Linear Algebra
4.89 min read

Right Invertibility and Surjectivity

A matrix ARm×nA \in \mathbb{R}^{m \times n} is right invertible if there exists a matrix RRn×mR \in \mathbb{R}^{n \times m} such that AR=ImAR = I_m. That is, applying AA after RR gives the identity.

The connection to surjectivity: AA is right invertible if and only if AA is surjective. Proof: if AR=IAR = I, then for any bRm\mathbf{b} \in \mathbb{R}^m, A(Rb)=(AR)b=Ib=bA(R\mathbf{b}) = (AR)\mathbf{b} = I\mathbf{b} = \mathbf{b}, so x=Rb\mathbf{x} = R\mathbf{b} is a solution. Conversely, right invertibility requires rank =m= m.

The right inverse is a "right undo" — composing AA on the right of RR recovers the identity. Geometrically: RR picks a "section" of AA — one pre-image for each output.

Formal View

Theorem 4.8
ARm×nA \in \mathbb{R}^{m \times n} is right invertible \Longleftrightarrow AA is surjective \Longleftrightarrow Rank(A)=m\text{Rank}(A) = m.

Why This Matters

Right invertibility is the algebraic certificate of surjectivity.

  • Pseudoinverse: the minimum-norm right inverse of a surjective matrix
  • Control systems: right invertibility of the control matrix means all states can be reached

Quiz

Question 1

A 3×53 \times 5 matrix with rank 3 is right invertible.

Common Mistakes

  • Confusing right inverse (undoes AA when composed on the right) with left inverse (undoes AA when composed on the left).