Skip to content

6.3-exam3-2009-spring

Quiz 3, 2009 (spring) · Section 6.3: Symmetric Positive Definite Matrices Source: exam3_1.pdf, p. 162

Problem

Problem 4: \(A\) is a \(3 \times 3\) real-symmetric matrix. Two of its eigenvalues are \(\lambda_1 = 1\) and \(\lambda_2 = -1\) with eigenvectors \(v_1 = (1, 1, 1)^T\) and \(v_2 = (1, -1, 0)^T\), respectively. The third eigenvalue is \(\lambda_3 = 0\).

(I) Give an eigenvector \(v_3\) for the eigenvalue \(\lambda_3\). (Hint: what must be true of \(v_1\), \(v_2\), and \(v_3\)?)

(II) Using your result from (I), write the matrix \(e^A\) as the product of three matrices, and explicitly give the three matrices. (You need not work out the arithmetic, but your answer should contain no matrix inverses or matrix exponentials. If you find yourself doing a lot of arithmetic, you are forgetting a useful property of this matrix!)

Solution (I) For a real-symmetric matrix, its eigenvectors are orthogonal to each other. So, by inspection, in order for \(v_3\) to be perpendicular to \(v_2\), we need its first two components the same. Hence, we should take \(v_3\) to be \((1, 1, -2)^T\). To ease the second part, we can normalize the eigenvectors: \[ q_1 = v_1/\|v_1\| = (1, 1, 1)^T/\sqrt{3}, \] \[ q_2 = v_2/\|v_2\| = (1, -1, 0)^T/\sqrt{2}, \] \[ q_3 = v_3/\|v_3\| = (1, 1, -2)^T/\sqrt{6}. \] Alternatively, we can use Gram-Schmidt to find (a multiple of) \(v_3\) as follows. We start with \(v = (1, 0, 0)\): \[ v_3 = v - (v \cdot q_1)q_1 - (v \cdot q_2)q_2 = (1, 0, 0)^T - \frac{1}{3}(1, 1, 1)^T - \frac{1}{2}(1, -1, 0)^T = \left(\frac{1}{6}, \frac{1}{6}, -\frac{1}{3}\right)^T. \] (II) We can write \(A = S\Lambda S^{-1} = Q\Lambda Q^T\): \[ A = \begin{bmatrix} 1/\sqrt{3} & 1/\sqrt{2} & 1/\sqrt{6} \\ 1/\sqrt{3} & -1/\sqrt{2} & 1/\sqrt{6} \\ 1/\sqrt{3} & 0 & -2/\sqrt{6} \end{bmatrix} \begin{bmatrix} 1 & 0 & 0 \\ 0 & -1 & 0 \\ 0 & 0 & 0 \end{bmatrix} \begin{bmatrix} 1/\sqrt{3} & 1/\sqrt{3} & 1/\sqrt{3} \\ 1/\sqrt{2} & -1/\sqrt{2} & 0 \\ 1/\sqrt{6} & 1/\sqrt{6} & -2/\sqrt{6} \end{bmatrix}. \] Hence, \[ e^A = Qe^{\Lambda}Q^T \] \[ = \begin{bmatrix} 1/\sqrt{3} & 1/\sqrt{2} & 1/\sqrt{6} \\ 1/\sqrt{3} & -1/\sqrt{2} & 1/\sqrt{6} \\ 1/\sqrt{3} & 0 & -2/\sqrt{6} \end{bmatrix} \begin{bmatrix} e & 0 & 0 \\ 0 & 1/e & 0 \\ 0 & 0 & 1 \end{bmatrix} \begin{bmatrix} 1/\sqrt{3} & 1/\sqrt{3} & 1/\sqrt{3} \\ 1/\sqrt{2} & -1/\sqrt{2} & 0 \\ 1/\sqrt{6} & 1/\sqrt{6} & -2/\sqrt{6} \end{bmatrix}. \]

Keywords: symmetric matrix, orthonormal eigenvectors, matrix exponential, orthogonal diagonalization