6.5-exam3-2002¶
Quiz 3, 2002 · Section 6.5: Solving Linear Differential Equations Source: exam3_1.pdf, pp. 68, 74
Problem¶
2 (20 pts.) (a) If \(A\) is similar to \(B\) show that \(e^A\) is similar to \(e^B\). First define "similar" and \(e^A\)!!
(b) If \(A\) has 3 eigenvalues \(\lambda = 0, 2, 4\), find the eigenvalues of \(e^A\).
Using part (a) explain this connection with determinants:
\[
\text{determinant of } e^A = e^{\text{trace of } A}.
\]
Solution
(a) \(B\) is similar to \(A\) when \(B = M^{-1}AM\), with \(M\) invertible. The exponential of \(A\) is \[ e^A = I + A + \frac{1}{2}A^2 + \frac{1}{6}A^3 + \dots. \] Every power \(B^k\) of \(B\) is similar to the same power \(A^k\) of \(A\): \[ B^k = M^{-1}AM M^{-1}AM \dots M^{-1}AM = M^{-1}A^k M. \] Then \[ e^B = I + B + \frac{1}{2}B^2 + \dots = M^{-1} \left( I + A + \frac{1}{2}A^2 + \dots \right) M = M^{-1} e^A M. \] It is also OK to show this using \(e^A = Se^{\Lambda}S^{-1}\), although that assumes that the matrices are diagonalizable. (b) The exponential of \(A\) is \[ e^A = Se^{\Lambda}S^{-1} = S \begin{bmatrix} e^0 & 0 & 0 \\ 0 & e^2 & 0 \\ 0 & 0 & e^4 \end{bmatrix} S^{-1}. \] But this is an eigenvalue decomposition of \(e^A\), so the eigenvalues are \(1, e^2, e^4\). More generally, the eigenvalues of \(e^A\) are the exponentials of the eigenvalues of \(A\), and \[ \det(e^A) = e^{\lambda_1}e^{\lambda_2} \dots e^{\lambda_n} = e^{\lambda_1 + \lambda_2 + \dots + \lambda_n} = e^{tr(A)}. \]Keywords: similar matrices, matrix exponential, eigenvalues, determinant