Skip to content

6.5-exam3-2013-b

Quiz 3, 2013 · Section 6.5: Solving Linear Differential Equations Source: exam3_2.pdf, pp. 61-62, 66-67

Problem

3. (38 pts.) Let \(A = \begin{bmatrix} -1 & 1 \\ 1 & -1 \end{bmatrix}\).

a) (10 pts.) Find a nonzero solution \(y(t)\) in \(\mathbb{R}^2\) to \(dy/dt = Ay\) that is independent of \(t\), in other words, \(y(t)\) is a constant vector in \(\mathbb{R}^2\). (Hint: why would a vector in the nullspace of \(A\) have this property?)

b) (10 pts.) Show that \(e^{At}\) is Markov for every value of \(t \geq 0\).

c) (10 pts.) What is the limit of \(e^{At}\) as \(t \to \infty\)?

d) (8 pts.) What is the steady state vector of the Markov matrix \(e^A\)?

Solution a) If \(y(t)\) is constant, then \(\frac{dy}{dt} = 0\). So \(Ay = \frac{dy}{dt} \iff Ay = 0\), and therefore \(y\) lies in the nullspace of \(A\). We can take \[ y = \begin{bmatrix} 1 \\ 1 \end{bmatrix}. \] b) We calculate \(e^{At}\). The matrix \(A\) has eigenvalues \(-2, 0\) with corresponding eigenvectors \(\begin{bmatrix} 1 \\ -1 \end{bmatrix}, \begin{bmatrix} 1 \\ 1 \end{bmatrix}\). So \[ A = S\begin{bmatrix} -2 & 0 \\ 0 & 0 \end{bmatrix}S^{-1}, \qquad S = \begin{bmatrix} 1 & 1 \\ -1 & 1 \end{bmatrix}, \] and \[ e^{At} = S\begin{bmatrix} e^{-2t} & 0 \\ 0 & 1 \end{bmatrix}S^{-1} = \frac{1}{2}\begin{bmatrix} 1 + e^{-2t} & 1 - e^{-2t} \\ 1 - e^{-2t} & 1 + e^{-2t} \end{bmatrix}. \] Each column sums to 1: \[ \frac{1}{2}\big((1 + e^{-2t}) + (1 - e^{-2t})\big) = 1, \] and the entries are nonnegative for \(t \geq 0\), so \(e^{At}\) is Markov. c) As \(t \to \infty\), \(e^{-2t} \to 0\), so \[ e^{At} \to \begin{bmatrix} 1/2 & 1/2 \\ 1/2 & 1/2 \end{bmatrix}. \] d) The steady state of \(e^A\) is \[ \lim_{n \to \infty} e^{nA}v = \lim_{t \to \infty} e^{tA}v, \] where \(v\) is any probability vector, e.g. \(\begin{bmatrix} 1 \\ 0 \end{bmatrix}\). Thus \[ \lim_{t \to \infty} e^{tA}v = \begin{bmatrix} 1/2 & 1/2 \\ 1/2 & 1/2 \end{bmatrix}\begin{bmatrix} 1 \\ 0 \end{bmatrix} = \begin{bmatrix} 1/2 \\ 1/2 \end{bmatrix}. \]

Keywords: differential equation, matrix exponential, Markov matrix, steady state, nullspace