GitLab

Changed behavior of environment to change action_dict of agents after their input. --> Giacomo please check that this is ok with your implementation

agents keep going forward if motion started, until stopped + RailEnvActions enum + agent.moving flag for the motion status of each agent + penalties for starting/stopping agents, but set to 0

also implies that agents are drawn in their old locations, ie from previous timestep

added a print.out question : why will an agent not move, once he reach it's destination? May he have to move for passing

updated tests after failing

Added load and save to env (left GridTransitionMap load and save for now)