Suppose A and B are two mxn matrices such that rank (A) > rank (B) , then I have to prove that rowspace of B is a subset of rowspace of A.

What I tried was using property of 'Simplified Span Method' , i.e. Finding reduced row echelon form after using vectors as rows of matrix, that the final rref has linearly independent non zero rows. So A will have more linearly independent vectors than B as rank(A) > rank(B). But I do not know how to proceed if these vectors are different i.e.

Suppose n is 6, so maximum possible cardinal number for a subset of mxn matrices is 6 and A contains 3 of these and B contains 2 and the other one is left out. In this case, rowspace of B will not be a subset of rowspace of A.

  • 1
    $\begingroup$ It's not hard to find two $3\times3$ matrices $A,B$ such that the rank of $A$ is 2, the rank of $B$ is 1, but the rowspace of $B$ is not a subset of the rowspace of $A$. $\endgroup$ – Gerry Myerson Feb 19 '17 at 10:27
  • $\begingroup$ Did you mean in your question to ask how to disprove? $\endgroup$ – Maciej Caputa Feb 19 '17 at 10:39
  • $\begingroup$ The question asked to prove or disprove the fact, I completely ignored that. 😅 $\endgroup$ – Quark Feb 19 '17 at 10:41

This statement is not correct and can be disproved by a counter example as you mentioned.

Let $A$ and $B$ be a $3$ by $3$ matrices with $$Rowspace(A) = span\left\{<1,0,0>,<0,1,0>\right\}$$ and $$Rowspace(B) = span\left\{<0,0,1>\right\}$$

Hence $Rank(A) > Rank(B)$, but $Rowspace(B) \not\subset Rowspace(A)$

However, following statement is true.

$$A, B \in M_{m,n}(\mathbb{R}) : rank(A) = m\ \wedge\ rank(A) > rank(B) \implies Rowspace(B) \subseteq Rowspace(A) $$

This can be read as:

Let $A$ and $B$ be m by n matrices with real entries. If rowspace of A spans $\mathbb{R}^m$ ($dim(Rowspace(A)) = m$) and $rank(A) > rank(B)$ then rowspace of B is a proper subset of rowspace of A.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.