Skip to content

6.2-exam3-2001

Quiz 3, 2001 · Section 6.2: Diagonalizing a Matrix Source: exam3_1.pdf, p. 54

Problem

For full credit, carefully explain your reasoning, as always!

1 (36 pts.) Let \(A\) be the square matrix

\[ A = \begin{bmatrix} 2 & 1 \\ x & y \end{bmatrix}. \]

(a) With \(x = 2\) and \(y = 1\) diagonalize \(A\). That is, compute \(A = S\Lambda S^{-1}\), where \(\Lambda\) is a diagonal matrix. (12p)

(b) With \(y = 2\) pick \(x\) so that \(S\) can be orthogonal in a diagonalization of \(A\). Compute then one such \(S\). (12p)

(c) If \(y = 2\), can you find \(x > 0\) such that \(A\) and \(\begin{bmatrix} 2 & 1 \\ 2 & 1 \end{bmatrix}\) are similar? (Hint: look at the eigenvalues.) (12p)

Solution (a) With \(x = 2\) and \(y = 1\), \(A = \begin{bmatrix} 2 & 1 \\ 2 & 1 \end{bmatrix}\). Since the columns are identical, \(\det A = 0\), so one eigenvalue is \(0\); the trace is \(3\), so the other eigenvalue is \(3\). Eigenvectors: for \(\lambda = 3\), \((A - 3I)v = 0\) gives \(-v_1 + v_2 = 0\), so \(v_1 = (1, 1)\); for \(\lambda = 0\), \(Av = 0\) gives \(2v_1 + v_2 = 0\), so \(v_2 = (1, -2)\). Thus \[ A = S\Lambda S^{-1} = \begin{bmatrix} 1 & 1 \\ 1 & -2 \end{bmatrix} \begin{bmatrix} 3 & 0 \\ 0 & 0 \end{bmatrix} \frac{1}{3} \begin{bmatrix} 2 & 1 \\ 1 & -1 \end{bmatrix}. \] (b) With \(y = 2\), the matrix is \(A = \begin{bmatrix} 2 & 1 \\ x & 2 \end{bmatrix}\). It is symmetric exactly when \(x = 1\), and a symmetric matrix has orthogonal eigenvectors. For \(A = \begin{bmatrix} 2 & 1 \\ 1 & 2 \end{bmatrix}\) the eigenvalues are \(3\) and \(1\), with eigenvectors \((1, 1)\) and \((1, -1)\). Normalizing them gives an orthogonal \(S\): \[ S = \frac{1}{\sqrt{2}} \begin{bmatrix} 1 & 1 \\ 1 & -1 \end{bmatrix}, \qquad A = S \begin{bmatrix} 3 & 0 \\ 0 & 1 \end{bmatrix} S^T. \] (c) No. The matrix \(B = \begin{bmatrix} 2 & 1 \\ 2 & 1 \end{bmatrix}\) has eigenvalues \(3\) and \(0\), so its trace is \(3\). Any matrix similar to \(B\) must have the same trace. But with \(y = 2\), the trace of \(A\) is always \(2 + 2 = 4 \neq 3\). Hence no \(x > 0\) (in fact no \(x\) at all) makes \(A\) similar to \(B\).

Keywords: diagonalization, orthogonal matrix, eigenvalues, similar matrices