Skip to content

4.4-final-2003

Final Exam, 2003 · Section 4.4: Orthonormal Bases and Gram-Schmidt Source: finalexam_1.pdf, pp. 97-98

Problem

2. Let \(F\) be the subspace of \(R^4\) given by

\[ F = \{(x, y, z, w) : x - y + 2z + 3w = 0\}. \]

Let \(P\) be the projection matrix for projecting onto \(F\). (Many of the subquestions can be answered independently of the others.)

(a) Give an orthonormal basis \(\{v_1, \dots, v_k\}\) for the orthogonal complement to \(F\).

(b) Find an orthonormal basis \(\{w_1, \dots, w_l\}\) for \(F\). Explain how you proceed.

(c) What are the eigenvalues of \(P\)? Give them with their multiplicities.

(d) What is the projection of \(\begin{bmatrix} 1 \\ 1 \\ 1 \\ 1 \end{bmatrix}\) onto \(F\)?

Solution (a) \(F\) is the nullspace of the row vector \((1, -1, 2, 3)\), so \(F^\perp\) is the line spanned by the normal vector: \(v_1 = \frac{1}{\sqrt{15}}(1, -1, 2, 3)\). (\(k = 1\).) (b) \(F\) has dimension 3. A basis of \(F\) is \((1, 1, 0, 0)\), \((-2, 0, 1, 0)\), \((-3, 0, 0, 1)\) (free variables \(y, z, w\), with \(x = y - 2z - 3w\)). Apply Gram-Schmidt: \[ w_1 = \frac{1}{\sqrt{2}}\begin{bmatrix} 1 \\ 1 \\ 0 \\ 0 \end{bmatrix}, \] \[ (-2, 0, 1, 0) - (( -2, 0, 1, 0) \cdot w_1)\, w_1 = (-2, 0, 1, 0) + (1, 1, 0, 0) = (-1, 1, 1, 0), \] \[ w_2 = \frac{1}{\sqrt{3}}\begin{bmatrix} -1 \\ 1 \\ 1 \\ 0 \end{bmatrix}, \] \[ (-3, 0, 0, 1) - ((-3, 0, 0, 1) \cdot w_1)w_1 - ((-3, 0, 0, 1) \cdot w_2)w_2 = \left(-\frac{1}{2}, \frac{1}{2}, -1, 1\right) \sim (-1, 1, -2, 2), \] \[ w_3 = \frac{1}{\sqrt{10}}\begin{bmatrix} -1 \\ 1 \\ -2 \\ 2 \end{bmatrix}. \] All three lie in \(F\) (each satisfies \(x - y + 2z + 3w = 0\)) and are orthonormal, so \(\{w_1, w_2, w_3\}\) is an orthonormal basis of \(F\). (c) \(P\) projects onto the 3-dimensional subspace \(F\), so its eigenvalues are 1 (multiplicity 3) and 0 (multiplicity 1): every vector in \(F\) is fixed and every vector in \(F^\perp\) is sent to zero. (d) \(b = (1, 1, 1, 1)\). Subtract its projection onto the normal direction \(n = (1, -1, 2, 3)\): \(b \cdot n = 5\), \(n \cdot n = 15\), so \[ Pb = b - \frac{5}{15}n = \begin{bmatrix} 1 \\ 1 \\ 1 \\ 1 \end{bmatrix} - \frac{1}{3}\begin{bmatrix} 1 \\ -1 \\ 2 \\ 3 \end{bmatrix} = \frac{1}{3}\begin{bmatrix} 2 \\ 4 \\ 1 \\ 0 \end{bmatrix}. \] (Check: \((2 - 4 + 2 + 0)/3 = 0\), so \(Pb \in F\), and \(b - Pb\) is a multiple of the normal vector.)

Keywords: orthogonal complement, Gram-Schmidt, projection matrix, eigenvalues