Inscribing Regular Polygons in Circles

Geometry Level 5

Let n 3 n \ge 3 . A n n -sided regular polygon is inscribed in a circle of radius 1 1 . Then, all the diagonals of the polygon are drawn as well. Thus every vertex of the polygon is connected by a line segment to every other vertex.

Let L n L_n be the total length of line segments that are drawn by this process (including the perimeter). To the nearest thousandth, find the value of lim n L n n 2 \displaystyle\lim_{n\to\infty}\frac{L_n}{n^2} .


The answer is 0.637.

This section requires Javascript.
You are seeing this because something didn't load right. We suggest you, (a) try refreshing the page, (b) enabling javascript if it is disabled on your browser and, finally, (c) loading the non-javascript version of this page . We're sorry about the hassle.

3 solutions

Ariel Gershon
Apr 4, 2015

Let O O be the centre of the circle. If we draw a line segment from O O to every vertex, this would divide the circle into n n equal parts. Therefore, for each pair of vertices A , B A,B there must be a integer k k with 1 k n 2 1 \le k \le \dfrac{n}{2} such that A O B = 2 π k n \angle AOB = \dfrac{2\pi k}{n} .

Suppose first that n n is odd. Then for each 1 k n 1 2 1 \le k \le \dfrac{n-1}{2} , there are exactly n n line segments A B AB such that A O B = 2 π k n \angle AOB = \dfrac{2\pi k}{n} . (Why that? Because there is exactly one line segment that starts from each vertex and goes counterclockwise by the specified angle.)

Now it's fairly simple to prove that if A O B = 2 π k n \angle AOB = \dfrac{2\pi k}{n} , then A B = 2 sin ( π k n ) AB = 2\sin{\left(\dfrac{\pi k}{n}\right)} . Since there are exactly n n line segments of each length, we get the formula: L n = k = 1 n 1 2 2 n sin ( π k n ) L_n = \sum_{k=1}^{\frac{n-1}{2}} 2n\sin{\left(\dfrac{\pi k}{n}\right)} L n n 2 = 2 n k = 1 n 1 2 sin ( π k n ) \frac{L_n}{n^2} = \frac{2}{n}\sum_{k=1}^{\frac{n-1}{2}} \sin{\left(\dfrac{\pi k}{n}\right)} Now let's rewrite this as a Riemann sum. Let f ( x ) = sin π x f(x) = \sin{\pi x} , and x k = k n 1 , x k = k n x_k = \frac{k}{n-1}, x^{*}_k = \frac{k}{n} for 0 k n 1 2 0\le k\le \dfrac{n-1}{2} . Then k 1 n 1 < k n < k n 1 \dfrac{k-1}{n-1} < \dfrac{k}{n} <\dfrac{k}{n-1} so x k 1 < x k < x k x_{k-1} < x^{*}_k < x_{k} . Furthermore, we have x k x k 1 = 1 n 1 x_k - x_{k-1} = \dfrac{1}{n-1} .

Therefore, our expression can be rewritten as follows: L n n 2 = 2 ( n 1 ) n k = 1 n 1 2 f ( x k ) ( x k x k 1 ) \dfrac{L_n}{n^2} = \dfrac{2(n-1)}{n} \sum_{k=1}^{\frac{n-1}{2}} f(x^{*}_k)(x_k - x_{k-1}) Therefore we do indeed have a Riemann sum. Now x 0 = 0 x_0 = 0 and x ( n 1 ) / 2 = 1 2 x_{(n-1)/2} = \frac{1}{2} , so as n n approaches infinity, we get: lim n L n n 2 = lim n 2 ( n 1 ) n 0 1 / 2 sin ( π x ) d x = 2 1 π = 2 π \lim_{n\to\infty} \frac{L_n}{n^2} = \lim_{n\to\infty}\frac{2(n-1)}{n} \int_{0}^{1/2} \sin{(\pi x)} dx = 2 * \frac{1}{\pi} = \boxed{\dfrac{2}{\pi}} Note that if n n is even, we get a similar formula: L n n 2 = 1 n ( 1 + k = 1 n / 2 2 sin π k n ) \frac{L_n}{n^2} = \frac{1}{n}\left(-1 + \sum_{k=1}^{n/2} 2\sin{\frac{\pi k}{n}}\right)

Alternately, you can also try to prove that L n = n cot π 2 n L_n = n \cot{\frac{\pi}{2n}} .

Ariel Gershon - 6 years, 2 months ago

Log in to reply

I tried to solve it but I am getting L n = n ( 4 cos 2 ( π 2 n ) cot ( π 2 n ) 3 cot ( π 2 n ) ) L_n=n(4 \cos^2 (\frac {\pi}{2n} )\cot (\frac {\pi}{2n}) - 3 \cot (\frac {\pi}{2n}))

Akshay Bodhare - 6 years, 1 month ago

Log in to reply

How did you get that? I used my formulas from above:

L n = { 2 n k = 1 n 1 2 sin ( π k n ) i f n i s o d d n + 2 n k = 1 n 2 sin ( π k n ) i f n i s e v e n L_n = \left\{ \begin{matrix} 2n \displaystyle\sum_{k=1}^{\frac{n-1}{2}} \sin\left(\dfrac{\pi k}{n}\right) & if& n& is& odd \\ -n + 2n \displaystyle\sum_{k=1}^{\frac{n}{2}} \sin{\left(\dfrac{\pi k}{n}\right)} & if& n& is& even \end{matrix} \right. and then simplified them to one expression: L n = n k = 1 n sin ( π k n ) L_n = n \sum_{k = 1}^{n} \sin \left(\frac{\pi k}{n}\right) Now all that's left is to prove that k = 1 n sin ( π k n ) = cot ( π 2 n ) \displaystyle\sum_{k = 1}^{n} \sin \left(\frac{\pi k}{n}\right) = \cot\left(\dfrac{\pi}{2n}\right) .

