4
$\begingroup$

How do I prove this conjecture?

Let $A$ be a matrix, and $B$ be a column vectore. If $Ax=B$ has two solutions, then there must be a third one.

Thanks in a advance!

$\endgroup$
2
  • $\begingroup$ One of the theorems you already know about the number of solutions to a system of equations can be applied.... $\endgroup$
    – user14972
    Apr 13, 2013 at 14:44
  • 2
    $\begingroup$ Given any two vectors $x,y$ in a vector space over an infinite field $F$, their span $\mbox{Span}\;\{x,y\}=Fx+Fy$ is infinite as soon as one of them is nonzero. $\endgroup$
    – Julien
    Apr 13, 2013 at 15:04

3 Answers 3

11
$\begingroup$

Hint: How could you linearly combine two solutions to make a third?

$\endgroup$
4
$\begingroup$

Suppose $x_1$ and $x_2$ are two (different) solutions, namely $Ax_1=B,Ax_2=B$. Then $A(x_2-x_1)=0$. Let $r$ be the nullity of $A$ and $y_1,\cdots,y_r$ be linearly independent solutions of $Ay=0$. Then $$ x=x_1+c_1y_1+\cdots+c_ry_r $$ is the general solution of $Ax=B$. Thus there are infinite many solutions (of course, there is a third one).

$\endgroup$
2
  • $\begingroup$ Thanks! I have a lot to work on in this course. Thanks again to all! $\endgroup$
    – ohad
    Apr 13, 2013 at 15:13
  • 1
    $\begingroup$ @ohad: It's worth noting that we don't require linearly independent solutions to prove this. $\endgroup$ Apr 13, 2013 at 15:33
3
$\begingroup$

You can do even better than that, assuming your field is infinite. Suppose that $Ax_1=b$ and $Ax_2=b$, where $x_1\ne x_2$. Now, for any $t$ in your scalar field, define $y_t:=tx_1+(1-t)x_2$. Now, $y_1=x_1$ and $y_0=x_2$, and generally, for $s\ne t$ we have $y_s\ne y_t$ (why?). What is $Ay_t$? In this way, we see that we have infinitely many solutions, so long as the scalar field is infinite.

$\endgroup$
8
  • $\begingroup$ Sorry... I feel dumb on this one. How is, for instance, $Ay_3$ a solution? $\endgroup$
    – ohad
    Apr 13, 2013 at 15:42
  • $\begingroup$ Don't forget that $A(x+y)=Ax+Ay$ for any (appropriately-sized) column vectors $x,y,$ and that $A(cx)=c(Ax)$ for any scalar $c$ and any (appropriately-sized) column vector $x$. $\endgroup$ Apr 13, 2013 at 16:09
  • $\begingroup$ In your particular example, $A(3x_1)=3b$ and $A(-2x_2)=-2b$, whence $Ay_3=3b+-2b=b.$ $\endgroup$ Apr 13, 2013 at 16:24
  • $\begingroup$ Thank you. I just fail to understand why this happens really. I also fail using the @ function. Thanks again!!! $\endgroup$
    – ohad
    Apr 13, 2013 at 17:02
  • $\begingroup$ When you comment on a question or answer, the original poster is notified automatically, so there's no need to use the @ in that case. As for why matrix multiplication distributes over vector addition, and why we can pull scalars out front, that's just a nice consequence of the way matrix multiplication is defined. For example, suppose $A$ is an $m\times n$ matrix with $a_{ij}$ indicating the entry of $A$ in the $i$th row, $j$th column. Let $x$ and $y$ be $n$-dimensional column vectors, with respective entries $x_1,\dots,x_n$ and $y_1,\dots,y_n$. Let $c$ be a scalar. (cont'd). $\endgroup$ Apr 13, 2013 at 18:39

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.