6.5-final-2014-strang¶
Final Exam, 2014 · Section 6.5: Solving Linear Differential Equations Source: finalexam_2.pdf, p. 140
Problem¶
6 (10 points)
(a) Find the eigenvalues of this matrix \(A\) (the numbers in each column add to zero):
\[
A = \begin{bmatrix} -1 & \frac{1}{2} & 0 \\ 1 & -1 & 1 \\ 0 & \frac{1}{2} & -1 \end{bmatrix}.
\]
(b) If you solve \(\frac{du}{dt} = Au\), is (1) or (2) or (3) true as \(t \to \infty\)?
(1) \(u(t)\) goes to zero?
(2) \(u(t)\) approaches a multiple of (what vector?)
(3) \(u(t)\) blows up?
Solution
(a) The numbers in each column add to zero, so \((1, 1, 1)\) is in the left nullspace \(N(A^T)\); hence \(\mathrm{rank}(A) < 3\) and \[ \lambda_1 = 0. \] Next, \(A + I = \begin{bmatrix} 0 & 1/2 & 0 \\ 1 & 0 & 1 \\ 0 & 1/2 & 0 \end{bmatrix}\) has two equal columns (columns 1 and 3), so \(\det(A + I) = 0\) and \[ \lambda_2 = -1. \] The trace gives the last eigenvalue: \[ \lambda_3 = \mathrm{trace}(A) - \lambda_1 - \lambda_2 = -3 - 0 - (-1) = -2. \] (b) Statement (2): \(u(t)\) approaches a multiple of \[ \begin{bmatrix} 1 \\ 2 \\ 1 \end{bmatrix}. \] Indeed \(A \begin{bmatrix} 1 \\ 2 \\ 1 \end{bmatrix} = 0\), so \((1, 2, 1)\) is the eigenvector for the zero eigenvalue. Since the other eigenvalues \(-1, -2\) are negative, their exponentials \(e^{-t}, e^{-2t}\) decay, and \(u(t) = e^{At}u(0)\) tends to the projection of \(u(0)\) onto the line spanned by \((1, 2, 1)\).Keywords: eigenvalues, differential equation, steady state, trace