Skip to content

6.3-exam3-2009-b

Quiz 3, 2009 · Section 6.3: Symmetric Positive Definite Matrices Source: exam3_1.pdf, pp. 153-154

Problem

Problem 3. Let \(A = \begin{bmatrix} 1 & 1 & 1 \\ 1 & 2 & 0 \\ 1 & 0 & 2 \end{bmatrix}\).

(A) Indicate which of the following statements are true and which are false:

(1) \(A\) is symmetric; (2) \(A\) is orthogonal;

(3) \(A\) is invertible; (4) \(\frac{1}{3} A\) is a Markov matrix

(B) Find the eigenvalues and the eigenvectors of \(A\). (Hint: Part (A) might help you.)

(C) Find an orthogonal matrix \(Q\) and a diagonal matrix \(\Lambda\) such that \(A = Q\Lambda Q^T\).

(D) Calculate the limit \(u_\infty\) of \(u_k = (\frac{1}{3} A)^k u_0\) as \(k \to \infty\), for \(u_0 = \begin{bmatrix} 1 \\ 0 \\ 0 \end{bmatrix}\).

Solution (A) (1) Yes, \(A\) is symmetric. (2) No: \(\begin{bmatrix} 1 \\ 1 \\ 1 \end{bmatrix} \cdot \begin{bmatrix} 1 \\ 2 \\ 0 \end{bmatrix} \neq 0\), and the columns aren't unit length. (3) No: \(2 \cdot row 1 = row 2 + row 3\). (4) Yes! The columns all add to 3. (B) \(A\) is singular, so \(0\) is an eigenvalue. \(\frac{1}{3} A\) is Markov \(\implies 1\) is an eigenvalue of \(\frac{1}{3} A \implies 3\) is an eigenvalue of \(A\). And \(tr(A) = 1 + 2 + 2 = 3 + 0 + \lambda \implies \lambda = 2\) is the other eigenvalue. So the eigenvalues are \(0, 2, 3\), with eigenvectors \[ \lambda = 0: \begin{bmatrix} 2 \\ -1 \\ -1 \end{bmatrix}, \qquad \lambda = 2: \begin{bmatrix} 0 \\ -1 \\ 1 \end{bmatrix}, \qquad \lambda = 3: \begin{bmatrix} 1 \\ 1 \\ 1 \end{bmatrix} \] (the last since the row sums are all 3). (Note: the printed solutions list \((-1, 1, 1)\) for \(\lambda = 2\), but \(A(-1, 1, 1) = (1, 1, 1)\), so the correct eigenvector is \((0, -1, 1)\).) (C) Since \(A\) is symmetric, the eigenvectors are orthogonal; we must normalize them: \[ Q = \begin{bmatrix} \frac{2}{\sqrt{6}} & 0 & \frac{1}{\sqrt{3}} \\ -\frac{1}{\sqrt{6}} & -\frac{1}{\sqrt{2}} & \frac{1}{\sqrt{3}} \\ -\frac{1}{\sqrt{6}} & \frac{1}{\sqrt{2}} & \frac{1}{\sqrt{3}} \end{bmatrix}, \qquad \Lambda = \begin{bmatrix} 0 & & \\ & 2 & \\ & & 3 \end{bmatrix}. \] (D) Since \(\frac{1}{3} A = Q \frac{1}{3}\Lambda Q^T\), \[ \left(\frac{1}{3}A\right)^k = Q \begin{bmatrix} 0 & & \\ & \frac{2}{3} & \\ & & 1 \end{bmatrix}^k Q^T \longrightarrow_{k \to \infty} Q \begin{bmatrix} 0 & & \\ & 0 & \\ & & 1 \end{bmatrix} Q^T. \] Therefore \[ u_\infty = Q \begin{bmatrix} 0 & & \\ & 0 & \\ & & 1 \end{bmatrix} Q^T \begin{bmatrix} 1 \\ 0 \\ 0 \end{bmatrix}. \] Now \(Q^T u_0 = \begin{bmatrix} \frac{2}{\sqrt{6}} \\ 0 \\ \frac{1}{\sqrt{3}} \end{bmatrix}\), so \[ u_\infty = Q \begin{bmatrix} 0 \\ 0 \\ \frac{1}{\sqrt{3}} \end{bmatrix} = \frac{1}{\sqrt{3}} \begin{bmatrix} 1 \\ 1 \\ 1 \end{bmatrix} = \begin{bmatrix} \frac{1}{3} \\ \frac{1}{3} \\ \frac{1}{3} \end{bmatrix}. \] (The printed solution has a sign error here and gives \((-1/3, -1/3, -1/3)\); since \(\frac{1}{3}A\) is Markov, the component sum is preserved, and \(\sum u_\infty = \sum u_0 = 1\), so the correct limit is \((1/3, 1/3, 1/3)\).)

Keywords: Markov matrix, eigenvalues, eigenvectors, orthogonal diagonalization, matrix powers