enter image description here

The question above has a typo where it should be i = 1,...,n.

Why is the matrix A invertible? I understand how to calculate it's inverse by multiplying both sides by A's inverse, but don't understand why A is invertible to begin with

My initial thoughts are that Av_i equaling the standard unit vectors somehow implies that the columns are linearly independent, but I don't really know how to put it into words. Any help would be great!


I'll give two approaches: the first is a little more intuitive than the second, which is clearer to beginners.

To see that $A$ is invertible, we first guess that it is invertible. (This is not mathematically incorrect unless we arrive at a contradiction). Then we define a transformation $B$, given by $Be_i = v_i$, where $v_i$ are the vectors which exist so that $Av_i = e_i$. Then, extend linearly: if $x = \sum x_ie_i$, then $Bx = \sum x_iv_i$.

Now, you can check that $AB= I$, since $(AB)e_i = (A)(Be_i) = Av_i= e_i$ for all $i = 1,2, \ldots , n$.

Now, (this is not true for infinite dimensions!) since we are in a finite dimensional space, $AB = I$ implies $BA = I$, hence $A$ is invertible, with inverse $B$.

There's another approach, but they all boil down to the same thing: We will show that $A$ is surjective. Let $x$ be a vector, then $x = \sum x_ie_i$ for some scalars $x_i$, and hence we can see that $A(\sum x_iv_i) = \sum x_i A(v_i) = \sum x_ie_i = x$. So $A$ is surjective, hence injective (rank-nullity theorem), hence is invertible.

Note that the converse is also true : if $A$ is invertible, then it is surjective, so every vector, let alone unit vectors, will have a pre-image.



Remember that a matrix is invertible if and only if its kernel has dimension zero and that the dimension of the kernel is $n$ minus the dimension of the range. So if you can prove the range has dimension $n$ you are done.


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.