence $A(0,0,0,b)=vb^2/2$. This yields finally

$$ A(a_2,a_4,a_6,b)=v(a_2a_4+a_4a_6+a_6a_2+b(a_2+a_4+a_6)+b^2/2). $$

The second question asks for the maximal value of $A$ when $b=1$ and $a_2+a_4+a_6=\frac32$. Lagrange multipliers method indicates that this happens when $a_2=a_4=a_6$, hence the maximal value is $$ A(1/2,1/2,1/2,1)=v\cdot11/4=11\sqrt3/8. $$ The third question asks for the maximal value when $a_2+a_4+a_6=3s$ is such that $2s+b=2$. For each value of $s$, $a_2a_4+a_4a_6+a_6a_2$ is maximal when $a_2=a_4=a_6=s$ hence one looks for the maximal value of $3s^2+3bs+b^2/2$. Since $b=2-2s$, this is $2+s(2-s)\le 3$ which reaches its maximal value when $s=1$ and $b=0$. Finally, the maximal value corresponds to the regular hexagon with side $1$ and is $$ A(1,1,1,0)=3v=3\sqrt3/2. $$

Edit Using your picture, one sees that $A$ is the area of the equilateral triangle with side $a_2+a_4+a_6+2b$ minus the areas of the three equilateral triangles with side $a_1=a_4+b$, $a_3=a_6+b$ and $a_5=a_2+b$. The area of an equilateral triangle with side $\ell$ being $\sqrt3\ell^2/4$, one gets directly $$ A(a_2,a_4,a_6,b)=\left((a_2+a_4+a_6+2b)^2-(a_4+b)^2-(a_2+b)^2-(a_6+b)^2\right)\sqrt3/4. $$ Expanding the squares, one falls back on the formulas above.

share|cite|improve this answer
@kso, Although I've figured out the shape of H, but I totally can't solve these problems is to be compared with There is another solution which use Schwarz's inequality... :-) – Did Oct 20 '11 at 9:13

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.