Skip to content

7.1-exam3-2010

Quiz 3, 2010 · Section 7.1: Singular Values and Singular Vectors Source: exam3_2.pdf, pp. 5-6, 12

Problem

2. (35 points) Let

\[ A = - \begin{bmatrix} 1/4 & 1/4 & 1/4 & 1/4 \\ 1/4 & 1/4 & 1/4 & 1/4 \\ 1/4 & 1/4 & 1/4 & 1/4 \\ 1/4 & 1/4 & 1/4 & 1/4 \end{bmatrix}. \]

(Note the minus sign in the definition of \(A\).)

(a) (15 pts.) Write down a valid SVD for \(A\). (No partial credit for this one so be careful.)

(b) (20 pts.) The \(4 \times 4\) matrix \(e^{At} = I + f(t)A\). Find the scalar function \(f(t)\) in simplest possible form. (Hint: the power series is one way; eigendecomposition is another.)

Solution (a) \(A\) is clearly rank 1, so a reduced SVD like we saw in class is the easiest: \[ A = U\Sigma V^T = \begin{bmatrix} 1/2 \\ 1/2 \\ 1/2 \\ 1/2 \end{bmatrix} (1) \begin{bmatrix} -1/2 & -1/2 & -1/2 & -1/2 \end{bmatrix}. \] (b) Answer 1: notice that \(A^2 = -A\), so \(e^{At} = I + A(t - t^2/2 + t^3/3! - t^4/4! + \cdots)\). Thus \(f(t) = 1 - e^{-t}\). Answer 2: the eigenvalues of the symmetric rank 1 matrix \(A\) are \(-1, 0, 0, 0\). In matrix language \(\Lambda = \begin{bmatrix} -1 & 0 \\ 0 & 0 \end{bmatrix}\) and \[ e^{\Lambda t} - I = \begin{bmatrix} e^{-t} - 1 & 0 \\ 0 & 0 \end{bmatrix} = (1 - e^{-t})\Lambda. \] If \(A = Q\Lambda Q^T\), then \(e^{At} - I = Q(e^{\Lambda t} - I)Q^T = (1 - e^{-t})Q\Lambda Q^T = (1 - e^{-t})A\), giving the same answer \(f(t) = 1 - e^{-t}\).

Keywords: singular value decomposition, matrix exponential, eigenvalues