The Schur complement is defined for a block matrix that is partitioned into four sub-blocks:
where and are square matrices.
Schur Complement of
If the block is invertible, the Schur complement of in is the matrix defined as:
Schur Complement of
Similarly, if the block is invertible, the Schur complement of in is the matrix defined as:
Decomposition
The cool thing is that if the complement is possible, then we can decompose
Decomposition via Schur Complement of
This decomposition is possible if the top-left block is invertible. The Schur Complement of in is:
The matrix can then be decomposed as:
Where:
- The first matrix is block lower triangular ().
- The second matrix is block diagonal () and contains the original block and the Schur complement .
- The third matrix is block upper triangular ().
This formula shows that is congruent to the block-diagonal matrix via a block triangular matrix.
Decomposition via Schur Complement of
This decomposition is possible if the bottom-right block is invertible. The Schur Complement of in is:
The matrix can then be decomposed as:
Where:
- The first matrix is block upper triangular ().
- The second matrix is block diagonal () and contains the Schur complement and the original block .
- The third matrix is block lower triangular ().
