Question 3.34: Give an affine transformation T(x^→ ) = x0^→ + x^→A, where A...

Give an affine transformation

T\left(\overrightarrow{x} \right)=\overrightarrow{x_{0}}+\overrightarrow{x}A,  where  A = \left[\begin{matrix} -\frac{1}{3} & \frac{2}{3} & -\frac{2}{3} \\ \\  \frac{2}{3} & \frac{2}{3} & \frac{1}{3} \\ \\  \frac{2}{3} & -\frac{1}{3} & -\frac{2}{3} \end{matrix} \right] .

Try to determine these \overrightarrow{x_{0}} so that T is a rotation. In this case, determine the axis and the angle of the rotation, also the rotational plane.

The blue check mark means that this solution has been answered and checked by an expert. This guarantees that the final answer is accurate.
Learn more on how we answer questions.

The three row (or column) vectors of A are of unit length and are perpendicular to each other. So A^{*}=A^{-1} and A is orthogonal. Direct computation shows that

det A = \frac{1}{27}\left(4+4+4+8+8-1\right)=\frac{1}{27}\cdot27=1.

Therefore the associated T is a rotation of R³ if \overrightarrow{x_{0}} is chosen from Im\left(I_{3}-A\right).

The characteristic polynomial of A is

det \left(A-tI_{3}\right)=-\frac{1}{3}\left(t-1\right)\left(3t^{2}+4t+3\right)

and thus A has a real eigenvalue 1 of multiplicity one. Solve

\overrightarrow{x} \left(A-I_{3}\right)=\left(x_{1},x_{2},x_{3}\right)\left[\begin{matrix} -\frac{4}{3} & \frac{2}{3} & -\frac{2}{3} \\ \\  \frac{2}{3} & -\frac{1}{3} & \frac{1}{3} \\ \\ \frac{2}{3} & -\frac{1}{3} & -\frac{5}{3} \end{matrix} \right] =\overrightarrow{0}

\Rightarrow 2x_{1}-x_{2}-x_{3}=0,  2x_{1}-x_{2}+5x_{3}=0

\Rightarrow \overrightarrow{x}=x_{1}\left(1,2,0\right)  for  x_{1} \in R.

Hence, take an eigenvector \overrightarrow{v_{1}}=\frac{1}{\sqrt{5} }\left(1,2,0\right).

Let \overrightarrow{v}=\left(\alpha _{1},\alpha _{2},\alpha _{3}\right). Then

\overrightarrow{v} \bot \overrightarrow{v_{1}}

\Leftrightarrow \alpha _{1} +2\alpha _{2}=0.

Therefore, we may choose \overrightarrow{v_{2}}=\frac{1}{\sqrt{5} }\left(-2,1,0\right). Again, solve \alpha _{1} +2\alpha _{2}= -2\alpha _{1} +\alpha _{2}=0, we may choose \overrightarrow{v_{3}} =\left(0,0,1\right)=\overrightarrow{e_{3}}. Now, B = \left\{\overrightarrow{v_{1}},\overrightarrow{v_{2}},\overrightarrow{v_{3}}\right\} forms an orthonormal basis for R³.
By computation,

\overrightarrow{v_{2}}A=\frac{1}{3\sqrt{5} }\left(4,-2,5\right)=\frac{2}{3}\overrightarrow{v_{2}}+\frac{\sqrt{5}}{3 }\overrightarrow{v_{3}},

\overrightarrow{v_{3}}A=\frac{1}{3}\left(2,-1,-2\right)=-\frac{\sqrt{5}}{3 }\overrightarrow{v_{2}}+\frac{2}{3}\overrightarrow{v_{3}}

\Rightarrow \left[A\right]_{B}=PAP^{-1}=\left[\begin{matrix} 1 & 0 & 0 \\ \\  0 & \frac{2}{3} & \frac{\sqrt{5}}{3 } \\ \\  0 & -\frac{\sqrt{5}}{3 } & \frac{2}{3} \end{matrix} \right] ,  where  P = \left[\begin{matrix} \frac{1}{\sqrt{5} } & \frac{2}{\sqrt{5} } & 0 \\ \\  -\frac{2}{\sqrt{5} } & \frac{1}{\sqrt{5} } & 0 \\ \\  0 & 0 & 1 \end{matrix} \right].

To find Im \left(I_{3}-A\right), let \overrightarrow{y}=\overrightarrow{x}\left(I_{3}-A\right).Then

\overrightarrow{y}=\overrightarrow{x}\left(I_{3}-A\right)

\Leftrightarrow y_{1}=\frac{2}{3}\left(2x_{1}-x_{2}-x_{3}\right),  y_{2}=\frac{1}{3}\left(-2x_{1}+x_{2}+x_{3}\right) and

y_{3}=\frac{1}{3}\left(2x_{1}-x_{2}+5x_{3}\right)

\Leftrightarrow y_{1}+2y_{2}=0.

