3.5-final-2008-b¶
Final Exam, 2008 · Section 3.5: Dimensions of the Four Subspaces Source: finalexam_1.pdf, p. 203
Problem¶
9 (20 pts.) Suppose an \(n\) by \(n\) matrix has the property that its nullspace is equal to its column space.
(a) Can the matrix be the zero matrix?
(b) Possibly in terms of \(n\), what is the rank of the matrix?
(c) What are the eigenvalues of this matrix? (Briefly explain your answer. Hint: It might be useful to consider applying \(A\) more than once in some way.)
(d) Give an example of a 2 by 2 such matrix.
(e) Perhaps using the previous case twice somehow, give an example of a 4 x 4 such matrix.
Solution
(a) No. For the zero matrix the nullspace is all of \(R^n\), while the column space is \(\{0\}\); these are equal only in the degenerate case \(n = 0\). (b) The rank is \(n/2\). Since \(\dim N(A) = n - r\) and \(\dim C(A) = r\), equality of the two subspaces forces \(n - r = r\), so \(r = n/2\). In particular such a matrix exists only when \(n\) is even. (c) All eigenvalues are 0. Since \(C(A) = N(A)\), we have \(Ax \in N(A)\) for every \(x\), so \(A(Ax) = 0\) for every \(x\): the matrix satisfies \(A^2 = 0\), hence \(A\) is nilpotent and all its eigenvalues are 0. (Equivalently, an eigenvector \(x\) of \(A\) lies in \(C(A) = N(A)\), so \(Ax = 0\) and the eigenvalue is 0.) (d) \(A = \begin{bmatrix} 0 & 1 \\ 0 & 0 \end{bmatrix}\): the column space is spanned by \((1, 0)^T\) (the first column is zero, the second is \((1, 0)^T\)), and the nullspace consists of all \((x_1, x_2)\) with \(x_2 = 0\), i.e., the span of \((1, 0)^T\). So \(N(A) = C(A)\). (e) Take the block diagonal matrix with two copies of the 2 by 2 example: \[ A = \begin{bmatrix} 0 & 1 & 0 & 0 \\ 0 & 0 & 0 & 0 \\ 0 & 0 & 0 & 1 \\ 0 & 0 & 0 & 0 \end{bmatrix}. \] Its column space is spanned by \((1, 0, 0, 0)^T\) and \((0, 0, 1, 0)^T\), and \(Ax = (x_2, 0, x_4, 0)^T = 0\) exactly when \(x_2 = x_4 = 0\), so the nullspace is the same span.Keywords: nullspace, column space, rank, nilpotent