3.5-exam1-2012-strang¶
Quiz 1, 2012 (strang) · Section 3.5: Dimensions of the Four Subspaces Source: exam1_2.pdf, p. 29
Problem¶
2 (16 pts.) Suppose the matrix \(A\) is \(m\) by \(n\) of rank \(r\), and the matrix \(B\) is \(M\) by \(N\) of rank \(R\). Suppose the column space \(C(A)\) is contained in (possibly equal to) the column space \(C(B)\). (This means that every vector in \(C(A)\) is also in \(C(B)\).) What relations must hold between \(m\) and \(M\), \(n\) and \(N\), and \(r\) and \(R\)? It might be good to write down an example of \(A\) and \(B\) where all the columns are different.
Solution
The column space of \(A\) is contained in \(R^m\), and the column space of \(B\) is contained in \(R^M\). If \(C(A) \subseteq C(B)\), this means they are contained in the same Euclidean space, so \(M = m\). The dimension of the column space is the rank of the matrix, so if \(C(A) \subseteq C(B)\), this means \(\dim C(A) \leq \dim C(B)\), hence \(r \leq R\). There are no relations between \(n\) and \(N\): \(n = N\) if \(A = B\), \(n \leq N\) if \(B = [A \ A]\), and \(n \geq N\) if \(A = [B \ B]\).Keywords: column space, rank, containment, dimensions