4

The problem itself.

Recently, a friend of mine has sent me this image, and asked me if I could figure out the area of the white region in this circle; initially, I thought it was simply a infinite summation of a n-gon area, but I found myself stuck, and as such, I wanted some help.

The puzzle is: We have a infinite set of circumscribed polygons, each one in a increasingly larger circle; a infinite amount of them form this larger circle. What is the area of all of the polygons?

My attemp boiled down to, first, describing the problem:

$$(\ \sum^{\infty}_{?}W_a +\sum^{\infty}_{?}C_a\ )=\pi r_1^2$$ This is just a statement that the infinite sum of the white shades and the circles give off the area of a circle, which is just $\pi r^2$; my next step was delineating this to a more specific sum. I decided that the white shades should be the series of a circumscribed n-gon:

$$ \sum^{\infty}_{3}\frac{1}{2}nr_n^2 sin(\frac{2 \pi}{n})$$

So, to find the actual area would just be: $$\sum^{\infty}_{3}\frac{1}{2}nr_n^2 sin(\frac{2 \pi}{n})=\pi r_1^2- ( \sum^{\infty}_{n} \pi r_n^2)$$ However, I realized that the actual radii would be changing in as well, in the form of a infinite series: $$\sum^{\infty}_0 r_n+y$$ After that, I couldn't figure it out anymore. Does anyone know how to deal with something like this?

1 Answers1

2

Notice that two consecutive circles are the inscribed and circumscribed circles of a regular n-gon. You can show that their radii are related by $$r_n \cos \frac{\pi}{n} = r_{n-1}.$$

Hence $$r_n = r_2 \prod_{k=3}^n \frac{1}{\cos \frac{\pi}{k}},$$ where $r_2$ is the radius of the inscribed circle of the triangle. To see that $\lim_{n\to\infty} r_n$ exists, consider $$\ln r_n = \ln r_2 - \sum_{k=3}^n \ln \cos\frac{\pi}{k}.$$ Using this answer one can see that $\ln r_n$ indeed converges, and so does $r_n$.

Edit: Actually, this question is closely related to the problem at hand.

Guus B
  • 648
  • (1) do you mean $\pi / n$? (2) It's not even obvious to me that $r_n$ converges. Sure the ratio $\to 1$, but that does not necessarily mean that $r_\infty$ (i.e. the infinite product of the ratios) is finite, right? Do you have any insight into that? – antkam Oct 30 '19 at 17:17
  • (1) Yes, I do. (2) I will edit the answer with some extra insights, thanks. – Guus B Oct 31 '19 at 08:23
  • Thanks a lot for the help my man. I tried using another method before you replied, which was effectively the same thing I did previously, but using the formula for the apothem of a circumscribed polygon; using a numerical method, the answer converged to $\frac{\pi}{10}$; I haven't checked the convergence of your answer, but I believe it's a nice result. – Artur Renato B.B Nov 08 '19 at 00:36