Second Point Of Intersection Of Circumcircles

Geometry Level pending

Let A B C \triangle ABC be a triangle with B C = 5 , C A = 6 , A B = 7. BC= 5, CA= 6, AB= 7. Let D , E D, E be the midpoints of A B , A C AB, AC respectively. The circumcircles of B C D \triangle BCD and A D E \triangle ADE meet for a second time at X X ( X A X \neq A ). Given that A X 2 = a b AX^2= \dfrac{a}{b} for some coprime positive integers a , b , a, b, find the last three digits of a + b . a+b.

Details and assumptions

  • This problem is inspired by a recent problem that appeared in the Chinese Girls Mathematical Olympiad 2014.


The answer is 883.

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

OK, this is a horrible problem with a very inelegant solution. Apologies. The original problem asked to show that the second intersection point of the circumcircles of B C D , A D E \triangle BCD, \triangle ADE and B C E , A D E \triangle BCE, \triangle ADE are equidistant from A A (which should be obvious from the following solution). The reader might refer to this helpful article on barycentric coordinates.


Set the normalized barycentric coordinates A = ( 1 , 0 , 0 ) , B = ( 0 , 1 , 0 ) , C = ( 0 , 0 , 1 ) . A= (1, 0, 0), B = (0, 1, 0), C = (0, 0, 1). Then, the coordinates of D , E D, E are ( 1 2 , 1 2 , 0 ) \left( \dfrac{1}{2}, \dfrac{1}{2}, 0 \right) and ( 1 2 , 0 , 1 2 ) \left( \dfrac{1}{2}, 0, \dfrac{1}{2} \right) respectively.

Suppose the equation of circle ( A D E ) (ADE) is a 2 y z + b 2 z x + c 2 x y = u x + v y + w z a^2yz+b^2zx+c^2xy = ux+vy+wz (normalized coordinates). Plugging the coordinates of A , A, we see that u = 0. u=0. Plugging the coordinates of D , D, we see that c 2 4 = u + v 2 v = c 2 2 . \dfrac{c^2}{4} = \dfrac{u+v}{2} \implies v = \dfrac{c^2}{2}. Similarly, plugging the coordinates of E E shows that w = b 2 2 . w= \dfrac{b^2}{2}. To sum it up, the equation of the circumcircle of A D E \triangle ADE is a 2 y z + b 2 z x + c 2 x y = c 2 y + b 2 z 2 . a^2yz+b^2zx+c^2xy= \dfrac{c^2y + b^2z}{2}.

Let the equation of ( B C D ) (BCD) be a 2 y z + b 2 z x + c 2 x y = p x + q y + r z . a^2yz + b^2zx + c^2xy = px+qy+rz. Plugging the coordinates for B B and C C show that q = r = 0. q=r=0. Plugging the coordinates of D , D, we see that c 2 4 = p 2 p = c 2 2 . \dfrac{c^2}{4} = \dfrac{p}{2} \implies p = \dfrac{c^2}{2}. To sum it up, the equation of the circumcircle of B C D \triangle BCD is a 2 y z + b 2 z x + c 2 x y = c 2 x 2 . a^2yz+b^2zx + c^2xy = \dfrac{c^2x}{2}.

Let ( x , y , z ) (x, y, z) be the point of intersection of ( A D E ) (ADE) and ( B C D ) (BCD) apart from D D (where x + y + z = 1 x+y+z=1 ).

The displacement vector A X \vec{AX} is ( x 1 , y , z ) , (x-1, y, z), so A X 2 = a 2 y z b 2 z ( x 1 ) c 2 ( x 1 ) y = ( a 2 y z + b 2 z x + c 2 x y ) + ( b 2 z + c 2 y ) = c 2 y + b 2 z 2 + b 2 z + c 2 y = b 2 z + c 2 y 2 = a 2 y z + b 2 z x + c 2 x y = c 2 x 2 . \begin{aligned} |AX|^2 & = - a^2 yz - b^2 z (x-1) - c^2 (x-1)y \\ & = - (a^2yz + b^2zx + c^2xy) + (b^2z + c^2y) \\ &= - \dfrac{c^2y + b^2z}{2} + b^2z + c^2y \\ &= \dfrac{b^2z + c^2y}{2} \\ & = a^2yz + b^2zx + c^2xy \\ & = \dfrac{c^2x}{2}. \end{aligned}

