Linear Algebra I: Homework 4

Due: Friday, February 23, 2018
\[\newcommand{\tr}{\operatorname{tr}}\]
  1. The rotation matrix \(R_\theta\) is a matrix whose multiplication rotates vectors in \(\mathbb R^2\) by \(\theta\) radians counterclockwise. \(R_\theta\) has the formula,

    \[\begin{pmatrix} \cos \theta & -\sin \theta \\ \sin \theta & \cos \theta \end{pmatrix}\]
    1. Calculate the determinant of \(R_\theta\). Does it depend on \(\theta\)?

      The determinant is \(\cos^2(\theta)+\sin^2(\theta) = 1\), so it doesn’t depend on \(\theta\).

      That the determinant is 1 means that the matrix \(R_\theta\) doesn’t change volumes of shapes.

    2. Without doing any Gaussian elimination or using the formula for inverses of \(2 \times 2\) matrices, find a concise formula for \((R_\theta)^{-1}.\) (Hint: Think about what \(R_\theta\) does geometrically).

      The matrix rotates everything simultaneously by \(\theta\) radians. How can you undo a rotation? Rotating the other way! You do that by rotating just the negative amount of radians.

      So, \(R_\theta^{-1} = R_{-\theta}\)

    3. Find a non-identity matrix \(B \ne I\) for which \(B^3 = BBB = I\).

      We can use a rotation matrix for an example here. The matrix \(R_{2\pi/3}\) will rotate things by a third of the way around; rotating by this much exactly three times yields a full rotation. But nothing changes after a full rotation! So \(R_{2\pi/3}^3 = I\).

  2. If \(A\) is an \(n \times n\) symmetric matrix (that is, \(A^T = A\)), and \(B\) is any \(n \times m\) matrix, show that the following are symmetric:

    1.  \(B^TB\)

      We check, \((B^TB)^T = B^T (B^T)^T = B^TB\)

    2.  \(BB^T\)

      We check, \((BB^T)^T = (B^T)^T B^T = BB^T\)

    3.  \(B^TAB\)

      We check, \((B^TAB)^T = B^T A^T (B^T)^T = B^TAB\) because \(A^T = A\).

  3. Find the following determinants:

    1. \[\det \begin{pmatrix}2 & 3 & 7 & 0 \\ 0 & -3 & 1 & 3 \\ 0 & 0 & -1 & 1 \\ 0 & 0 & 0 & -2\end{pmatrix}\]

      The matrix is upper triangular, so the determinant is the product of its diagonals. So, the determinant is \((2)(-3)(-1)(-2) = -12\).

    2. \[\det \begin{pmatrix}1 & -4 & 1 \\ 0 & 3 & 1 \\ 2 & -8 & 2\end{pmatrix}\]

      The first and last rows are multiples of each other. So, the matrix can’t be invertible, and its determinant is \(0\).

    1. \(A\) is an \(n \times n\) matrix and \(\det(A) = 2\), find \(\det(3A)\). (Hint: First calculate \(\det(3I)\), then use that \(\det(AB) = \det(A)\det(B)\).)

      \(3I_n\) is the \(n \times n\) matrix with 3s on its diagonal, so its determinant is the product of the diagonals. There are \(n\) copies of 3, so \(\det(3I)=3^n\).

      The determinant is multiplicative, so \(\det(3A) = \det(3IA) = \det(3I)\det(A) = 3^n 2\).

    2. Is \(\det\) a linear function on matrices?

      No. Part (a) gave a counterexample to homogeneity (determinant doesn’t work well with scaling)

    1. For an \(n \times n\) matrix \(A = (a^i_j)\) and a number \(\lambda\), compare \(\tr(\lambda A)\) and \(\lambda \tr(A)\).

      They are the same, since sums are linear. So, trace works well with scaling.

    2. For a pair of \(n \times n\) matrices \(A = (a^i_j)\) and \(B = (b^i_j)\), compare \(\tr(A+B)\) and \(\tr(A) + \tr(B)\).

      They are the same, since sums are linear.

    3. Is \(\tr\) a linear function on matrices?

      Yes. We just proved the two things to check above.