1
$\begingroup$

I'm trying to solve the following equation:

$$ AX=B $$

where $A=\begin{bmatrix}1 & 5\\2 & 3\\-4 & 1\end{bmatrix}$, $B=\begin{bmatrix}1 & 2 & 5\\2 & 4 & 3\\-4 & -8 & -1\end{bmatrix}$

Usually I multiply both sides of the equation by $A^{-1}$ which results in $X=BA^{-1}$

However, matrix $A$ is not a square matrix so I cannot get its inverse. What I tried to do is multiply both sides of the equation by $B^{-1}$ and I get $AB^{-1}X=1$. But again I have no idea how to solve this.

Any tips on how to do this appreciated. Thx

$\endgroup$
2
$\begingroup$

Let $$X = \left[\begin{array}{ccc}a&b&c\\d&e&f\end{array}\right] $$

Then it should hold

$$\begin{array}{lrcl} (1) & c+5f &=& 5 \\ (2) & 2c+3f&=&3 \\ (3) &-4c+f &=& -1 \end{array}$$

Then $(3)+2\cdot(2)$ reveals $$f = \frac{5}{7}$$ but $(3)+4\cdot(1)$ reveals that $$f=\frac{19}{21}$$ a contradiction. Thus you system has no solution.

$\endgroup$
  • $\begingroup$ thx. In your answer I believe (1) should be $a+5d = 1$ . Wondering if it's the simplest method to solve this eq, cause solving this would need solving 3 equation systems if the eq was solvable? $\endgroup$ – user45430 Jan 12 '14 at 15:06
  • $\begingroup$ You are right, there has been an error in my calculation. I made an edit to my answer. $\endgroup$ – user127.0.0.1 Jan 13 '14 at 14:35
0
$\begingroup$

You should get $X=A^{-1}B$ when you do what you say you do. But to answer your question, if $A'A$ is invertible then $X=(A'A)^{-1}A'B$ is the value of $X$ that minimizes the norm of $AX-B$ so if there is a solution then $X$ would be it. More generally, read up on Moore-Penrose inverses.

$\endgroup$

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.