...
The main idea is, that we have a universal model of the track segments. From these single track segments, track sections, that connect train stations are modeled and from those track sections, all final schedules for the trains should be build. On each track segment, there should be at most one train: In order to drive over a track, a train must request this track, and afterwards it must free it again (details below). If two trains request the same track section, the priority, which is derived from the train number, decides, which train gets the track section. A train must request all tracks until the next possibility to stop an wait in order to avoid collisions. If the train does not get all track segments, it must free them again in order to avoid deadlocks or delays of other trains. Deadlocks might occur at the exists exits of all train stations, and additionally, if one train with a low priority exists exits the Kicking-Horse-path while another train with a higher priority sends an entry-request. It remains to be seen, if additional deadlocks occur. Deadlocks can be resolved by using a superior Mutex-Controller.
...