0
$\begingroup$

If $r$ is the radius of the circle, $\Delta{ABC}$ a triangle inscribed in it and $x,y,z$ are the angles at $A,B,C$ then how to find the triangle that has the maximum area.

In my book it's mentioned that the area$=\frac12ab \sin z$ where $a,b$ are the sides opposite to $A,B$, also $a=2r\sin z, b=2r\sin z .$ Can anyone tell me how to find expression for $a,b$ and the area mentioned above?

$\endgroup$
5

4 Answers 4

2
$\begingroup$

If $x$, $y$ and $z$ are given then $$S_{\Delta ABC}=2r^2\sin{x}\sin{y}\sin{z}$$ if it's not so,

we'll get much interesting problem.

Let $a$, $b$ and $c$ be sides-lengths of the triangle.

For $a=b=c$ we have $r=\frac{a}{\sqrt3}$ and $$S_{\Delta ABC}=\frac{a^2\sqrt3}{4}=\frac{3\sqrt3}{4}r^2.$$

We'll prove that $$S_{\Delta ABC}\leq\frac{3\sqrt3}{4}r^2.$$ Indeed, since $$S_{\Delta ABC}=\sqrt{p(p-a)(p-b)(p-c)}=\frac{1}{4}\sqrt{\sum\limits_{cyc}(2a^2b^2-a^4)}$$ and $$S_{\Delta ABC}=\frac{abc}{4r},$$ we need to prove that $$\frac{1}{4}\sqrt{\sum\limits_{cyc}(2a^2b^2-a^4)}\leq\frac{3\sqrt3}{4}\frac{a^2b^2c^2}{16\left(\frac{1}{4}\sqrt{\sum\limits_{cyc}(2a^2b^2-a^4)}\right)^2}$$ or $$\left(\sqrt{\sum\limits_{cyc}(2a^2b^2-a^4)}\right)^3\leq3\sqrt3a^2b^2c^2$$ or $$\sum\limits_{cyc}(2a^2b^2-a^4)\leq3\sqrt[3]{a^4b^4c^4}$$ or $$\sum_{cyc}(a^4+\sqrt[3]{a^4b^4c^4}-2a^2b^2)\geq0.$$ Now, let $a^2=x^3$, $b^2=y^3$ and $c^2=z^3$.

Hence, we need to prove that $$\sum_{cyc}(x^6+x^2y^2z^2-2x^3y^3)\geq0,$$ which follows from Schur: $\sum_{cyc}(x^6-x^4y^2-x^4z^2+x^2y^2z^2)\geq0.$

Indeed, by Schur we obtain: $$\sum_{cyc}(x^6+x^2y^2x^2-2x^3y^3)\geq\sum_{cyc}(x^4y^2+x^4z^2-2x^3y^3)=\sum_{cyc}x^2y^2(x-y)^2\geq0$$ and we are done!

$\endgroup$
1
$\begingroup$

I would think of it without trigonometry and calculus: If the triangle has maximum area, then we can't make the area larger by moving a single corner.

Pick one side of the triangle as base (rotate the circle so that the base is horizontal for clearer view). Then the largest area you can get is by putting the point opposite the base as far away from from the base as possible. This happens to be halfway along the longest arc between the end points of the base. In other words, the opposite point must be exactly in the middle of the two end points of the base for the area to be maximal.

Now, if the triangle is not equilateral, there is at least one point which is not in the middle of the two others, so we can increase the area of the triangle. An equilateral triangle is the only one which cannot be increased this way, so it must be maximal.

To find the area of an equilateral triangle inscribed in a circle, the height of the triangle is $\frac32r$. This is because in an equilateral triangle, heights, medians and side bisectors coincide, and we know that

  1. The center of the circumcircle of a triangle is where the side bisetors intersect
  2. The medians partition one another in segmenst of ratio $2:1$
  3. The $2$ part of the medians is the radius in the circle, so its length is $r$

Now you have an equilateral triangle where you know the height is $\frac32r$, so you can work out that the side length is $\sqrt3r$, which gives the triangle an area of $\frac{3\sqrt3}2r^2$.

