This map is wider than the screen. The view follows the drones while they fly; scroll sideways to look around.
Speed
Fly-in is a routing puzzle from the 42 Berlin curriculum. A fleet of drones has to get from a start zone to a goal, through a network of zones and the corridors between them. Every zone and corridor holds only so many drones at once, and the solver has to bring them all in using as few turns as possible.
This page is a visualisation of that solver. The routes come from the Python code of Jason Koch (jasonkochai); the page only draws what the program decided, one turn at a time.
How I learned it: the interactive notes behind this project →
This map is wider than the screen. The view follows the drones while they fly; scroll sideways to look around.
The solver plans the drones one after another. For each drone it searches over places and turns together, so waiting a turn counts as a move, and it always looks at the earliest turn first. The first time that search reaches the goal is the earliest arrival that does not clash with any drone planned before it.
That route is then booked turn by turn, zones and corridors alike, and the next drone plans around the bookings. This is why you see drones wait at a station: the place they need next is still booked for that turn.
On every map that ships with the 42 project, the plan uses exactly the number of turns the project gives as the optimum.