6.2-exam3-2002-b¶
Quiz 3, 2002 · Section 6.2: Diagonalizing a Matrix Source: exam3_1.pdf, pp. 72, 75
Problem¶
4 (22 pts.) Multinational companies in the US, Asia, and Europe have assets of 12 trillion dollars. At the start, 6 trillion are in the US, 6 trillion in Europe. Each year half the US money stays home, \(^1/_4\) each goes to Asia and Europe. For Asia and Europe, half stays home and half is sent to the US.
\[
\begin{bmatrix} US \\ Asia \\ Europe \end{bmatrix}_{year\, k+1} = \begin{bmatrix} .5 & .5 & .5 \\ .25 & .5 & 0 \\ .25 & 0 & .5 \end{bmatrix} \begin{bmatrix} US \\ Asia \\ Europe \end{bmatrix}_{year\, k}
\]
(a) The eigenvalues and eigenvectors of this singular matrix \(A\) are
(b) The limiting distribution of the 12 trillion dollars as the world ends is
US =
Asia =
Europe =
Solution
(a) \(A\) is singular, so one eigenvalue is 0. It is also a Markov matrix, so another eigenvalue is 1 (motivation: each column of \(A\) sums to 1, so each column of \(A - I\) sums to 0; \(A - I\) then has an eigenvalue 0, and \(A\) has an eigenvalue 1). The last eigenvalue is 0.5 since \(trace(A) = \sum_i \lambda_i = 1.5\). The eigenvectors are found by solving the following systems: \[ \lambda_1 = 1 : (A - \lambda_1 I)x_1 = \begin{bmatrix} -.5 & .5 & .5 \\ .25 & -.5 & 0 \\ .25 & 0 & -.5 \end{bmatrix} x_1 = 0 \implies x_1 = (2, 1, 1), \] \[ \lambda_2 = 0.5 : (A - \lambda_2 I)x_2 = \begin{bmatrix} 0 & .5 & .5 \\ .25 & 0 & 0 \\ .25 & 0 & 0 \end{bmatrix} x_2 = 0 \implies x_2 = (0, 1, -1), \] \[ \lambda_3 = 0 : (A - \lambda_3 I)x_3 = \begin{bmatrix} .5 & .5 & .5 \\ .25 & .5 & 0 \\ .25 & 0 & .5 \end{bmatrix} x_3 = 0 \implies x_3 = (2, -1, -1). \] (b) Write the initial value as a linear combination of the eigenvectors: \[ u_0 = (6, 0, 6) = 3x_1 - 3x_2. \] The distribution after \(k\) steps is then \[ u_k = A^k u_0 = 3\lambda_1^k x_1 - 3\lambda_2^k x_2 = 3x_1 - 3 \cdot 0.5^k x_2 \to 3x_1 = (6, 3, 3) \quad \text{as } k \to \infty. \]Keywords: Markov matrix, eigenvalues, eigenvectors, steady state