Full-rank square matrix is invertible
Dependencies:
- Rank of a matrix
- RREF is unique
- Inverse of a matrix
- Rank of a homogenous system of linear equations
- Matrix multiplication is associative
- Row equivalence matrix
- Full-rank square matrix in RREF is the identity matrix
Let $A$ be an $n$ by $n$ matrix. Then $\operatorname{rank}(A) = n$ iff $A$ has an inverse.
Proof of 'if' part
Let $A$ have an inverse $B$. Then $AB = BA = I_n$.
$AX = 0$ is a homogeneous system of $n$ linear equations in $n$ variables.
\[ AX = 0 \implies B(AX) = B0 \implies (BA)X = 0 \implies I_nX = 0 \implies X = 0 \] Therefore, $X = 0$ is the only solution to $AX = 0$. Therefore, $\operatorname{rank}(A) = n$.
Proof of 'only-if' part
Let $\operatorname{rank}(A) = n$.
Since $\operatorname{RREF}(A)$ is row-equivalent to $A$, $\operatorname{RREF}(A) = RA$, where $R$ is an invertible matrix.
Since $\operatorname{rank}(A) = n$, all rows of $\operatorname{RREF}(A) = RA$ are non-zero. Since $RA$ is a square matrix in RREF, $RA = I_n$. Since $R$ is invertible, all left inverses of $R$ are the same as all right inverses of $R$. Since $A$ is a right inverse of $R$, it is also a left inverse. Therefore, $AR = I_n = RA$, so $A$ is invertible.
Dependency for:
- General multivariate normal distribution
- A is diagonalizable iff there are n linearly independent eigenvectors
- RREF([A|I]) = [I|inv(A)] iff A is invertible
- AB = I implies BA = I
- Determinant of product is product of determinants
Info:
- Depth: 9
- Number of transitive dependencies: 38
Transitive dependencies:
- /linear-algebra/vector-spaces/condition-for-subspace
- /linear-algebra/matrices/gauss-jordan-algo
- /sets-and-relations/equivalence-relation
- Group
- Ring
- Polynomial
- Integral Domain
- Comparing coefficients of a polynomial with disjoint variables
- Field
- Vector Space
- Linear independence
- Span
- Semiring
- Matrix
- Stacking
- System of linear equations
- Product of stacked matrices
- Matrix multiplication is associative
- Reduced Row Echelon Form (RREF)
- Matrices over a field form a vector space
- Row space
- Elementary row operation
- Every elementary row operation has a unique inverse
- Row equivalence of matrices
- Row equivalent matrices have the same row space
- RREF is unique
- Identity matrix
- Inverse of a matrix
- Inverse of product
- Elementary row operation is matrix pre-multiplication
- Row equivalence matrix
- Equations with row equivalent matrices have the same solution set
- Rank of a matrix
- Basis of a vector space
- Linearly independent set is not bigger than a span
- Homogeneous linear equations with more variables than equations
- Rank of a homogenous system of linear equations
- Full-rank square matrix in RREF is the identity matrix