6.3-final-2011¶
Final Exam, 2011 · Section 6.3: Symmetric Positive Definite Matrices Source: finalexam_2.pdf, pp. 45-46
Problem¶
6 (20 pts.) Let
(a) (4 pts.) There are two values of \(c\) that make \(A\) a projection matrix. Find them by guessing, calculating, or understanding projection matrices. Check that \(A\) is a projection matrix for these two \(c\).
(b) (4 pts.) There are two values of \(c\) that make \(A\) an orthogonal matrix. Find them and check that \(A\) is orthogonal for these two \(c\).
(c) (4 pts.) For which values of \(c\) is \(A\) diagonalizable?
(d) (4 pts.) Find the eigenvalues of \(A^{-1}\) (if it exists) in terms of \(c\). (Hint: find the eigenvalues of \(E_3\) first.)
(e) (4 pts.) For which values of \(c\) is \(A\) positive definite?
Solution
(a) A projection satisfies \(A^2 = A\). Since \(E^2 = 3E\) (with \(E = E_3\) the all-ones matrix), \[ A^2 = I - 2cE + c^2E^2 = I - 2cE + 3c^2E = I - cE = A \] exactly when \(3c^2 = c\), i.e. \(c = 0\) or \(c = 1/3\). Thus \[ A = I \quad \text{or} \quad A = \frac{1}{3} \begin{bmatrix} 2 & -1 & -1 \\ -1 & 2 & -1 \\ -1 & -1 & 2 \end{bmatrix}, \] and each of these squares to itself. (b) \(A\) is symmetric, so \(A^T A = A^2\). Orthogonality requires \[ I = A^T A = A^2 = I - 2cE + 3c^2E, \] i.e. \(3c^2 = 2c\), so \(c = 0\) or \(c = 2/3\). Thus \[ A = I \quad \text{or} \quad A = \frac{1}{3} \begin{bmatrix} 1 & -2 & -2 \\ -2 & 1 & -2 \\ -2 & -2 & 1 \end{bmatrix}, \] and the second matrix squares to the identity, so it is orthogonal. (c) The matrix is symmetric for every \(c\), so all values of \(c\) make \(A\) diagonalizable. (d) \(E_3\) is rank 1 with trace 3, so its eigenvalues are \(3, 0, 0\). Then \(A = I - cE\) has eigenvalues \(1 - 3c, 1, 1\), and \(A^{-1}\) (which exists for \(c \ne 1/3\)) has eigenvalues \[ \frac{1}{1 - 3c}, \ 1, \ 1. \] (e) \(A\) is positive definite when all eigenvalues are positive: \(1 - 3c > 0\) (and \(1 > 0\)), so \(c < 1/3\).Keywords: projection matrix, orthogonal matrix, positive definite, eigenvalues