0
$\begingroup$

$A$ is a rectangular matrix of size $m\times n$ with real entries, and $\{v_1,v_2,\ldots,v_k\}$ forms a basis of the row space of $A$. Is the set $\{Av_1,Av_2,\ldots,Av_k\}$ linearly independent?

So we have that $\{v_1,\ldots,v_k\}$ is a basis of row space of $A$. We've to check if $\{Av_1,\ldots,Av_k\}$ is a linearly independent set. I started with $$\sum_{i=1}^k\alpha_i Av_i =0 \stackrel{??}{\implies} \alpha_i = 0$$ but couldn't conclude anything about $\alpha_i's$. I also haven't found a counterexample to the result yet, so it may be true.

When $A$ is square and invertible, I could prove the claim.

Any hints to point me in the right direction?

$\endgroup$
1
$\begingroup$

you're working over reals so you have a nice symmetric bilinear form given by the dot product, which is positive definite.

since $\{\mathbf v_1, \mathbf v_2,\ldots,\mathbf v_k\}$ forms a basis, it is a linearly independent set.

Suppose for contradiction
$\{A\mathbf v_1, A\mathbf v_2,\ldots,A\mathbf v_k\}$ is linearly dependent
$\implies \mathbf 0 = \sum_{r=1}^k \alpha_r A\mathbf v_r = A\big(\sum_{r=1}^k \alpha_r \mathbf v_r\big)$
where at least some $\alpha_j\neq 0 \implies \mathbf w :=\big(\sum_{r=1}^k \alpha_r \mathbf v_r\big) \neq \mathbf 0$ since $\mathbf v_r$ are linearly independent.,
$\implies \mathbf w \in \ker A$ but there must exist some x such that $\mathbf x^T A=\mathbf w^T$ giving

$ 0 = \mathbf x^T\mathbf 0= \mathbf x^T \big(A\mathbf w\big) = \big(\mathbf x^T A\big)\mathbf w = \mathbf w^T \mathbf w \gt 0$
which is a contradiction.

(i.e. to further expand the logic $\mathbf w^T\in \text{row space }A \iff \mathbf w\in \text{image }A^T$ which it is since it is written as a linear combination of basis vectors for such space, which means there is at least one vector in pre-image, i.e. some $\mathbf x$ satisfying $A^T\mathbf x = \mathbf w$.)

$\endgroup$

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.