I have a cost function that I try to minimise:
$$ \Pi \sim ||y - Ax||^2 + ||\nabla x || ^2 $$
The gradient is there to constraint that my solution at minimum $x_{min}$ is continuous in 1st derivative.
Now dimension-wise, $y$ is a vector with $m \times 1$, $A$ is a matrix mapping x to y-space with dimension $m \times n$ and $x$ is a vector of $n \times 1$.
Ideally, I am trying to make this optimise problem into a matrix form of:
$$ Kx = b $$
and then hopefully can apply CG on it. But I fail at deriving the matrix $K$ itself.
Any suggestions?