On Modelling Multi-Agent Path Finding as a Classical Planning Problem

Abstrakt

Multi-Agent Path Finding (MAPF) deals with finding collision-free paths for a set of agents. It is a special case of a planning problem with only two actions - move and wait - and with one major constraint of no collision between the agents.

The paper addresses the question of how to model MAPF as a classical sequential planning problem. Several models in the PDDL language are proposed and empirically compared.

Klíčová slova

• Multi-Agent Path Finding
• Classical Planning
• Modeling