6.2-final-2003¶
Final Exam, 2003 · Section 6.2: Diagonalizing a Matrix Source: finalexam_1.pdf, p. 101
Problem¶
4. (a) Give two 3 by 3 matrices \(A\) and \(B\) such that \(AB\) is not equal to \(BA\).
(b) Suppose \(A\) and \(B\) are \(n \times n\) matrices with the same set of linearly independent eigenvectors \(v_1, v_2, \dots, v_n\). However, the eigenvalues might be different: \(v_i\) is the eigenvector for the eigenvalue \(\lambda_i\) of \(A\) and the eigenvector for the eigenvalue \(\mu_i\) of \(B\). Show that \(AB = BA\).
Solution
(a) Many examples work; for instance \[ A = \begin{bmatrix} 1 & 1 & 0 \\ 0 & 1 & 0 \\ 0 & 0 & 1 \end{bmatrix}, \quad B = \begin{bmatrix} 1 & 0 & 0 \\ 0 & 1 & 1 \\ 0 & 0 & 1 \end{bmatrix}. \] Then \(AB = \begin{bmatrix} 1 & 1 & 1 \\ 0 & 1 & 1 \\ 0 & 0 & 1 \end{bmatrix}\) but \(BA = \begin{bmatrix} 1 & 1 & 0 \\ 0 & 1 & 1 \\ 0 & 0 & 1 \end{bmatrix}\), which are different. (Equivalently: \(AB\) has 1 in the \((1, 3)\) position, \(BA\) has 0.) (b) For each \(i\), \(v_i\) is an eigenvector of both matrices: \[ AB\,v_i = A(\mu_i v_i) = \mu_i \lambda_i v_i, \qquad BA\,v_i = B(\lambda_i v_i) = \lambda_i \mu_i v_i, \] so \(ABv_i = BAv_i\) for every \(i\). Since \(v_1, \dots, v_n\) is a basis of \(R^n\), the linear transformations \(AB\) and \(BA\) agree on a basis, hence agree everywhere: \(AB = BA\). (Each of \(AB\) and \(BA\) is diagonal with entries \(\lambda_i\mu_i\) in the \(v\)-basis.)Keywords: matrix multiplication, eigenvectors, commuting matrices, diagonalization