Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

Let $A$ be an $n\times m$ matrix with rank $n$. Then I would like to conclude that there is a matrix $B$ such that $AB=I_n$.

This is easy, on the intuitive level. The image of $A$ in $\mathbb{R}^m$ is an $n$-dimensional subspace, spanned by the pivotal columns of $A$. Taking $B$ to be the $m\times n$ matrix that sends each pivotal column to the standard basis in $\mathbb{R}^n$ should yield the answer.

I feel like my solution lacks rigor. Any tips for how to make this more precise?

share|cite|improve this question
up vote 1 down vote accepted

Let $A:\mathbb{R}^m\to\mathbb{R}^n$ be rank $n$. Then, you know that $A(\mathbb{R}^m)=\mathbb{R}^n$. Pick a basis for $\mathbb{R}^n$, say $\{b_1,\cdots,b_n\}$ and choose $a_i\in A^{-1}(b_i)$. Then, define $B(b_i)=a_i$, and extend by linearity. When you make the matrix for $B$ you will have your right inverse!

share|cite|improve this answer

We can use one of the handy formulations of matrix multiplication: if $A,B$ are matrices and $B$ can be written as a matrix of columns $[b_1|b_2|\ldots|b_k]$, then $AB=[Ab_1|Ab_2|\ldots|Ab_k]$. Now since we want $AB=I$, we would like $Ab_i=e_i$, $i=1,...,k$, $e_i$ the $i$-th standard basis vector. So the problem reduces to finding these vectors $b_i$. Can you do this?

share|cite|improve this answer

Let $e_k$ be the vector with zeros except for a one at the $k$th position. I will abuse notation slightly by using $e_k$ regardless of the dimension $n$ of the space it is in (as long as $k\leq n$).

Since $\text{rk} A = n$, we can find $i_1,...,i_n$ such that $\{A e_{i_1},...,A e_{i_n}\}$ are linearly independent. If we define $\Pi : \mathbb{R}^n \to \mathbb{R}^m$ by $\Pi = \sum_{k=1}^n e_{i_k} e_k^T$, then we have $A \Pi = \begin{bmatrix} A e_{i_1} & \cdots & A e_{i_n}\end{bmatrix}$, which is invertible. Hence we have $A \Pi (A \Pi)^{-1} = I$. Letting $B=\Pi (A \Pi)^{-1}$ gives the desired result.

share|cite|improve this answer

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

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