Building a model that can successfully navigate trains in a railway setting without causing deadlocks is a difficult and complex task. From their position often the trains have multiple roads to the station, with different length, congestion and number of trains. By using heuristics or algorithmic approaches, it is difficult to program a solution that works with different railway scenarios. Additionally, the complexity of the solution should be low, because it is required that the trains are able to reschedule, meaning to adapt to different railway scenarios and blocked passage in real time. For example, if a train malfunctions and blocks a railway, the other trains need to update their routes and schedules.