|
线性约束问题:可以加$\lambda$化为无约束的问题:
第13页 | We shall prove that our constrained minimization problem has a unique solution given by the system of linear equations
\[
\begin{aligned}
C^{-1} y+A \lambda & =b, \\
A^{\top} y & =f,
\end{aligned}
\]
which can be written in matrix form as
\[
\left(\begin{array}{cc}
C^{-1} & A \\
A^{\top} & 0
\end{array}\right)\left(\begin{array}{l}
y \\
\lambda
\end{array}\right)=\left(\begin{array}{l}
b \\
f
\end{array}\right) .
\]Note that the matrix of this system is symmetric. Eliminating $y$ from the first equation
\[
C^{-1} y+A \lambda=b,
\]
we get
\[
y=C(b-A \lambda),
\]
and substituting into the second equation, we get
\[
A^{\top} C(b-A \lambda)=f,
\]
that is,
\[
A^{\top} C A \lambda=A^{\top} C b-f .
\]However, by a previous remark, since $C$ is symmetric positive definite and the columns of $A$ are linearly independent, $A^{\top} C A$ is symmetric positive definite, and thus invertible. |
|
|