|
|
|
|
|
|
|
|
CTK Exchange
Bui Quang Tuan
Member since Jun-23-07
|
Dec-12-07, 09:23 PM (EST) |
|
"Geometrical Limit Around A Circle"
|
Dear All My Friends, Given one circle (O) and one point P outside this circle. Choose one direction around this circle say d, for example d is clockwise. From P we can draw two tangent lines of (O), say x and y. Two lines x, y bound one angle say A(P) <= 180 and (O) is inside of it. We can always draw third line z tangent with (O) and intersects x, y at X, Y respectively such that: PX = PY and (O) is incircle of triangle PXY. We choose X, Y such that PXY is following direction d. Now we choose point X as P1 and do with P1 exactly as with P then we can get angle A(P1). Continuing this steps we can get P2, A(P2), P3, A(P3),... Pn, A(Pn),... as much as we can. My questions: 1. What is the limit of A(Pn) sequence? 2. What we can say about the curve on which are all P, P1, P2,... Pn,... ? Thank you and best regards, Bui Quang Tuan
|
|
Alert | IP |
Printer-friendly page |
Reply |
Reply With Quote | Top |
|
|
mpdlc
Member since Mar-12-07
|
Dec-13-07, 10:02 PM (EST) |
|
1. "RE: Geometrical Limit Around A Circle"
In response to message #0
|
I do not know if I understood correctly the enunciate , but it looks too easy since by construction the resulted circumscribed triangle is always isosceles it means 2*A(Pn)= pi- A(Pn-1); 2*A(Pn+1)= pi- A(Pn); eliminating pi we get the difference equation A(Pn+1)=(A(Pn)+A(Pn-1))/2 which solution render P(k)= (P(0)+P(1)+P(2))/3 and since this sum amount 180 degree P(n)= 60 degrees. The points P, P1, P2 will be located in a kind of spiral being its limit the circumcicle containing the equilateral triangle circumscribed to the given circle. The trajectory it will depend obviously of the initial selection of P mpdlc |
|
Alert | IP |
Printer-friendly page |
Reply |
Reply With Quote | Top |
|
|
mr_homm
Member since May-22-05
|
Dec-13-07, 11:04 PM (EST) |
|
2. "RE: Geometrical Limit Around A Circle"
In response to message #0
|
Hello Bui Quang Tuan, Each triangle is isoceles by construction, with the current angle A(P_n) between the equal sides. Therefore, 2*A(P_n+1) + A(P_n) = pi. This can be rearranged into the symmetrical form (pi/3 - A(P_n)) / 2 = -(pi/3 - A(P_n+1)). This shows clearly that the angles will alternate between being greater than pi/3 and less than pi/3. Also, at each step, the difference from pi/3 will decrease by a factor of 2. Therefore, the sequence of angles goes exponentially to a stable limit pi/3. Thus the triangles approach being equilateral in the limit. If the initial angle is expressed as pi/3 + A, then the angles are pi/3 + A*(-1/2)^n, where the initial angle is given number n=0. The angle between the ray OP_0 and OP_1 is then pi - 1/2*(pi/3 + A + pi/3 - A/2) = 2pi/3 - A/4. For each succeeding pair of rays, the angle will be 2pi/3 + (A/2)*(-1/2)^n, where the initial angle is given number n=1. This shows that the points P_0, P_1, ... rotate around the center of the circle in unequal steps that converge to an angle of 2pi/3. (This is as it'should be, because the limiting triangle is isoceles, and so the rotation should be 2pi/3 in the limit.) The total angle of rotation between OP_0 and OP_n is then n*2pi/3 - (A/2)*(1/2)^n. The distances OP_n will then be R*csc(1/2*(pi/3 + A*(-1/2)^n)), where R is the radius of the circle. These distances will alternately be larger or smaller than 2R, and so they will lie on two uneven spirals, one spiral of decreasing radial distances, and another spiral of increasing radial distances. I think that because angles of 2pi can be discarded, the points will lie on 3 nice smooth spirals going inward and 3 nice smooth spirals going outward. This is because taking n = 1, 4, 7, etc makes angle n*2pi/3 equivalent to 2pi/3, and similarly form n = 2, 5, 8, etc and n = 3, 6, 9, etc. This is as much as I can see right now. I hope it is of interest. --Stuart Anderson |
|
Alert | IP |
Printer-friendly page |
Reply |
Reply With Quote | Top |
|
|
|
mr_homm
Member since May-22-05
|
Dec-16-07, 07:22 PM (EST) |
|
4. "RE: Geometrical Limit Around A Circle"
In response to message #3
|
Dear Bui Quang Tuan, The location of the final equilateral triangle can be found from algebra. Consider that the points P_n, P_n+1 and P_n+2 form the vertices of one triangle. Since the limit triangle is equilateral, it is only necessary to find the location of one of these points. This can be done by looking only at points P_0, P_3, P_6, ... P_3n, ..., which should form a converging sequence of points leading to the a vertex of the limit triangle. In my previous post, I said that the angle between OP_0 and OP_n was n*2pi/3 - (A/2)*(1/2)^n. That was an error, because I summed the series incorrectly. The correct sum of 2pi/3 + (A/2)*(-1/2)^n from 1 to n is n*2pi/3 - (A/2)*(1/3)*(1-(-1/2)^n) = n*2pi/3 - (A/6)*(1-(-1/2)^n)). Therefore, the angle between OP_0 and OP_3n is 3n*2pi/3 - (A/6)*(1-(-1/2)^3n) = n2pi - A/6*(1-(-1/2)^n). Since multiples of 2pi can be discarded, and since the limit of (1/2)^n = 0, the final angle is therefore -A/6. Thus a vertex of the limit triangle lies at an angle A/6 from the line OP_0 in the opposite direction from the way we choose to go around the circle from each P to the next P. Recall that A is defined as the difference between the vertex angle at P_0 and pi/3. This answers at least part of the question. I have still not thought much about the spirals. Regards, Stuart Anderson |
|
Alert | IP |
Printer-friendly page |
Reply |
Reply With Quote | Top |
|
|
|
|
|
|
mr_homm
Member since May-22-05
|
Dec-17-07, 00:41 AM (EST) |
|
6. "RE: Geometrical Limit Around A Circle"
In response to message #5
|
Hello Bui Quang Tuan, Yes, if the limit triangle could be constructed with ruler and compass, you could solve the ancient angle trisection problem, which we know cannot happen. This is an interesting application of the theorem against angle trisection. --Stuart Anderson |
|
Alert | IP |
Printer-friendly page |
Reply |
Reply With Quote | Top |
|
|
|
You may be curious to have a look at the old CTK Exchange archive. Please do not post there.
Copyright © 1996-2018 Alexander Bogomolny
|
|