So the image subspace is x_{1}+2x_{2}=0 with a unit normal vector \overrightarrow{v_{1}}.Note that, since A^{*}=A^{-1}, Im \left(A-I_{3}\right)^{\bot}= Ker \left(A^{*}-I_{3}\right)=Ker \left(A^{-1}-I_{3}\right)=Ker \left(A-I_{3}\right). This is the theoretical reason why \overrightarrow{v_{1}}\bot Im \left(A-I_{3}\right)=Ker  \left(A-I_{3}\right)^{\bot}= \ll \overrightarrow{v_{2}},\overrightarrow{v_{3}}\gg.

Take any point \overrightarrow{x_{0}} on Im \left(A-I_{3}\right), say \overrightarrow{x_{0}}=\overrightarrow{0} for simplicity. Then, the axis of rotation is \ll \overrightarrow{v_{1}}\gg, the rotational plane is \ll \overrightarrow{v_{2}},\overrightarrow{v_{3}}\gg = Im \left(A-I_{3}\right) and the angle of rotation is \theta = \tan ^{-1} \frac{\sqrt{5}}{ 2}. See Fig. 3.70.

We have two questions:

Q1 What is the image of the unit sphere x^{2}_{1}+x^{2}_{2}+x^{2}_{3}=1 under \overrightarrow{y}=\overrightarrow{x}A?

Q2 What is the image of the cylinder x^{2}_{1}+x^{2}_{2}=1 under \overrightarrow{y}=\overrightarrow{x}A?

For Q1 It is the unit sphere itself both by geometric intuition and by analytic proof.

For Q2 Geometric intuition tells as that it is still a cylinder with central axis along the vector \overrightarrow{e_{3}}A=\left( \frac{2}{3}, -\frac{1}{3}, -\frac{2}{3}\right) and the base circle lying on the plane 2x_{1}-x_{2}-2x_{3}=0. And computation shows that the image has the equation, a complicated one,

5x^{2}_{1}+8x^{2}_{2}+5x^{2}_{3}+4x_{1}x_{2}+8x_{1}x_{3}-4x_{2}x_{3}=9  \left(*_{}\right)

in the eyes of N = \left\{\overrightarrow{0},\overrightarrow{e_{1}},\overrightarrow{e_{2}},\overrightarrow{e_{3}}\right\}.

Hence, we raise a third question concerning Q2 and \left(*_{}\right).

Q3 What is the equation of the image cylinder \left(*_{}\right) in the orthonormal affine basis C = \left\{\overrightarrow{0},A_{1*},A_{2*},A_{3*}\right\} where A_{i*} denotes the ith row vector of A for i = 1, 2, 3?

In B, one needs to compute

\overrightarrow{x} \left[\begin{matrix} 1 & 0 & 0 \\ 0 & 1 & 0 \\ 0 & 0 & 0 \end{matrix} \right]\overrightarrow{x^{*}}

= \left(\overrightarrow{y}P^{-1}\right) \left(PAP^{-1}\right)P\left[\begin{matrix} 1 & 0 & 0 \\ 0 & 1 & 0 \\ 0 & 0 & 0 \end{matrix} \right]P^{-1}\left(PAP^{-1}\right)^{-1}\left(\overrightarrow{y}P^{-1}\right)^{*}=1,

where \overrightarrow{y}=\overrightarrow{x}A  and \overrightarrow{y}P^{-1} is the coordinate vector of \overrightarrow{y} in the basis B.

In C, one needs to compute

\overrightarrow{x} \left[\begin{matrix} 1 & 0 & 0 \\ 0 & 1 & 0 \\ 0 & 0 & 0 \end{matrix} \right]\overrightarrow{x^{*}}=\left(\overrightarrow{y}A^{-1}\right)\left[\begin{matrix} 1 & 0 & 0 \\ 0 & 1 & 0 \\ 0 & 0 & 0 \end{matrix} \right]\left(\overrightarrow{y}A^{-1}\right)^{*}= \alpha^{2}_{1}+\alpha^{2}_{2}=1,

where \overrightarrow{y}A^{-1}=\left(\alpha _{1} ,\alpha _{2} ,\alpha _{3}\right) is the coordinate vector of \overrightarrow{y}=\overrightarrow{x} A in the basis C. Hence, to view the image cylinder, under\overrightarrow{y}=\overrightarrow{x} A, in C is the same as to view the original cylinder in N.

3

Related Answered Questions

Question: 3.37

Verified Answer:

Both S_{1}  and  S_{2} are two-dime...
Question: 3.26

Verified Answer:

Analysis The characteristic polynomial is ...