Let $\mathbf G$ be a given $m \times n$ matrix, $\mathbf y$ a given $m \times 1$ column vector and $\mathbf x$ an unknown $n \times 1$ column vector such that $\mathbf x \ge 0$.
1) How do you find $\displaystyle \min_{\mathbf x} (\mathbf y - \mathbf G\mathbf x)^T(\mathbf y - \mathbf G\mathbf x)$?
2) Whether there is a solution of that problem similar to the simple pseudo-inverse $x=(G^TG)^{−1}G^Ty$ used for solving the least-squares problem?
3) And what if we have $n \times n$ matrix $\mathbf C$ and the constraint $\mathbf C\mathbf x \ge 0$ instead of $\mathbf x \ge 0$?