Ariel Gershon - 6 years, 1 month ago

Log in to reply

@Ariel Gershon Suppose that we start from vertex a 1 a_1 then we find the distances ( a 1 , 3 + a 1 , 4 + . . . . . . + a 1 , n 1 ) (a_{1,3}+a_{1,4}+...... +a_{1,n-1}) we then multiply the above by n and then divide it by 2 to get the expression for L n L_n . I think that I might have misinterpreted the problem since what I have found is the length of all diagonals rather that the length of each and every line.However, in the end we get the same answer.

Akshay Bodhare - 6 years, 1 month ago

Very Very elegant problem and solution. I solved it in a way completely different to the way you solved it. Though my way isn't to precise and I am unsure of one part of it.

Trevor Arashiro - 6 years, 1 month ago
汶良 林
May 21, 2015

Trevor Arashiro
Apr 20, 2015

This solution is kinda sketchy just saying.

As n approaches infinity, we are looking at a circle inscribed in a circle (it's circleception :3 lol)

Fix a point T T on the circle's circumference. Draw a radius from the circle's center O O and place another point S S on the circle's circumference. Label angle S O T = θ \angle SOT=\theta and line segment S T = x \overline{ST}=x . We are looking at the average length of x x

By law of cosines x 2 = 2 + 2 cos ( θ x^2=2+2\cos(\theta )

x 2 = r 2 + r 2 2 r 2 cos ( θ ) ) = 2 r 2 ( 1 ( 2 cos 2 ( θ 2 ) 1 ) ) = x^{2} = r^{2}+r^2- 2 r^2\cos(\theta)) = 2r^{2}(1 - (2\cos^{2}(\frac{\theta}{2}) - 1)) =

4 r 2 ( 1 cos 2 ( θ 2 ) ) = 4 r 2 sin 2 ( θ 2 ) 4r^{2}(1 - \cos^{2}(\frac{\theta}{2})) = 4r^{2}\sin^{2}(\frac{\theta}{2})

x = 2 r sin ( θ 2 ) for 0 θ π . \Longrightarrow x = 2r\sin(\frac{\theta}{2})~\text{for} 0 \le \theta \le \pi.

Using probability distribution function

1 π 0 π 2 r sin ( θ 2 ) d θ = 4 r π 0 π sin ( u 2 ) d u = 4 r π [ cos ( π 2 ) ( cos ( 0 ) ) ] = 4 r π \displaystyle\dfrac{1}{\pi}\int_{0}^{\pi} 2r\sin(\frac{\theta}{2}) d\theta = \dfrac{4r}{\pi}\int_{0}^{\pi} \sin(\frac{u}{2}) du = \dfrac{4r}{\pi}[-\cos(\frac{\pi}{2}) - (-\cos(0))] = \dfrac{4r}{\pi}

Here r = 1 r=1 so we have the average length equal to 4 π \dfrac{4}{\pi}

The number of diagonals in a figure is n ( n 3 ) 2 \dfrac{n(n-3)}{2} . To include the number of sides this becomes n ( n 1 ) 2 \dfrac{n(n-1)}{2} .

Multiplying the number of diagonals by the average diagonal length (remember that this only works as n approaches infinity).

L n = ( 4 π ) ( n ( n 1 ) 2 ) = 2 n ( n 1 ) π L_n=\left(\dfrac{4}{\pi}\right)\cdot\left(\dfrac{n(n-1)}{2}\right)= \dfrac{2n (n-1)}{\pi}

Dividing through by the number of sides gives us 2 ( n 1 ) n π \dfrac{2 (n-1)}{n\pi}

As lim n n = lim n ( n 1 ) \displaystyle \lim_{n\rightarrow \infty} n=\displaystyle \lim_{n\rightarrow \infty} (n-1)

Thus our result is

lim n L n n 2 = lim n 2 ( n 1 ) n π = 2 π 0.637 \displaystyle \lim_{n\rightarrow \infty}\dfrac{L_n}{n^2}=\displaystyle \lim_{n\rightarrow \infty}\dfrac{2 (n-1)}{n\pi}=\dfrac{2}{\pi}\approx \boxed{0.637}

This is a nice solution! I think it is valid if you explain it properly and rewrite it a bit. The average length is simply the total length divided by the number of line segments, so therefore multiplying the average by the number of line segments gives you the total length. And as n n \to \infty , the average approaches the distribution function with the value of 4 π \dfrac{4}{\pi} , while the number of line segments divided by n 2 n^2 approaches 1 2 \dfrac{1}{2} . In general the limit of a product is the product of the limits so it is okay to calculate the limits separately and then multiply them.

One correction though: in the question it asks for the total value of all line segments drawn, so it's not just the diagonals but the side lengths of the polygon as well. So instead of n ( n 3 ) 2 \dfrac{n(n-3)}{2} , you should have written n ( n 1 ) 2 \dfrac{n(n-1)}{2} . Fortunately this does not change the limit and so you got the same answer; but I thought it's worth mentioning.

Ariel Gershon - 6 years, 1 month ago

Log in to reply

Ohh, well yes the perimeter is negligent, I thought the question was only asking form diagonals, I've edited the question to include the sidelengths.

Btw, thanks for verifying my solution :)

Trevor Arashiro - 6 years, 1 month ago

0 pending reports

×

Problem Loading...

Note Loading...

Set Loading...