Source
Adapted from my own handwritten and hand-typed notes from Year 1 to Year 3 of my EEE degree
For a square matrix, many ideas from this module turn out to be the same condition. This summary ties linear systems, independence, inverses and determinants together.
Equivalent Statements
For an matrix , the following are all true or all false together:
- is invertible
- has a unique solution for every
- has only the trivial solution
- the columns of are linearly independent
- is row equivalent to the identity
- , equivalently
- the columns of form a basis of
- is not an eigenvalue of (since )
How to Use It
Pick whichever test is cheapest for the problem:
- small matrix: compute
- larger matrix: row reduce and count pivots
- already know the eigenvalues: check none is zero
If any one statement fails, they all fail. For example, a zero row in the RE form means , has no inverse, and has nontrivial solutions.
Example
reduces to a matrix with a zero row (column 3 = column 1 + column 2). So , is singular, , and has the nontrivial solution .
The inverse itself, when it exists, is computed as in Matrix Inverse.