This is a cool article! And it lead me down the path to another cool factorio compsci paper called 'The steady-states of splitter networks' where they really go into the theoretical properties of the belts and splitters.
But I can't help to feel that the following paragraph should be removed, since it really seems to be a bit confused about the concepts of both Np-complete and NP-hard.
> Finding a feasible solution is NP-hard because we can't do any better than start placing the components on a 2D grid and check if the properties are respected. Furthermore, finding the optimal solution (i.e., minimum number of components) is NP-complete because it will require enumerating all the possible solutions in order to find the best one.
But I can't help to feel that the following paragraph should be removed, since it really seems to be a bit confused about the concepts of both Np-complete and NP-hard.
> Finding a feasible solution is NP-hard because we can't do any better than start placing the components on a 2D grid and check if the properties are respected. Furthermore, finding the optimal solution (i.e., minimum number of components) is NP-complete because it will require enumerating all the possible solutions in order to find the best one.