Linear Algebra I: Homework 10

Due: Friday, November 10, 2017
  1. For vectors \(\vec v = \begin{pmatrix}v^1 \\ v^2\end{pmatrix}\) and \(\vec w = \begin{pmatrix}w^1 \\ w^2\end{pmatrix}\), \(\langle \vec v, \vec w \rangle = 3v^1w^1 + 2v^2w^2\) is an inner product.
    1. Find all unit vectors in \(\mathbb R^2\) with respect to this new inner product.

      See Homework 7.2.a; this is similar but with a different magnitude. Unit vectors are the ellipse determined by \(1 = 3x^2 + 2y^2\), that is, the set of vectors;

      \[\left\{ \begin{pmatrix} \frac{1}{\sqrt 3} \cos \theta \\ \frac{1}{\sqrt 2} \sin \theta \end{pmatrix} \middle| \theta \in [0, 2\pi) \right\}.\]
    2. Find two different orthonormal bases for \(\mathbb R^2\) with respect to this new inner product.

      Our solution for (a) says that two vectors determined by angles \(\alpha\) and \(\beta\) have dot product,

      \[\frac{3}{\sqrt 9}\cos \alpha \cos \beta + \frac{2}{\sqrt 4}\sin \alpha \sin \beta = \cos\alpha\cos\beta + \sin\alpha \sin \beta\]

      This means that two of our unit vectors are orthogonal if \(\alpha\) and \(\beta\) form a right angle (why?). So to come up with orthonormal bases we just have to pick an angle, and then another angle which is orthogonal to it:

      \[\left\{ \begin{pmatrix} \frac{1}{\sqrt 3} \cos \theta \\ \frac{1}{\sqrt 2} \sin \theta \end{pmatrix}, \begin{pmatrix} \frac{1}{\sqrt 3} \cos (\theta+\pi/2) \\ \frac{1}{\sqrt 2} \sin (\theta+\pi/2) \end{pmatrix} \right\}.\]

      For example, we get two different bases if we pick \(\theta = 0\) and \(\theta = \pi/4\).

  2. If \(\vec v = \begin{pmatrix} 3 \\ 2 \\ 1\end{pmatrix}\) and \(\vec u = \begin{pmatrix} -1 \\ 0 \\ 1\end{pmatrix}\), find a decomposition

    \[\vec v = \vec v^{||} + \vec v^{\perp}\]

    where \(\vec v^{||}\) is parallel to \(\vec u\) and \(\vec v^{\perp}\) is orthogonal to \(\vec u\).

    Gram-Schmidt says that

    \[\vec v^{||} = \frac{\vec v \cdot \vec u}{\vec u \cdot \vec u}\vec u = \frac{-3 + 1}{1 + 1}\begin{pmatrix} -1 \\ 0 \\ 1\end{pmatrix} = \begin{pmatrix} 1 \\ 0 \\ -1\end{pmatrix}.\]

    Then,

    \[\vec v^{\perp} = \vec v - \vec v^{||} = \begin{pmatrix} 3 \\ 2 \\ 1\end{pmatrix} - \begin{pmatrix} 1 \\ 0 \\ -1\end{pmatrix} = \begin{pmatrix} 2 \\ 2 \\ 2\end{pmatrix}.\]
  3. For two \(m \times n\) matrices \(M, N\) we can define the inner product,

    \[\langle M, N \rangle = \mathop{tr}(M^T N).\]

    Are the vectors,

    \[\begin{pmatrix}2&1 \\ -1&3\end{pmatrix} \quad\textrm{and}\quad \begin{pmatrix}-3&0 \\ 0&2\end{pmatrix}\]

    orthogonal? Explain why or why not.

    Yes, because \(\langle M, N \rangle = \mathop{tr}(M^T N) = 0\).