6.5-exam3-2013¶
Quiz 3, 2013 · Section 6.5: Solving Linear Differential Equations Source: exam3_2.pdf, pp. 59-60, 64-66
Problem¶
2. (30 pts.) (3 points each: Please circle true or false, and either way, explain briefly.)
a) If \(A\) and \(B\) are invertible, then so is \((A + B)/2\). True? False?
b) If \(A\) and \(B\) are Markov, then so is \((A + B)/2\). True? False?
c) If \(A\) and \(B\) are positive definite, then so is \((A + B)/2\). True? False?
d) If \(A\) and \(B\) are diagonalizable, then so is \((A + B)/2\). True? False?
e) If \(A\) and \(B\) are rank 1, then so is \((A + B)/2\). True? False?
f) If \(A\) is symmetric then so is \(e^A\). True? False?
g) If \(A\) is Markov then so is \(e^A\). True? False?
h) If \(A\) is symmetric, then \(e^A\) is positive definite. True? False?
i) If \(A\) is singular, then so is \(e^A\). True? False?
j) If \(A\) is orthogonal, then so is \(e^A\). True? False?
Solution
a) No. For example, \(A = I\), \(B = -I\): \((A + B)/2 = 0\) is not invertible. b) True. The \(k\)-th column sum of \((A + B)/2\) is \[ \frac{(\text{\(k\)-th column sum of }A) + (\text{\(k\)-th column sum of }B)}{2} = \frac{1 + 1}{2} = 1, \] and the entries stay nonnegative, so \((A + B)/2\) is Markov. c) True. \[ x^T\left(\frac{A + B}{2}\right)x = \frac{x^TAx + x^TBx}{2} > 0 \quad \text{if } x^TAx, x^TBx > 0. \] d) False. \(A = \begin{bmatrix} 2 & 0 \\ 0 & 0 \end{bmatrix}\) and \(B = \begin{bmatrix} 0 & 2 \\ 0 & 2 \end{bmatrix}\) are both diagonalizable, but \((A + B)/2 = \begin{bmatrix} 1 & 1 \\ 0 & 1 \end{bmatrix}\) is not. e) False. \(A = \begin{bmatrix} 2 & 0 \\ 0 & 0 \end{bmatrix}\) and \(B = \begin{bmatrix} 0 & 0 \\ 0 & 2 \end{bmatrix}\) are both rank 1, but \((A + B)/2 = I\) is not. f) True. In the sum \[ e^A = I + A + \frac{A^2}{2!} + \dots \] every term is symmetric, so the sum is symmetric. g) False. \(A = \begin{bmatrix} 1 & 0 \\ 0 & 1 \end{bmatrix}\) is Markov but \(e^A = \begin{bmatrix} e & 0 \\ 0 & e \end{bmatrix}\) is not. h) True. If \(\lambda_1, \dots, \lambda_n\) are the eigenvalues of \(A\), then \(e^{\lambda_1}, \dots, e^{\lambda_n}\) are the eigenvalues of \(e^A\), and \(e^{\lambda_i}\) is positive since \(\lambda_i\) is real. i) False. \(A = \begin{bmatrix} 0 & 0 \\ 0 & 0 \end{bmatrix}\) is singular but \(e^A = I\) is not. j) False. \(A = \begin{bmatrix} 1 & 0 \\ 0 & 1 \end{bmatrix}\) is orthogonal but \(e^A = \begin{bmatrix} e & 0 \\ 0 & e \end{bmatrix}\) is not.Keywords: matrix exponential, Markov matrix, positive definite, symmetric matrix, eigenvalues