I'm approaching CT reconstruction using optimization methods instead of the standard radon transform methods. The typical problem I am working with at the moment has approx 109 voxels/variables so it appears to me to be impractical to use slack-variable Lagrangian methods to enforce non-negativity on these variables. Another option I have been considering is, in each iteration of the optimization, to set all negative variables to zero.
Are there more practical ways to enforce non-negativity in large problems such as this?