Hello @darnstrom !
I want to create C code generation tools based on LinearMPC.jl. I was not able to undestand how known disturbances $d$ are handled in compute_control function ? Is it assumed to be current value $d_k$ ? If yes, you probably assumed that it is constant over the prediction horizon $\hat{d}_{k+1} = \hat{d}_{k+2} = ... = \hat{d}_{k+N_p} = d_k$ ? I would mention it in the documentation. Custom disturbance predictions could be an interesting feature to add, it is really similar to the reference preview feature.
