--- title: Shur complement tags: Linear algebra GA: G-77TT93X4N1 --- # Shur complement ## Derivation We consider a square matrix with block structure $$ \tag{1} \begin{bmatrix} A & B \\ C & D \end{bmatrix} \begin{bmatrix} X\\ Y \end{bmatrix}= \begin{bmatrix} U\\ V \end{bmatrix}. $$ If $A$ is invertible, we can apply one-step block Gaussian elimination to eliminate $C$. We mulplity both side by an elimination matrix as follows: $$ \begin{bmatrix} I & 0 \\ -CA^{-1} & I \end{bmatrix} \begin{bmatrix} A & B \\ C & D \end{bmatrix} \begin{bmatrix} X\\ Y \end{bmatrix}= \begin{bmatrix} I & 0 \\ -CA^{-1} & I \end{bmatrix} \begin{bmatrix} U\\ V \end{bmatrix}. $$ Direct calculation leads to $$ \tag{2} \begin{bmatrix} A & B \\ 0 & D-CA^{-1}B \end{bmatrix} \begin{bmatrix} X\\ Y \end{bmatrix}= \begin{bmatrix} U\\ V-CA^{-1}U \end{bmatrix}. $$ Therefore, to solve (1) is equivalent to solve (2), and we can split the procedure into two steps: 1. Solve the following system to find $Y$: $$ \tag{3} (D-CA^{-1}B) Y = V-CA^{-1}U. $$ 2. Solve the following system to find $X$: $$ \tag{4} AX = U-BY. $$ --- ## Example Consider solving the linear system $$ \tag{5} \left[\begin{array}{@{}c|c@{}} I_{10} & B \\ \hline C & D \end{array}\right] \begin{bmatrix} x_1\\ \vdots\\ x_{12} \end{bmatrix}= \begin{bmatrix} b_1\\ \vdots\\ b_{12} \end{bmatrix}, $$ where $I_{10}$ is the $10\times 10$ identity matrix and $$ B\in\mathbb{M}_{10\times 2}, \quad C\in\mathbb{M}_{2\times 10}, \quad D\in\mathbb{M}_{2\times 2}. $$ Since $(I_{10})^{-1}x = x$ for any $x$. We can take advantages of the Schur complement to solve this system. We solve the system in 2 steps: 1. Solve a $2\times 2$ system: $$ \tag{6} (D-CB) \begin{bmatrix} x_{11}\\ x_{12} \end{bmatrix} = \begin{bmatrix} b_{11}\\ b_{12} \end{bmatrix}-C\begin{bmatrix} b_{1}\\ \vdots\\ b_{10} \end{bmatrix}. $$ 2. Calculate the remaining part of the solution directly: $$ \tag{7} \begin{bmatrix} x_{1}\\ \vdots\\ x_{10} \end{bmatrix} = \begin{bmatrix} b_{1}\\ \vdots\\ b_{10} \end{bmatrix}-B\begin{bmatrix} x_{11}\\ x_{12} \end{bmatrix}. $$ --- ## Remark * Schur complement is useful when the linear system $Ax=b$ can be solved easily. * (NOTE) Never compute $A^{-1}$. To evaluate $A^{-1}b$, we solve $Ax=b$.