Decentralized Multi-agent Path Finding based on Deep Reinforcement Learning

DOI

10.2991/978-94-6463-300-9_19How to use a DOI?

Keywords

multi-agent path finding; deep reinforcement learning; robotics

Abstract

Multi-agent path finding (MAPF) problem has long been a focus of reinforcement learning researchers due to is potential applications to real world robot deployment. Nowadays, many efforts have been made to develop decentralized MAPF algorithms since decentralized ways tend to scale better in larger robot team compared with centralized algorithms. However, reviews on this topic are still lacking. This paper reviews some state-of-the-art decentralized MAPF algorithms. These algorithms are classified into three categories, i.e., imitation learning (IL) algorithms, graph neuro networks (GNN) and task decomposition algorithms. IL-based algorithm learns from expert data, GNN-based algorithms learn by incorporating GNN, and task decomposition methods decompose MAPF into easier subtasks. For each algorithm, first its formulation of MAPF problem, i.e., the structure of observations, actions and rewards, is introduced. Then its essential part is analyzed in detail. Finally, its advantages and limitations are investigated and comparisons with other algorithms are made. In the end, the paper is summarized and provides outlook to the field.

Copyright

© 2023 The Author(s)

Open Access

Open Access This chapter is licensed under the terms of the Creative Commons Attribution-NonCommercial 4.0 International License (http://creativecommons.org/licenses/by-nc/4.0/), which permits any noncommercial use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license and indicate if changes were made.

Download article (PDF)

TY  - CONF
AU  - Heng Lu
PY  - 2023
DA  - 2023/11/27
TI  - Decentralized Multi-agent Path Finding based on Deep Reinforcement Learning
BT  - Proceedings of the 2023 International Conference on Image, Algorithms and Artificial Intelligence (ICIAAI 2023)
PB  - Atlantis Press
SP  - 185
EP  - 192
SN  - 2352-538X
UR  - https://doi.org/10.2991/978-94-6463-300-9_19
DO  - 10.2991/978-94-6463-300-9_19
ID  - Lu2023
ER  -