Now, we know that x , y , z x, y, z simultaneously satisfy the equations a 2 y z + b 2 z x + c 2 x y = c 2 x 2 a 2 y z + b 2 z x + c 2 x y = c 2 y + b 2 z 2 x + y + z = 1. \begin{aligned}a^2yz+b^2zx + c^2xy & = \dfrac{c^2x}{2} \\ a^2yz+b^2zx+c^2xy &= \dfrac{c^2y + b^2z}{2} \\ x + y + z & = 1. \end{aligned} Equating a 2 y z + b 2 z x + c 2 x y a^2yz+b^2zx+c^2xy from the first two equations and plugging z = 1 x y , z = 1-x-y, we obtain c 2 x = c 2 y + b 2 ( 1 x y ) y = ( b 2 + c 2 ) x b 2 c 2 b 2 . c^2x = c^2y + b^2 (1 - x - y) \implies y = \dfrac{(b^2+c^2)x-b^2}{c^2-b^2}.
This implies z = 1 ( b 2 + c 2 ) x b 2 c 2 b 2 x = c 2 b 2 c 2 ( 2 x 1 ) . z = 1 - \dfrac{(b^2+c^2)x-b^2}{c^2-b^2} - x = \dfrac{c^2}{b^2-c^2}(2x-1).

Plugging this in the second equation, we obtain 2 a 2 ( b 2 + c 2 ) x b 2 c 2 b 2 c 2 b 2 c 2 ( 2 x 1 ) + 2 b 2 ( b 2 + c 2 ) x b 2 c 2 b 2 x + 2 c 2 ( b 2 + c 2 ) x c 2 c 2 b 2 x = 0. 2a^2 \dfrac{(b^2+c^2)x-b^2}{c^2-b^2} \dfrac{c^2}{b^2-c^2}(2x-1) + 2b^2 \dfrac{(b^2+c^2)x-b^2}{c^2-b^2} x + 2c^2\dfrac{(b^2+c^2)x-c^2}{c^2-b^2}x = 0.

We interpret this as a quadratic in x . x. We know that x = 1 2 x = \dfrac{1}{2} is one solution (since D = ( 1 2 , 1 2 , 0 ) D = \left( \dfrac{1}{2}, \dfrac{1}{2}, 0 \right) is a point of intersction of the circles in question). To find the other solution, it suffices to find the ratio of the constant term and coefficient of x 2 . x^2. Collecting terms, we find out that this ratio is [ x 0 ] [ x 2 ] = a 2 b 2 2 a 2 ( b 2 + c 2 ) ( b 2 c 2 ) 2 . \dfrac{[x^0]}{[x^2]} = \dfrac{a^2b^2}{2a^2(b^2+c^2) - (b^2-c^2)^2}. Hence, the other solution for x x is 2 a 2 b 2 2 a 2 ( b 2 + c 2 ) ( b 2 c 2 ) 2 . \dfrac{2a^2b^2}{2a^2(b^2+c^2) - (b^2-c^2)^2}. Now, A X 2 = c 2 x 2 = a 2 b 2 c 2 2 a 2 ( b 2 + c 2 ) ( b 2 c 2 ) 2 |AX|^2 = \dfrac{c^2x}{2} = \dfrac{a^2b^2c^2}{2a^2(b^2+c^2) - (b^2-c^2)^2} .

Plugging the values shows us that the last three digits of a + b a+b are 883 . \boxed{883}.

Note that the expression of A X 2 AX^2 is symmetric in b , c , b, c, which solves the original problem.

