Source
Adapted from my own handwritten and hand-typed notes from Year 1 to Year 3 of my EEE degree
For an matrix , the inverse (if it exists) satisfies
If exists, is invertible (non-singular). Otherwise it is singular. The inverse, when it exists, is unique.
Properties
For invertible matrices and :
- (order reverses)
- in general
- is invertible if and only if
Proof of the first one: .
2 by 2 Formula
Augmented Matrix Method
For any size, row reduce . If the left block becomes , the right block is the inverse:
It works because the row operations that turn into are the same as multiplying by , and applying them to gives . If a zero row appears on the left, is singular.
Example with :
So , which matches the 2 by 2 formula since .
Adjoint Formula
Using cofactors from Determinants and Cofactor Expansion:
The adjoint is the transpose of the cofactor matrix. For a 2 by 2 matrix this gives back the formula above.
Solving Ax = b
If is invertible, then for every the system has the unique solution
With the above and : .
In practice Gaussian elimination is cheaper than computing , but the formula is useful in theory.