Skip to content

3.4-final-2010

Final Exam, 2010 · Section 3.4: Independence, Basis, and Dimension Source: finalexam_2.pdf, pp. 22-24

Problem

Bonus problem (From an earlier version of the exam.)

A basis for the nullspace of the matrix \(A\) consists of the three vectors:

\[ \begin{bmatrix} 1 \\ -2 \\ 1 \\ 0 \\ 0 \end{bmatrix}, \quad \begin{bmatrix} 3 \\ -2 \\ 0 \\ 1 \\ 0 \end{bmatrix}, \quad \begin{bmatrix} 6 \\ 5 \\ 0 \\ 0 \\ 1 \end{bmatrix}. \]

A basis for the column space of the matrix \(A\) consists of the two vectors:

\[ \begin{bmatrix} 1 \\ -1 \\ -1 \end{bmatrix}, \quad \begin{bmatrix} 2 \\ -1 \\ 0 \end{bmatrix}. \]

(a) (2 pts.) How many rows and how many columns are in \(A\)?

(b) (6 pts.) Provide a basis for \(N(A^T)\) and \(C(A^T)\).

(c) (6 pts.) Write down an example of such a matrix \(A\).

Solution (a) If \(A\) is an \(m \times n\) matrix, then the nullspace \(N(A)\) is a subspace of \(R^n\), while the column space \(C(A)\) is a subspace of \(R^m\). In our case, \(\dim N(A) = 3\) tells us \(n = 5\) and \(\dim C(A) = 2\) tells us \(m = 3\). So \(A\) is a \(3 \times 5\) matrix: it has 3 rows and 5 columns. (b) The left-nullspace \(N(A^T)\) is the orthogonal complement of \(C(A)\) in \(R^3\). Take the cross product of the two vectors in the basis for \(C(A)\): \[ (1, -1, 1) \times (2, -1, 0) = (1, 2, 1). \] The vector \((1, 2, 1)\) is perpendicular to the plane \(C(A)\), so it is a basis for the line \(N(A^T)\). If you didn't think of the cross product, solve \[ \begin{bmatrix} 1 & -1 & 1 \\ 2 & -1 & 0 \end{bmatrix} \begin{bmatrix} x_1 \\ x_2 \\ x_3 \end{bmatrix} = \begin{bmatrix} 0 \\ 0 \end{bmatrix} \] by row reduction: \[ \begin{bmatrix} 1 & -1 & 1 \\ 2 & -1 & 0 \end{bmatrix} \to \begin{bmatrix} 1 & -1 & 1 \\ 0 & 1 & -2 \end{bmatrix} \to \begin{bmatrix} 1 & 0 & -1 \\ 0 & 1 & -2 \end{bmatrix}; \] with \(x_3\) free, \(x = x_3 (1, 2, 1)\), giving the same special solution \((1, 2, 1)\), a basis for \(N(A^T)\). Similarly, the row space \(C(A^T)\) is the orthogonal complement of \(N(A)\), so it consists of the solutions \((x_1, \dots, x_5)\) to \[ \begin{bmatrix} 1 & -2 & 1 & 0 & 0 \\ 3 & -2 & 0 & 1 & 0 \\ 6 & 5 & 0 & 0 & 1 \end{bmatrix} x = 0. \] If you reverse the order of the columns, the \(3 \times 5\) matrix is already in reduced row-echelon form, so the special solutions can be found by setting \((x_1, x_2)\) to \((1, 0)\) or \((0, 1)\) and solving for \((x_3, x_4, x_5)\). In this way we get the special solutions \((1, 0, -1, -3, -6)\) and \((0, 1, 2, 2, -5)\); these two vectors form a basis for \(C(A^T)\). (c) From part (b), \(A\) has the same column space as \[ \begin{bmatrix} 1 & 0 \\ 0 & 1 \\ -1 & -2 \end{bmatrix} \] (you could also just say \(\begin{bmatrix} 1 & 2 \\ -1 & -1 \\ 1 & 0 \end{bmatrix}\); the method above makes the subsequent computations a bit easier), and the same row space as \[ \begin{bmatrix} 1 & 0 & -1 & -3 & -6 \\ 0 & 1 & 2 & 2 & -5 \end{bmatrix}. \] So \(A\) is any matrix of the form \[ \begin{bmatrix} 1 & 0 \\ 0 & 1 \\ -1 & -2 \end{bmatrix} B \begin{bmatrix} 1 & 0 & -1 & -3 & -6 \\ 0 & 1 & 2 & 2 & -5 \end{bmatrix}, \] where \(B\) is an invertible \(2 \times 2\) matrix (if you don't see why, please ask). For example, if we take \(B\) to be the identity, we get \[ A = \begin{bmatrix} 1 & 0 & -1 & -3 & -6 \\ 0 & 1 & 2 & 2 & -5 \\ -1 & -2 & -3 & -1 & 16 \end{bmatrix}. \] *Note on the source:* the printed solution computes the cross product \((1, -1, 1) \times (2, -1, 0)\), while the printed problem lists the first column-space basis vector as \((1, -1, -1)\); the entire printed solution is consistent with \((1, -1, 1)\). With the vector \((1, -1, -1)\) as printed, the cross product would be \((1, -1, -1) \times (2, -1, 0) = (-1, -2, 1)\), i.e. the line spanned by \((1, 2, -1)\).

Keywords: nullspace, column space, left nullspace, row space, basis