$\endgroup$
5
  • $\begingroup$ you only proved that an equilateral triangle is a local maximum with respect to the way you change the triangle. You did not prove that this is the maximum triangle that can be inscribed in a circle. $\endgroup$
    – miracle173
    Commented Aug 3, 2017 at 7:06
  • $\begingroup$ I proved it is the only local maximum. If there is only one local maximum, and the space is compact (which it is if allow degenerate triangles), then that is the global maximum. $\endgroup$
    – Arthur
    Commented Aug 3, 2017 at 7:10
  • $\begingroup$ that is true but if you incorporate this in your proof (including a proof that the space is compact) then your proof will not be simple at all. If you omit it your proof will be incomplete and wrong. $\endgroup$
    – miracle173
    Commented Aug 3, 2017 at 7:26
  • $\begingroup$ Here is a proof that adresses the missing part: cut-the-knot.org/triangle/InscribedEquilateral.shtml#proof . It needs some calculation. $\endgroup$
    – miracle173
    Commented Aug 3, 2017 at 14:24
  • $\begingroup$ Here is a more elegant geometric proof of the missing part in your proof. math.stackexchange.com/a/1403815/11206 $\endgroup$
    – miracle173
    Commented Aug 7, 2017 at 22:31
1
$\begingroup$

HINT

First, I suppose only $r$ is given, otherwise it doesn't make much sense.

Since $S_{\Delta ABC}=2r^2\sin{x}\sin{y}\sin{z}$, we'll have to find the maximum of $X = \sin{x}\sin{y}\sin{z}$. Now, let's fix $z$. Then $X=\frac {\cos(x - y) - \cos(x + y)}{2}\sin z$. Since $z$ it's fixed, $x + y$ is fixed and $X$ will reach a maximum when $\cos(x - y)$ reaches a maximum, that is when $x=y$

$\endgroup$
5
  • $\begingroup$ So you proved that an isoceles triangle is a local maximum, but not that the equilateral triangle is the solution $\endgroup$
    – miracle173
    Commented Aug 3, 2017 at 14:29
  • $\begingroup$ @miracle173 Just take, let's say, x as fixed and repeat it $\endgroup$
    – user261263
    Commented Aug 3, 2017 at 14:58
  • $\begingroup$ nevertheless there is the same problem as with the solution of Arthur $\endgroup$
    – miracle173
    Commented Aug 3, 2017 at 15:37
  • $\begingroup$ @miracle173 What problem? $\endgroup$
    – user261263
    Commented Aug 3, 2017 at 15:40
  • $\begingroup$ check math.stackexchange.com/a/1403815/11206 $\endgroup$
    – miracle173
    Commented Aug 7, 2017 at 22:30
0
$\begingroup$

The maximum area will be for an equilateral triangle. Draw a radial line from each vertex to the circle's centre, to make three isosceles triangles. The central angle $θ$ is 120 degrees or 2π/3 radians. The area of each triangle is $\frac {r^2\sinθ}{2}$ and the whole area will be 3 times that.

Edit after a comment from @miracle173.

Why is it an equilateral triangle? Consider such an inscribed triangle with one vertex centrally at the top and a horizontal base line. The area of a triangle is height * base / 2. Now, if you move the top vertex either way, the height decreases and the base remains the same, so the area decreases.

The side lengths $a$ and $b$ can also be found from simple trigonometry

$a=b=2r\sin(θ/2)$

$\endgroup$
6
  • $\begingroup$ Why is it an equilateral triangle? $\endgroup$
    – miracle173
    Commented Aug 3, 2017 at 6:42
  • $\begingroup$ @miracle173 I have edited the answer thank you. $\endgroup$ Commented Aug 3, 2017 at 6:50
  • $\begingroup$ that's not sufficient. This is true for an isosceles triangle, too. $\endgroup$
    – miracle173
    Commented Aug 3, 2017 at 7:01
  • $\begingroup$ @miracle173 not if you put a different side as the base. Moving the top vertex upward will then increase the area. I am not going to prove what is "self evident". $\endgroup$ Commented Aug 3, 2017 at 7:04
  • $\begingroup$ see my comment to Arthur's answer. Nothing is self evident, the proof does not work. $\endgroup$
    – miracle173
    Commented Aug 3, 2017 at 7:08

Not the answer you're looking for? Browse other questions tagged .