Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

Find the matrix for the transformation which first reflects across the main diagonal, then projects onto the line $2y+\sqrt{3}x=0$, and then reflects about the line $\sqrt{3}y=2x$.

Reflection about the line $y=x: T(x,y)=(y,x)$, so the standard matrix for this would be the matrix

$$\begin{pmatrix}0 &1 \\ 1 &0 \end{pmatrix}.$$

However I'm not sure how to deal with equations rather than axis. I assume in the second projection, you can simplify it to $y=\frac{-\sqrt{3}}{2} x$. Can you then separate these into a scalar operation $\frac{-\sqrt{3}}{2}$ and orthogonal operation $y=x$? Even so, I wouldn't know how to go further than this since I only know how to do it among the axis.

Edit: $T_1=\left[\begin{array}{cc}0&1\\1&0\end{array}\right]$ by reflecting across main diagonal

$2y+\sqrt{3}x=0\rightarrow y=\frac{-\sqrt{3}}{2}x$

$T_2=\frac{-\sqrt{3}}{2}\left[\begin{array}{cc}\cos-45&-\sin-45\\\sin-45&\cos-45\end{array}\right]\left[\begin{array}{cc}0&0\\0&1\end{array}\right]\left[\begin{array}{cc}\cos45&-\sin45\\\sin45&\cos45\end{array}\right]$ by rotating $45^\circ$, projecting along the y axis, then rotating $-45^\circ$ (transformations are applied right to left)

$\sqrt{3}y=2x\rightarrow y=\frac{2}{\sqrt{3}}x$

$T_3=\frac{2}{\sqrt{3}}\left[\begin{array}{cc}0&1\\1&0\end{array}\right]$ by separating fraction and then reflecting along $y=x$

share|cite|improve this question
You got a sign wrong for the equation of the line for the second transformation. – J. M. Apr 2 '13 at 18:28
Fixed, thanks for catching that. – Manu Juneja Apr 2 '13 at 18:34
I tried this method and edited the original. Can you please confirm that I did it correctly? I projected onto y axis instead because rotating 45 degrees would put it on the y axis – Manu Juneja Apr 2 '13 at 19:02
Trying that now, thanks. I edited my answer, I didn't realize that transformations are applied that way. – Manu Juneja Apr 2 '13 at 19:12
I've added my completed answer. Please let me know if it is correct. Additionally, do you know how I should format the answer? I'm not sure what format to put it in other than separating the three transformations – Manu Juneja Apr 2 '13 at 19:26

Okay, let's start with projections. The projection matrix onto a line $a x + b y = 0$ is a linear transformation expressible by a matrix, mapping the world onto points on that line. A typical point on that line has the form $t [b ;\; -a ]$ for some $t$, as this generates $a(b t) + b (-a t) = 0$. So the unit vector pointing in the direction of that line is $\hat u = [b ;\; -a] / \sqrt{a^2 + b^2}$ and the projection of a vector $\vec v$ is $\operatorname{proj}_{\hat u}~\vec v = \hat u (\hat u \cdot \vec v)$ which we can write as a matrix:$$\operatorname{proj}_{\hat u} = \frac{1}{a^2 + b^2} \begin{bmatrix}b\\-a\end{bmatrix}\begin{bmatrix}b & -a\end{bmatrix} = \frac{1}{a^2 + b^2} \begin{bmatrix}b^2 & -ba\\-ba & a^2\end{bmatrix}.$$So that's the projection matrix.

Once you have projections onto a line, you have reflections about the line. This is because if $\operatorname{proj}_{\hat u} \vec v = \vec v_u $ then we know $\vec v = \vec v_u + \vec c$ for some vector $\vec c$, and then the reflection about that line is just $\vec v_u - \vec c$: you flip the sign of the deviation, but you do not change the projection. Some thinking gives you an explicit construction as: $$\operatorname{flip}_{\hat u} \vec v = \operatorname{proj}_{\hat u} \vec v - (\vec v - \operatorname{proj}_{\hat u} \vec v) = 2 \operatorname{proj}_{\hat u} \vec v - \vec v$$so that the actual matrix which does this is $$\operatorname{flip}_{\hat u} = 2 \operatorname{proj}_{\hat u} - I.$$

Now indeed, if we're flipping around the line $y = x$ then $a = 1, b = -1$, and our projection matrix is $\frac 12 \begin{bmatrix}1&1\\1&1\end{bmatrix}$ and indeed, doubling this and subtracting the identity matrix gives $\begin{bmatrix}0&1\\1&0\end{bmatrix}.$

Your end result is therefore probably not right, since I do not see any factors of $\frac 17,$ which comes naturally from these $(a^2 + b^2)^{-1}$ prefactors.

share|cite|improve this answer

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

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