Ability to control the allowed state transitions during a simulation #283
Description
For now all state transitions are considered possible at each step of the simulation, and transitions are only driven by utility differences between states (a state a is a combination of a sequence of activities, destinations and modes). We know some of the transitions take more time than others to take place: it's easier for people to change modes than to change work, for example.
The long term solution to improve our model is to explicitly model different transition costs (see #196). In the meantime, we can start enabling or disabling transitions manually, so that we can recreate simplified state change dynamics: mode changes are allowed from iteration 5, then destination changes from iteration 8, then activity sequences from iteration 11.