Robust Prioritised Planning for Multi-agent Pathfinding

For this project we propose a new generation of prioritised planners: there is a priori-
ones which learn from failure and which guarantee to find a prioritised plan tized solution. But
if one exists. To achieve this we will leverage Lazy CBS – a recent break- not if a 1 its other,
through algorithm which combines ideas from Constraint Programming also optimal, path.
with those from Artificial Intelligence – and we will develop an entirely
new approach to MAPF: one that computes feasible prioritised plans fast, optimal prioritised
plans eventually and which can reliably scale to problems with up to thousands of moving
agents.