6.4-final-2008-strang¶
Final Exam, 2008 (strang) · Section 6.4: Complex Numbers and Vectors and Matrices Source: finalexam_1.pdf, p. 214
Problem¶
5 (10 pts.) The numbers \(D_n\) satisfy \(D_{n+1} = 2D_n - 2D_{n-1}\). This produces a first-order system for \(u_n = (D_{n+1}, D_n)\) with this 2 by 2 matrix \(A\):
\[
\begin{bmatrix} D_{n+1} \\ D_n \end{bmatrix} = \begin{bmatrix} 2 & -2 \\ 1 & 0 \end{bmatrix} \begin{bmatrix} D_n \\ D_{n-1} \end{bmatrix} \quad \text{or} \quad u_n = Au_{n-1}.
\]
(a) Find the eigenvalues \(\lambda_1, \lambda_2\) of \(A\). Find the eigenvectors \(x_1, x_2\) with second entry equal to 1 so that \(x_1 = (z_1, 1)\) and \(x_2 = (z_2, 1)\).
(b) What is the inner product of those eigenvectors? (2 points)
(c) If \(u_0 = c_1x_1 + c_2x_2\), give a formula for \(u_n\). For the specific \(u_0 = (2, 2)\) find \(c_1\) and \(c_2\) and a formula for \(D_n\).
Solution
(a) The eigenvalues of \[ A = \begin{bmatrix} 2 & -2 \\ 1 & 0 \end{bmatrix} \] satisfy the equation \(\lambda^2 - 2\lambda + 2 = 0\), so \(\lambda_1 = 1 + i\) and \(\lambda_2 = 1 - i\). We find the eigenvectors by taking the appropriate nullspaces: \[ A - \lambda_1I = \begin{bmatrix} 1 - i & -2 \\ 1 & -1 - i \end{bmatrix} \] has nullspace generated by \(x_1 = (1 + i, 1)\), and \[ A - \lambda_2I = \begin{bmatrix} 1 + i & -2 \\ 1 & -1 + i \end{bmatrix} \] has nullspace generated by \(x_2 = (1 - i, 1)\). If you pick a different vector in the nullspace, you just rescale so that the bottom entry is 1. (b) The inner product is \(x_1^Hx_2 = (1 - i)^2 + 1 = 1 - 2i\), or its conjugate expression \(x_2^Hx_1 = 1 + 2i\). (c) If \(u_0 = c_1x_1 + c_2x_2\), then \(u_n = c_1\lambda_1^n x_1 + c_2\lambda_2^n x_2\). A matrix always acts on its eigenvectors in a diagonal way. In particular, \((2, 2) = x_1 + x_2\). So we find \[ u_n = (1 + i)^n \begin{bmatrix} 1 + i \\ 1 \end{bmatrix} + (1 - i)^n \begin{bmatrix} 1 - i \\ 1 \end{bmatrix} \] with second entry \[ D_n = (1 + i)^n + (1 - i)^n. \]Keywords: complex eigenvalues, eigenvectors, recurrence, matrix powers