Yes u are right

Vagish Jha - 6 years, 9 months ago

Very nice solution. I'm adding one in Cartesian coordinates. Embed A B C \triangle ABC into the first quadrant of a Cartesian coordinate system with origin in B B and the directed line B C BC as x x -axis. Apply cosine rule to triangle A B C \triangle ABC to get 25 + 49 2 5 7 cos A B C = 36 , 25+49-2\cdot 5 \cdot 7 \cdot \cos \angle ABC = 36, hence cos A B C = 19 35 \cos \angle ABC = \frac{19}{35} and consequently sin 2 A B C = 1 1 9 2 3 5 2 = 864 1225 . \sin^{2} \angle ABC = 1 - \frac{19^{2}}{35^{2}} = \frac{864}{1225}. Use Apollonius' theorem to the median C D CD , obtaining C D 2 = 2 5 2 + 2 6 2 7 2 4 = 73 4 CD^{2} = \frac{2\cdot 5^{2} + 2 \cdot 6^{2} - 7^{2}}{4} = \frac{73}{4} If R R is the circumradius of B C D \triangle BCD , by sine rule we have 4 R 2 = C D 2 sin 2 A B C \hbox , w h e n c e R 2 = 125195 13824 4R^{2} = \frac{CD^{2}}{\sin^{2} \angle ABC} \hbox{, whence } R^{2} = \frac{125195}{13824} Call F F the circumcenter of B C D \triangle BCD and J J its x x -projection. Then F J 2 = R 2 25 4 = 3025 13824 , FJ^{2} = R^{2} - \frac{25}{4} = \frac{3025}{13824}, hence the coordinates of F F are ( 5 / 2 , 55 6 / 288 ) (5/2,55\sqrt{6}/288) . Call I I the x x -projection of A A , to get B I = A B cos A B C = 7 19 35 = 19 5 . |BI| = |AB|\cos \angle ABC = 7\cdot \frac{19}{35} = \frac{19}{5}. Now if L L is the x x -projection of E E , then B L = 1 2 ( 5 + 19 2 ) = 22 5 . |BL| = \frac{1}{2}\cdot \left(5 +\frac{19}{2}\right) = \frac{22}{5}. Now, if G G is the circumcenter of A D E \triangle ADE and N N its x x -projection, abscissa of G G is given by B N = B K + B L 2 = 19 + 44 20 = 63 20 , |BN| = \frac{|BK|+|BL|}{2} = \frac{19 + 44}{20} = \frac{63}{20}, where K K is the x x -projection of D D . Let's call now r r the circumradius of A D E \triangle ADE . We get 4 r 2 = 3 2 sin 2 A D E = 3 2 sin 2 A B C \hbox , h e n c e r 2 = 1225 384 4r^{2} = \frac{3^{2}}{\sin^{2} \angle ADE} = \frac{3^{2}}{\sin^{2} \angle ABC} \hbox{, hence } r^{2} = \frac{1225}{384} If O O is the mid-point of D E DE , we get G O 2 = r 2 D E 2 4 = 1225 384 25 16 = 625 384 , GO^{2} = r^{2} -\frac{DE^{2}}{4} = \frac{1225}{384}- \frac{25}{16} = \frac{625}{384}, hence G O = 25 6 48 |GO| = \frac{25\sqrt{6}}{48} . Moreover, we have D K 2 = ( 7 2 ) 2 sin 2 A B C = 216 25 , DK^{2} = {\left(\frac{7}{2}\right)}^{2}\sin^{2} \angle ABC = \frac{216}{25}, hence D K = 6 6 5 DK=\frac{6\sqrt{6}}{5} . The coordinates of point G G are given by G : ( 63 20 , 6 6 5 + 25 6 48 ) = ( 63 20 , 413 6 240 ) . G:\left(\frac{63}{20},\frac{6\sqrt{6}}{5}+\frac{25\sqrt{6}}{48}\right) = \left(\frac{63}{20},\frac{413\sqrt{6}}{240}\right). The system of equations of the two circumcircles is { ( x 5 2 ) 2 + ( y 55 6 288 ) 2 = 89425 13824 ( x 63 20 ) 2 + ( y 413 6 240 ) 2 = 1225 384 \left\{ \begin{array}{l} \left(x-\frac{5}{2}\right)^{2} + {\left(y - \frac{55\sqrt{6}}{288}\right)}^{2} = \frac{89425}{13824}\\ \left(x-\frac{63}{20}\right)^{2} + \left(y - \frac{413\sqrt{6}}{240}\right)^{2} = \frac{1225}{384} \end{array} \right. The solution different from D D is the point X : ( 15905 4081 , 4020 6 4081 ) X : \left(\frac{15905}{4081}, \frac{4020\sqrt{6}}{4081}\right) and its squared distance to A A is given by d 2 = ( 15905 4081 19 5 ) 2 + ( 4320 6 4081 12 6 5 ) 2 = = 6300 583 = a b . d^{2} = \left(\frac{15905}{4081} - \frac{19}{5}\right)^{2} +\left(\frac{4320\sqrt{6}}{4081} - \frac{12\sqrt{6}}{5}\right)^{2} =\\ = \frac{6300}{583} = \frac{a}{b}. Now the last three digits of a + b = 6883 a+b=6883 are 883 883 .

pierantonio legovini - 6 years, 7 months ago

I hope. This answer is wrong. Make second approach triangle BEC similar to triangle ADX so AX.AX=931/100 so answer will be 031

Vagish Jha - 6 years, 9 months ago

Log in to reply

I find no particular reason why B E C \triangle BEC should be similar to A D X . \triangle ADX. They have only one common angle ( A C D = A E D = E C B \angle ACD = \angle AED = \angle ECB ). The following image disproves your claim.

Image link: http://s23.postimg.org/4axu1538b/Untitled.png Image link: http://s23.postimg.org/4axu1538b/Untitled.png

Sreejato Bhattacharya - 6 years, 9 months ago

0 pending reports

×

Problem Loading...

Note Loading...

Set Loading...