6.2-final-2011-strang¶
Final Exam, 2011 (strang) · Section 6.2: Diagonalizing a Matrix Source: finalexam_2.pdf, p. 52
Problem¶
5. (a) Find the eigenvalues \(\lambda_1, \lambda_2, \lambda_3\) and a full set of independent eigenvectors \(x_1, x_2, x_3\) (if possible) for
\[
A = \begin{bmatrix} 2 & 1 & 1 \\ 1 & 2 & 1 \\ 1 & 1 & 2 \end{bmatrix}.
\]
(b) Suppose \(u_0 = 3x_1 + 7x_2 + 5x_3\) is a combination of your eigenvectors of \(A\): Find \(A^k u_0\). If \(\|v\|\) is the length of \(v\), find the limit of \(\frac{\|A^{k+1}u_0\|}{\|A^k u_0\|}\) as \(k \to \infty\).
Solution
(a) The matrix is \(I + \) (all-ones matrix) \(= I + uu^T\) with \(u = (1, 1, 1)\). Since \(uu^T\) has eigenvalues \(3, 0, 0\), the eigenvalues of \(A\) are \(\lambda_1 = 4, \lambda_2 = 1, \lambda_3 = 1\) (check: trace is \(6 = 4 + 1 + 1\)). Eigenvectors: \[ x_1 = \begin{bmatrix} 1 \\ 1 \\ 1 \end{bmatrix} \ (\lambda = 4), \qquad x_2 = \begin{bmatrix} -1 \\ 1 \\ 0 \end{bmatrix}, \quad x_3 = \begin{bmatrix} -1 \\ 0 \\ 1 \end{bmatrix} \ (\lambda = 1, \lambda = 1). \] The two vectors for \(\lambda = 1\) span the plane \(x_1 + x_2 + x_3 = 0\), so a full set of independent eigenvectors exists. (b) Since \(Ax_i = \lambda_i x_i\), \[ A^k u_0 = 3 \cdot 4^k x_1 + 7 \cdot 1^k x_2 + 5 \cdot 1^k x_3 = 3 \cdot 4^k x_1 + 7x_2 + 5x_3. \] The term \(3 \cdot 4^k x_1\) dominates as \(k \to \infty\), so the ratio of lengths has the limit \[ \lim_{k \to \infty} \frac{\|A^{k+1}u_0\|}{\|A^k u_0\|} = 4. \]Keywords: eigenvalues, eigenvectors, repeated eigenvalue, powers of A