1

Problem

Find the locus of the centroid of an equilateral triangle inscribed in the ellipse $x^2 / a^2 + y^2 / b^2 = 1$

My attempt

I assumed 3 parametric points on ellipse P, Q and R. And assumed the centroid to be $G(h,k)$ As we know the formula for centroid (taking average of three coordinate assumed). Hence got the relation $$ h= a(cos \theta_{1} + cos \theta_{2} + cos \theta_{3} )/3 $$ $$ k = b(sin \theta_{1} + sin \theta_{2} + sin \theta_{3} )/3 $$ Now I tried using formula for circumcentre and no matter whicher formula I used I used to get the same above equation. I tried writing equation of three sides and manually solving the centroid.

I also got an intuition that in an equilateral triangle all 4 centres are same. So centroid and circumcentre are same. So $ PG = QG = RG $. So I wrote down all the equation and equated them by comparing the LHS equation and RHS equation to get $$ cos \theta_{1} = cos \theta_{2} = cos \theta_{3} $$ $$ sin \theta_{1} = sin \theta_{2} = sin \theta_{3} $$ which is obviously wrong as that would mean all the tree points assumed intersected into one single point.

Where exactly am I wrong? What did I miss that I had to think about which I didn't think. I am looking for some nice elegant human solution to this problem as I have seen a similar question in mathematical stack exchange in which they have directly solved equations and got very big equation for h and k.

Below I attach solution that is provided in textbook but in the textbook solution there is a formula written directly. If anyone could explain how I could prove the formula as no information is provided from where the formula came.Solution 1. Solution 2

Thanks for helping. I have been working on this for 3 days with no improvement. This is my first question on mathematics stack exchange so any comments on how should I state something is appreciated if I stated the wrong way.

Mark
  • 13
  • Can you elaborate on "Now I tried using formula for circumcircle and no matter whicher formula I used I used to get the same above equation."? Which is the above equation? Are you saying that $x^2/a^2 + y^2/b^2 = 1$ is always a circle? – Calvin Lin Feb 02 '24 at 01:33
  • @CalvinLin No by formula I meant circumcentre formula here sorry I wrote circumcircle by mistake and by above equation I am saying the h and k equation that I got. – Mark Feb 02 '24 at 01:36
  • The 3 points on ellipse P, Q and R are the corners of an equilateral triangle. Could you pls show how you get point R starting from P and Q to achieve this? – m-stgt Feb 02 '24 at 02:09
  • @m-stgt I just assumed any three points on ellipse by parametric coordinate on Ellipse and assumed that three points form an equilateral triangle. – Mark Feb 02 '24 at 02:12
  • Just consider P and Q quite close on the ellipse (close in relation to either axis), the distance of P and Q define the length of all sides of an equilateral triangle. Only on distinct ellipses R may "sit" on the circumerence of it -- unless your triangle defines the ellipse. But up to now I regarded the shape of the ellipse as independent. – m-stgt Feb 02 '24 at 02:22
  • 1
    Please refer to another post here. – Ng Chung Tak Feb 02 '24 at 03:27
  • @NgChungTak I appreciate the answer but I don't know much complex analysis besides if you see the solution images you will find that the textbook provided solved it without complex analysis. Any idea about how can I prove the formula that book provided? – Mark Feb 02 '24 at 03:49
  • Lookup Corollary 2 in the "Concyclic points" section of your textbook, you should find there the answer. Which book is that, by the way? – Intelligenti pauca Feb 02 '24 at 10:51
  • @Intelligentipauca Arihant Skills in Mathematics for JEE Mains and Advance Coordinate Geometry by Dr. SK Goyal.In that this question is on page no. 551 subjective example 27 – Mark Feb 02 '24 at 13:47

1 Answers1

0

If you start with the coordinates of the vertices of the triangle as: $$ A=(a\cos\alpha,b\sin\alpha),\quad B=(a\cos\beta,b\sin\beta),\quad C=(a\cos\gamma,b\sin\gamma), $$ then you can find the coordinates of the circumcenter as the intersection of the perpendicular bisectors of two sides. The equation of the perpendicular bisector of side $AB$ is: $$ y-\frac{b}{2} (\sin \alpha+b \sin\beta) =\frac{a \cos \alpha-a \cos\beta}{b\sin\beta-b \sin \alpha} \left(x-\frac{a}{2} (\cos \alpha+\cos \beta)\right) $$ and you can find with the obvious substitutions the equation of the perpendicular bisector of side $BC$. Solving the system of those two equations gives a quite long solution, which can be put into the simple form: $$ x= \frac{a^2-b^2}{a}{ \cos \frac{\alpha+\beta}{2} \cos \frac{\alpha+\gamma}{2} \cos \frac{\beta+\gamma}{2}}, \quad y= \frac{b^2-a^2}{b}{ \sin \frac{\alpha+\beta}{2} \sin \frac{\alpha+\gamma}{2} \sin \frac{\beta+\gamma}{2}}. $$ Finally, using trig addition formulas, these can be rewritten in the form given by your textbook: $$ \begin{align} x&= \frac{a^2-b^2}{4a} [\cos \alpha+\cos\beta+\cos\gamma+\cos(\alpha+\beta+\gamma)], \\ \ \\ y&= \frac{b^2-a^2}{4b} [\sin \alpha+\sin\beta+\sin\gamma-\sin(\alpha+\beta+\gamma)]. \end{align} $$

EDIT. In your textbook this formula is derived in a smarter way, from the property that the eccentric angles of the four intersection points between an ellipse and a circle add up to a multiple of $2\pi$. See pages 497-498.

Intelligenti pauca
  • 50,470
  • 4
  • 42
  • 77
  • Thanks. I had tried writing equation of perpendicular bisectors as well but because the calculation was going too long. I discontinued and tried if I can use some geometrical trickery. Thanks BTW!! – Mark Feb 02 '24 at 13:57