Skip to content

4.4-exam2-2008-edelman

Quiz 2, 2008 (edelman) · Section 4.4: Orthonormal Bases and Gram-Schmidt Source: exam2_1.pdf, pp. 173, 157

Problem

2 (30 pts.)

(a) \(P\) is the projection matrix onto the column space of \(A\) which has independent columns. \(Q\) is a square orthogonal matrix with the same number of rows as \(A\). In simplest form, in terms of \(P\) and \(Q\), what is the projection matrix onto the column space of \(QA\)?

(b) The vectors \(a, b\), and \(c\) are independent. The matrix \(P\) is the projection matrix onto the span of \(a\) and \(b\). Suppose we apply Gram-Schmidt onto the vectors \(a, b\), and \(c\) producing orthonormal vectors \(q_1, q_2\), and \(q_3\). Write the unit vector \(q_3\) in simplest form in terms of \(P\) and \(c\) only.

(c) The vectors \(a, b\), and \(c\) are independent. The matrix \(A = [a \; b \; c]\) has these three vectors as its columns. The QR decomposition writes \(A = QR\) where \(Q\) is orthogonal and \(R\) is \(3 \times 3\) upper triangular. Write \(\|c\|\) in terms of only the elements of \(R\) in simplest form.

Solution a) The projection matrix of a matrix \(A\) is \(P = A(A^TA)^{-1}A^T\). So the projection matrix of \(QA\) is \[ (QA)(A^TQ^TQA)^{-1}A^TQ^T = QA(A^TA)^{-1}A^TQ^T = QPQ^T, \] where we have used that \(Q^TQ = I\). b) By definition \(c - Pc\) is orthogonal to the space \(\operatorname{span}\{a, b\} = \operatorname{span}\{q_1, q_2\}\). So we can choose \[ q_3 = \frac{c - Pc}{\|c - Pc\|}. \] c) Let \(s_1, s_2, s_3\) denote the rows of \(Q\), and \(r_1, r_2, r_3\) denote the columns of \(R\). Then \[ c = \begin{bmatrix} s_1 \cdot r_3 \\ s_2 \cdot r_3 \\ s_3 \cdot r_3 \end{bmatrix} = Qr_3. \] Since orthogonal matrices preserve the lengths of vectors, this implies \(\|c\| = \|r_3\|\). In terms of the elements of \(R\), \(\|c\| = \sqrt{r_{13}^2 + r_{23}^2 + r_{33}^2}\).

Keywords: projection matrix, Gram-Schmidt, QR decomposition, orthogonal matrix