Distributed multi-agent deep reinforcement learning for cooperative multi-robot pursuit

As a popular research topic in the area of distributed artificial intelligence, the multi-robot pursuit problem is widely used as a testbed for evaluating coordinated and cooperative strategies in multi-robot systems. This study the problem of multi-robot pursuit game using reinforcement learning (R...

Full description

Saved in:
Bibliographic Details
Published in:Journal of engineering (Stevenage, England) Vol. 2020; no. 13; pp. 499 - 504
Main Authors: Yu, Chao, Dong, Yinzhao, Li, Yangning, Chen, Yatong
Format: Journal Article
Language:English
Published: The Institution of Engineering and Technology 01.07.2020
Subjects:
control engineering computing
decentralised‐execution scheme
deep RL methods
distributed artificial intelligence
distributed control
distributed multiagent deep reinforcement learning
environmental agents
game theory
individual action output
individual leaning update process
learning (artificial intelligence)
learning systems
multiagent deep RL approach
multirobot pursuit game
multirobot systems
multi‐agent systems
multi‐robot systems
pursuit‐evasion problem
The 3rd Asian Conference on Artificial Intelligence Technology (ACAIT 2019)
pursuit-evasion problem
multi-agent systems
individual leaning update process
distributed control
game theory
multirobot pursuit game
deep RL methods
multi-robot systems
decentralised-execution scheme
distributed artificial intelligence
learning systems
environmental agents
individual action output
distributed multiagent deep reinforcement learning
multiagent deep RL approach
learning (artificial intelligence)
control engineering computing
multirobot systems
ISSN:2051-3305, 2051-3305
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:As a popular research topic in the area of distributed artificial intelligence, the multi-robot pursuit problem is widely used as a testbed for evaluating coordinated and cooperative strategies in multi-robot systems. This study the problem of multi-robot pursuit game using reinforcement learning (RL) techniques is studied. Unlike most existing studies that apply fully centralised deep RL methods based on the centralised-learning and decentralised-execution scheme, the authors propose a fully decentralised multi-agent deep RL approach by modelling each agent as an individual deep RL agent that has its own individual learning system (i.e. individual action-value function, individual leaning update process, and individual action output). To realise coordination among agents, the limited information of other environmental agents is used as input of the learning process. Experimental results show that both distributed and centralised approaches can ultimately solve the pursuit-evasion problem in different dimensions, but the learning efficiency and coordination performance of the proposed distributed approach are much better than the traditional centralised approach.
ISSN:2051-3305
2051-3305
DOI:10.1049/joe.2019.1200