(Level VI, part II)Geometricum geometriorum (HP III, Prisoner of "the circle")

Geometry Level 5

The perimeter of a circumference of radius 5 5 units and center O O is divided into n n equal parts. Find, as n n tends to infinitum, the limit of the arithmetic mean of the lengths of all circumferences tangent to the first one, passing through the points of division and a fixed point B B that is 4 4 units to O O


Try Part I

Bonus: Generalise this problem when the radius 5 5 is a a and 4 4 is b b with 0 b < a 0 \leq b < a . The answer has a closed form, submit your answer to 3 decimal places.

Note: sorry, the picture is clearly not very good, but I hope it helps to understand better the problem.


The answer is 21.99114857.

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.

1 solution

I'm going to prove the general case. (In this case, a = 5 a = 5 and b = 4 b = 4 ).

Suppose WLOG, O = ( 0 , 0 ) O = (0,0) , B = ( b , 0 ) B = (b,0) and P j = ( a cos ( ( j 1 ) 2 π n ) , a sin ( ( j 1 ) 2 π n ) ) P_j = (a \cos \left((j - 1) \frac{2\pi}{n} \right), a \sin \left((j - 1) \frac{2\pi}{n} \right)) for j = 1 , 2 , . . . , n . j = 1, 2, ... , n.

The centers O j O_j of the tangent circumferences will be in the line O P j OP_j , i.e, O j = ( a λ j cos ( ( j 1 ) 2 π n ) , a λ j sin ( ( j 1 ) 2 π n ) ) O_j = (a \lambda_j \cos\left((j - 1) \frac{2\pi}{n} \right), a \lambda_j \sin \left((j - 1) \frac{2\pi}{n} \right)) Because of d ( O j , B ) = d ( O j , P j ) d(O_j, B) = d(O_j, P_j) we have ( a λ j cos ( ( j 1 ) 2 π n ) b ) 2 + a 2 λ j 2 sin 2 ( ( j 1 ) 2 π n ) = a 2 ( 1 λ j ) 2 ( cos 2 ( ( j 1 ) 2 π n ) + sin 2 ( ( j 1 ) 2 π n ) ) \left(a \lambda_j \cos \left((j - 1) \frac{2\pi}{n} \right) - b \right)^2 + a^2 \lambda_j^2 \sin^2 \left((j - 1) \frac{2\pi}{n} \right) = a^2(1 - \lambda_j)^2 \left(\cos^2 \left((j - 1) \frac{2\pi}{n} \right) + \sin^2 \left((j - 1) \frac{2\pi}{n} \right)\right) ,i.e, b 2 + a 2 ( λ j ) 2 2 a b λ j cos ( ( j 1 ) 2 π n ) = a 2 2 λ j a 2 + a 2 λ j 2 b^2 + a^2(\lambda_j)^2 - 2ab\lambda_j \cos \left((j - 1) \frac{2\pi}{n} \right) = a^2 - 2\lambda_j a^2 + a^2 \lambda_j^2 \Rightarrow λ j = a 2 b 2 2 a ( a b cos ( ( j 1 ) 2 π n ) ) \lambda_j = \frac{a^2 - b^2}{2a\left(a - b \cos \left((j - 1) \frac{2\pi}{n} \right)\right)} The radii of the circumferencies are r j = d ( O j , p j ) = a ( 1 λ j ) r_j = d(O_j, p_j) = a(1 - \lambda_j) and then, as n n tends to infinitum, the limit of the arithmetic mean of the lengths of all circumferences tangent to the first one, passing through the points of division and a fixed point B is: M = lim n 1 n j = 1 2 π a ( 1 λ j ) = lim n 2 π n j = 1 [ a a 2 b 2 2 ( a b cos ( ( j 1 ) 2 π n ) ) ] = M = \displaystyle \lim_{n \to \infty} \frac{1}{n} \sum_{ j = 1}^{\infty} 2\pi a(1 - \lambda_j) = \lim_{n \to \infty} \frac{2\pi}{n} \sum_{ j = 1}^{\infty} \left[ a - \frac{a^2 - b^2}{2\left(a - b \cos \left((j - 1) \frac{2\pi}{n} \right)\right)}\right] = = 2 π a lim n π n j = 1 [ a 2 b 2 2 ( a b cos ( ( j 1 ) 2 π n ) ) ] = 2 π a 1 2 0 2 π a 2 b 2 a b cos ( t ) d t = \displaystyle = 2\pi a - \lim_{n \to \infty} \frac{\pi}{n} \sum_{ j = 1}^{\infty} \left[ \frac{a^2 - b^2}{2\left(a - b \cos \left((j - 1) \frac{2\pi}{n} \right)\right)}\right] = 2\pi a - \frac{1}{2} \int_0^{2\pi} \frac{a^2 - b^2}{a - b\cos(t)} dt = = 2 π a π a 2 b 2 = 2\pi a - \pi \sqrt{a^2 - b^2} Substituing a = 5 , b = 4 a=5 , b = 4 we get, M = 7 π M = 7 \pi

Note.- For doing the integral, you can make the change t = arctan ( x / 2 ) t = \arctan(x/2) . If someone wishes it, I can do it...

Used the same procedure to solve. Very nice problem!

Samrat Mukhopadhyay - 4 years, 7 months ago

Log in to reply

Thank you, the proof is not difficult to understand but I think it's a very,very hard problem...

Guillermo Templado - 4 years, 7 months ago

Log in to reply

Yes, the problem is hard if not approached with the right angle of attack. But otherwise, the problem is beautiful.

Samrat Mukhopadhyay - 4 years, 7 months ago

0 pending reports

×

Problem Loading...

Note Loading...

Set Loading...