1
$\begingroup$

I was plotting some equations and I got with the curious relation

If we build the triange enter image description here Such that it follows the following relation: $$AD=a$$ $$DB=b$$ $$AC=ak$$ $$CB=bk$$

Then when we vary $k$ from the smallest possible value to the greatest possible one according to triangle inequality, the point C plots a circle around the midpoint of the maximum and minimum solutions, with it's diameter being the segment between them.

In a more formal way

If $C$ is the center of the hypotetic circle, $$\frac{AD}{DB}=\frac{AE}{EB}$$ and according to triangle inequality $$\frac{AD}{DB}=\frac{FA}{FB}$$ Prove that $$CD=CE$$ That is, prove that the point E in fact lies in the hypotetic circle. enter image description here This came out of curiosity when i plotted the point C as k varies, I was expecting a more weird shape but I got with a circle!.I have empirically verified it in geogebra and such, but i have not been able to prove it.The closest i have gotten is that $CA=\frac{AD^2}{DB-AD}$, imposing that $DB>AD$. If they are equal, a division by cero occurs, but if $DB<AD$, the circles moves to the other side.

$\endgroup$
3
$\begingroup$

Here's a brute-force coordinate approach ...

Edit. My point names aren't consistent with OP's (which vary between his diagrams anyway :). In case of edits to the question, I'll explain the roles of my points here.

We take point $D$ on segment $AB$, with $a := |AD|$ and $b := |BD|$. A variable point $P$ is defined by $|AP|=k |AD|$ and $|BP|=k|BD|$. Note that, if $a=b$ the locus of $P$ is the perpendicular bisector of $AB$ (which happens to be the limiting case of circles), so we take $a > b > 0$.

Coordinatizing with points $A( -a, 0 )$, $B( b, 0 )$, $D( 0, 0 )$, $P(x,y)$, we have $$\begin{align} |AP| &= k |AD|: \qquad ( x + a )^2 + y^2 = a^2 k^2 \\ |BP| &= k |BD|: \qquad ( x - b )^2 + y^2 = b^2 k^2 \end{align}$$

Equating $k^2$ obtained from both equations gives $$\frac{(x+a)^2 + y^2}{a^2}=k^2=\frac{(x-b)^2+y^2}{b^2}$$ so that $$b^2 (x+a)^2 + b^2 y^2 = a^2 ( x-b)^2 + a^2 y^2$$

$$( a^2 - b^2 )x^2 - 2 x a b ( a + b ) + ( a^2 - b^2 ) y^2 = 0$$

$$( a - b ) x^2 - 2 x a b + ( a - b ) y^2 = 0$$

Since the (non-zero!) coefficients of $x^2$ and $y^2$ match (and there's no $xy$ term), we definitely have a circle; without a constant term, we see that the circle passes through the origin (point $D$). For more information, we put the equation into standard form:

$$\left( x - \frac{a b}{a-b} \right)^2 + y^2 = \left(\frac{ab}{a-b}\right)^2$$

So, the circle has center $C = \left(\frac{ab}{a-b},0\right)$ and radius $|CD| = \frac{ab}{a-b}$, and we can verify your conjecture:

$$|CA| = \frac{ab}{a-b}+a = \frac{a^2}{a-b} \qquad |CB| = \frac{ab}{a-b}-b=\frac{b^2}{a-b}$$

$\endgroup$
  • $\begingroup$ Thank you! I was very curious about it. Also, this "theorem" helps me a lot to simplify my equations for the app I was developing and plotting around. $\endgroup$ – chubakueno May 31 '13 at 1:07

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.