3.2-exam1-2014-strang¶
Exam I, 2014 (strang) · Section 3.2: Computing the Nullspace by Elimination: A = CR Source: exam1_2.pdf, pp. 72, 78-79
Problem¶
3 Suppose the nullspace of \(A\) (5 by 4 matrix) is spanned by \(v\) and \(w\), which are special solutions to \(Ax = 0\):
(a) What is the row reduced echelon form \(R = \text{rref}(A)\)? We don't have to know \(A\). (b) Which vectors in \(R^4\) can be rows of \(A\)? How many of the 5 rows of \(A\) will be independent? (c) One point question: What is the dimension of the matrix space containing all 5 by 4 matrices \(A\) that have those vectors \(v\) and \(w\) in their nullspace? (d) If \(C\) is any 4 by 7 matrix of rank \(r = 4\), find the column space of \(C\). Explain clearly why \(Cx = b\) always has infinitely many solutions.
Solution
(a) It is clear that \(x_2\) and \(x_4\) are free variables. We can reconstruct part of \(R\) right away: \[ R = \begin{bmatrix} 1 & a & 0 & b \\ 0 & 0 & 1 & c \\ 0 & 0 & 0 & 0 \\ 0 & 0 & 0 & 0 \\ 0 & 0 & 0 & 0 \end{bmatrix}. \] To find \(a\), \(b\), and \(c\) we use the fact that \(r_1 \cdot v = r_1 \cdot w = r_2 \cdot v = r_2 \cdot w = 0\). We get \(a = -4\), \(b = -1\), \(c = -2\). Thus, \[ R = \begin{bmatrix} 1 & -4 & 0 & -1 \\ 0 & 0 & 1 & -2 \\ 0 & 0 & 0 & 0 \\ 0 & 0 & 0 & 0 \\ 0 & 0 & 0 & 0 \end{bmatrix}. \] (b) The dimension of the nullspace is 2 and is equal to \(n - r\). So \(r = 4 - 2 = 2\). Therefore the dimension of the row space is 2, and exactly two of the five rows of \(A\) are independent. The row space of \(A\) is the same as the row space of \(R\), so any row of \(A\) must be a linear combination of the first two rows of \(R\): \[ c[1, -4, 0, -1] + d[0, 0, 1, -2]. \] (c) Each row can be any vector in the 2-dimensional space orthogonal to \(v\) and \(w\). There are five rows, so the dimension of the space of all such matrices is \(5 \times 2 = 10\). (d) The rank of the matrix is equal to the dimension of the column space, so the column space of \(C\) has dimension 4 and spans all of \(R^4\): \(C(C) = R^4\) (with basis \(e_1, e_2, e_3, e_4\), say). Hence, for any vector \(b\) there exists a solution. In addition, \(\dim N(C) = 7 - 4 = 3\), so each solution carries 3 free parameters: there are infinitely many solutions.Keywords: rref, nullspace, row space, matrix space