Source
ICAPS
DATE OF PUBLICATION
05/01/2022
Authors
Alexander Panov Alexey Skrynnik Konstantin Yakovlev Anton Andreychuk
Share

POGEMA: Partially Observable Grid Environment for Multiple Agents

Abstract

We introduce POGEMA1 a sandbox for challenging partiallyobservable  multi-agent  pathfinding  (PO-MAPF)  problems  .This  is  a  grid-based  environment  that  was  specifically  de-signed to be a flexible, tunable and scalable benchmark. It canbe tailored to a variety of PO-MAPF, which can serve as anexcellent testing ground for planning and learning methods,and their combination, which will allow us to move towardsfilling the gap between AI planning and learning.

Join AIRI on social media