Is there an easier way to solve a "Find the locus" problem?

212 Views Asked by At

Note: I am not concerned with the accuracy of my solution so you don't need to redo any of my calculations.

Original question: Suppose $ABC$ is an equilateral triangular lamina of side length unity, resting in two-dimensions. If $A$ and $B$ were constrained to move on the $x$- and $y$-axis respectively, then what is the locus of the centre of the triangle?

My solution: Let the vertices of $\Delta ABC$ be $A=(x_0,0)$, $B=(0,y_0)$ and $C=(x,y)$. Let the centre of the triangle be $D=(X,Y)$. We wish to find the locus of $D$ under the constraints: \begin{align} x_0^2 + y_0^2 &= 0\\ (x-x_0)^2 + y^2 &= 0\\ x^2 + (y-y_0)^2 &= 0 \end{align} Since $D$ is the centre of the triangle, we know $$X = \frac{x+x_0}{3}\quad\text{and}\quad Y = \frac{y+y_0}{3}$$

Parametrizing: Let $\theta$ be the angle that the edge $AB$ makes with the $x$-axis. We use (\theta) as our parameter of choice to write, \begin{align*} x_0 &= \cos\theta\\ y_0 &= \sin\theta \end{align*} for $\theta\in[0,2\pi)$.

This implies, $$(x-\cos\theta)^2 + y^2 = 0$$ We can parametrize this with another parameter $\phi\in[0,2\pi)$ as \begin{align*} x &= \cos\theta + \cos\phi\\ y &= \sin\phi \end{align*} Plugging these back into third equation gives $$(\cos\theta+\cos\phi)^2+(\sin\phi-\sin\theta)^2 = 1$$ After lots of algebraic manipulations... \begin{align*} X &= \frac{1}{2}\cos\theta\pm\frac{1}{2\sqrt{3}}\sin\theta & Y &= \frac{1}{2}\sin\theta\pm\frac{1}{2\sqrt{3}}\cos\theta\\ \end{align*}

Deparametrizing: To obtain equation for the locus of $D$, we eliminate the parameter. Solving for $\sin\theta$ and $\cos\theta$ in terms of $X$ and $Y$, we get $$\cos\theta = 3X\mp\sqrt{3}Y \quad\text{and}\quad \sin\theta = 3Y\mp\sqrt{3}X$$ Therefore, $X$ and $Y$ satisfy $$(3X\mp\sqrt{3}Y)^2+(3Y\mp\sqrt{3}X)^2=1$$ i.e., $$12X^2+12Y^2\mp12\sqrt{3}XY=1$$ These are two oblique ellipses.

My actual questions:

  1. Is there an easier way to solve the question i.e. was there no need to parametrize, seeing as I am deparametrizing anyway?
  2. Even before doing any of the calculations, is there some result/theorem that guarantees that a relationship among $X, Y$ and $Z$ exists. (Subquestion: is it possible that making the problem more complicated by say looking at equilateral tetrahedrons in 3D or something might not even be solvable?)

Note: I am not concerned with the accuracy of my solution so you don't need to redo any of my calculations.

1

There are 1 best solutions below

3
On

It is probably easier to embed the construction in the complex plane and set $$ A=\sin\theta,\qquad B=i\cos\theta $$ to have $$ C = (A-B)e^{\pm i\pi/3}+B = e^{-i\theta\pm i\pi/3}+B$$ so that: $$ 3D = A+B+C = e^{i\theta}+e^{-i\theta\pm i\pi/3}+i\cos\theta. $$ That leads to the same equations for the ellipses.
If three vertices of a tetrahedron lie on the $x-,y-$ and $z-$ axis we may assume they are $A(a,0,0),B(0,b,0),C(0,0,c)$ and $a^2+b^2=a^2+c^2=b^2+c^2$ must hold, so $|a|=|b|=|c|$ and there are just $16$ possible locations for $D$ and the centroid $G$ of $ABCD$.