Skip to content

6.1-exam3-2013-strang

Quiz 3, 2013 (strang) · Section 6.1: Introduction to Eigenvalues : Ax = λx Source: exam3_2.pdf, pp. 68, 74

Problem

(1) (40 pts.) In all of this problem, the 3 by 3 matrix \(A\) has eigenvalues \(\lambda_1, \lambda_2, \lambda_3\) with independent eigenvectors \(x_1, x_2, x_3\).

(a) What are the trace of \(A\) and the determinant of \(A\)?

(b) Suppose: \(\lambda_1 = \lambda_2\). Choose the true statement from 1, 2, 3: 1. \(A\) can be diagonalized. Why? 2. \(A\) can not be diagonalized. Why? 3. I need more information to decide. Why?

(c) From the eigenvalues and eigenvectors, how could you find the matrix \(A\)? Give a formula for \(A\) and explain each part carefully.

(d) Suppose \(\lambda_1 = 2\) and \(\lambda_2 = 5\) and \(x_1 = (1, 1, 1)\) and \(x_2 = (1, -2, 1)\). Choose \(\lambda_3\) and \(x_3\) so that \(A\) is symmetric positive semidefinite but not positive definite.

Solution (a) The trace of \(A\) is \(\lambda_1 + \lambda_2 + \lambda_3\) and the determinant is \(\lambda_1\lambda_2\lambda_3\). (b) (1) is the correct option, because we know that there exists a full set of independent eigenvectors. (c) We can recover \(A\) using \(A = S\Lambda S^{-1}\), where \(S\) is a matrix whose columns are \(x_1, x_2, x_3\), and \(\Lambda\) is a diagonal matrix whose diagonal entries are \(\lambda_1, \lambda_2, \lambda_3\). (d) If we want \(A\) to be symmetric, the third eigenvector \(x_3\) had better be orthogonal to the other two. The quick way to find a vector orthogonal to two given ones in \(\mathbb{R}^3\) is via the cross product: \(x_3 = x_1 \times x_2 = (3, 0, -3)\). Alternately, you can use elimination: \(x_3\) should be in the nullspace of \[ \begin{bmatrix} 1 & 1 & 1 \\ 1 & -2 & 1 \end{bmatrix}. \] You could also just notice the first and last entries match and guess the answer from that. Either way, \(x_3\) should be a multiple of \((1, 0, -1)\). As for the eigenvalue, to get a matrix that's positive semidefinite but not positive definite, we need to use \(\lambda_3 = 0\). It doesn't actually ask you to compute \(A\), but here's one that works: \[ A = \begin{bmatrix} 1 & 1 & 1 \\ 1 & -2 & 0 \\ 1 & 1 & -1 \end{bmatrix}\begin{bmatrix} 2 & 0 & 0 \\ 0 & 5 & 0 \\ 0 & 0 & 0 \end{bmatrix}\begin{bmatrix} 1 & 1 & 1 \\ 1 & -2 & 0 \\ 1 & 1 & -1 \end{bmatrix}^{-1} = \frac{1}{2}\begin{bmatrix} 3 & -2 & 3 \\ -2 & 8 & -2 \\ 3 & -2 & 3 \end{bmatrix}. \]

Keywords: eigenvalues, eigenvectors, trace, determinant, positive semidefinite