In this talk, the speaker discusses automated short-term train planning and how it is handled in OSRD, an open-source railway design tool. The problem is finding a path for a train that does not disrupt the established timetable. The speaker explains the rules that need to be followed, such as not causing delays for other trains and not encountering yellow signals. The solution involves representing the problem as a graph with dimensions of position, time, and speed, and using A* algorithm with a cost function and optimization heuristic to find the optimal solution. The speaker also discusses challenges and problems faced, such as speed and adding delay, and mentions future improvements, including integrating the signaling system and allowing the user to set departure and arrival times. The project is currently a work in progress, but the speaker demonstrates a video showing the tool's functionality. Finally, there are a few questions from the audience, including one about using artificial intelligence for this problem, and the speaker mentions that it could be a good heuristic, but not a critical component. Overall, the talk is about utilizing open source software to address the challenges of short-term train planning.