4.4-final-1999-strang¶
Final Exam, 1999 (strang) · Section 4.4: Orthonormal Bases and Gram-Schmidt Source: finalexam_1.pdf, pp. 35, 41
Problem¶
6. (a) (3 points) The vectors \(q_1 = \frac{1}{\sqrt{50}} \begin{bmatrix} 3 \\ 4 \\ 5 \end{bmatrix}\) and \(q_2 = \frac{1}{5} \begin{bmatrix} 4 \\ -3 \\ 0 \end{bmatrix}\) are orthonormal.
Find one more vector to complete an orthonormal basis for \(R^3\).
(b) (3 points) In solving part (a), you might start with a vector like \(a_3 = (0, 0, 1)\) and find \(q_3\). Which vectors \(a_3\) would not work as starting vectors to find \(q_3\) by Gram-Schmidt? How many different real vectors \(q_3\) will give a correct answer to part (a)?
(c) (3 points) Project the vector \(a_3 = (0, 0, 1)\) onto the plane spanned by \(q_1\) and \(q_2\). Find its projection \(p\).
Solution
(a) One way is to solve for \(x\) perpendicular to \(q_1\) and \(q_2\): \[ \begin{bmatrix} 3 & 4 & 5 \\ 4 & -3 & 0 \end{bmatrix} \begin{bmatrix} x_1 \\ x_2 \\ x_3 \end{bmatrix} = \begin{bmatrix} 0 \\ 0 \end{bmatrix} \] Another way is Gram-Schmidt and we might as well start with \(a_3 = (0, 0, 1)\). Then Gram-Schmidt subtracts off projections: \[ a_3 - (a_3^T q_1)q_1 - (a_3^T q_2)q_2 = \begin{bmatrix} 0 \\ 0 \\ 1 \end{bmatrix} - \frac{5}{50} \begin{bmatrix} 3 \\ 4 \\ 5 \end{bmatrix} - 0 = \begin{bmatrix} -0.3 \\ -0.4 \\ 0.5 \end{bmatrix} \] Normalizing to a unit vector gives \[ q_3 = \frac{1}{\sqrt{50}} \begin{bmatrix} -3 \\ -4 \\ 5 \end{bmatrix} \] (b) \(a_3\) will not work if it is in the plane of \(q_1\) and \(q_2\). The only possible vectors \(q_3\) are + (our \(q_3\)) and - (our \(q_3\)). (c) The projection is the vector that was subtracted off in part (a): \[ p = \frac{5}{50} \begin{bmatrix} 3 \\ 4 \\ 5 \end{bmatrix} = \begin{bmatrix} 0.3 \\ 0.4 \\ 0.5 \end{bmatrix} \]Keywords: Gram-Schmidt, orthonormal basis, projection