6.2-exam3-1997¶
Quiz 3, 1997 · Section 6.2: Diagonalizing a Matrix Source: exam3_1.pdf, pp. 1, 3
Problem¶
1.
(a) (20 points) Compute \(u_{20} = A^{20}u_0\) starting from
\[
A = \begin{bmatrix} .7 & .5 \\ .3 & .5 \end{bmatrix} \quad \text{and} \quad u_0 = \begin{bmatrix} 1 \\ 0 \end{bmatrix}.
\]
(b) (10 points) For the same \(A\), which real numbers \(c\) have the property that \((A - cI)^n u_0\) approaches zero as \(n \to \infty\)?
(c) (10 points) Find the eigenvalues and eigenvectors of \(A^{-1} + A^{20}\) without computing \(A^{-1}\) (for the same \(A\) as above).
Solution
(a) \(A\) is a Markov matrix, so \(\lambda = 1\) is an eigenvalue. Then \(\lambda = .2\) is the other eigenvalue because the trace is \(1.2\). Eigenvectors: \[ \lambda = 1: \quad A - I = \begin{bmatrix} -.3 & .5 \\ .3 & -.5 \end{bmatrix} \to x = \begin{bmatrix} 5 \\ 3 \end{bmatrix}, \qquad \lambda = .2: \quad A - .2I = \begin{bmatrix} .5 & .5 \\ .3 & .3 \end{bmatrix} \to x = \begin{bmatrix} 1 \\ -1 \end{bmatrix}. \] \[ S = \begin{bmatrix} 5 & 1 \\ 3 & -1 \end{bmatrix} \quad \text{has inverse} \quad \frac{1}{8} \begin{bmatrix} 1 & 1 \\ 3 & -5 \end{bmatrix} = S^{-1}, \quad \text{so that} \quad S^{-1}u_0 = \begin{bmatrix} 1 \\ 3 \end{bmatrix} \frac{1}{8}. \] We want \[ A^{20}u_0 = S\Lambda^{20}S^{-1}u_0 = \begin{bmatrix} 5 & 1 \\ 3 & -1 \end{bmatrix} \begin{bmatrix} 1 & 0 \\ 0 & .2^{20} \end{bmatrix} \begin{bmatrix} 1 \\ 3 \end{bmatrix} \frac{1}{8} = \begin{bmatrix} 5 + 3(.2^{20}) \\ 3 - 3(.2^{20}) \end{bmatrix} \frac{1}{8}. \] Check: change the 20th power to the 0th and we get \(\begin{bmatrix} 1 \\ 0 \end{bmatrix} = u_0\). (b) \(A\) has eigenvalues \(\lambda = .2\) and \(1\). \(A - cI\) has eigenvalues \(.2 - c\) and \(1 - c\). We need \(|.2 - c| < 1\) (need \(c < 1.2\)) and \(|1 - c| < 1\) (need \(c > 0\)). Therefore the condition for \(|\lambda| < 1\) and stability is \[ 0 < c < 1.2. \] (c) Same eigenvectors for \(A^{-1} + A^{20}\) as in part (a) for \(A\) itself. Eigenvalues: \[ \frac{1}{1} + 1^{20} = 2 \qquad \text{and} \qquad \frac{1}{.2} + (.2)^{20}. \]Keywords: Markov matrix, matrix powers, eigenvalues, stability