Assured Multi-Agent Reinforcement Learning for Safety-Critical Scenarios

Simple item page
dc.contributor.authorRiley, Joshua
dc.date.accessioned2024-05-04T16:04:49Z
dc.date.available2024-05-04T16:04:49Z
dc.date.issued2022-01-11T16:35:25Z
dc.description.abstractMulti-agent reinforcement learning involves and facilitates a team of agents to solve complex decision-making problems in shared environments. This learning process is largely successful in many areas, but its inherently stochastic nature is problematic when applied to safety-critical domains.To solve this limitation, we propose our Multi-Agent Reinforcement Learning (AMARL), which uses a model checking technique called quantitative verification. Quantitative verification provides formal guarantees of agent compliance to safety, performance, and other non-functional requirements, while reinforcement learning occurs and after a policy has been learned.Our AMARL approach is demonstrated using three separate navigation domains, which contain patrolling problems. The multi-agent systems must learn to visit patrol points to satisfy mission objectives while limiting exposure to risky areas in these domains. Different reinforcement learning algorithms have been utilised within these domains: temporal difference learning, game theory, and direct policy search. The performance of these algorithms, while combined with our approach, are presented. Lastly, we demonstrate AMARL with differing system sizes in both homogeneous and heterogeneous multi-agent systems through our extensive experimentation. This experimentation shows that the use of AMARL leads to faster and more efficient performance than standard reinforcement learning and consistently meets safety requirements.
dc.description.sponsorshipThe DSTL
dc.identifier.citationRiley, Joshua (2022). Assured Multi-Agent Reinforcement Learning for Safety-Critical Scenarios. Cranfield Online Research Data (CORD). Poster. https://doi.org/10.17862/cranfield.rd.18166637.v1
dc.identifier.doi10.17862/cranfield.rd.18166637.v1
dc.identifier.urihttps://dspace.lib.cranfield.ac.uk/handle/1826/21383
dc.publisherCranfield University
dc.rightsCC BY 4.0
dc.rights.urihttps://creativecommons.org/licenses/by/4.0/
dc.subjectAssurance'
dc.subject'Safety'
dc.subject'Multi-Agent Reinforcement Learning'
dc.subject'DSDS21'
dc.subject'DSDS21 Poster'
dc.subject'Adaptive Agents and Intelligent Robotics'
dc.titleAssured Multi-Agent Reinforcement Learning for Safety-Critical Scenarios
dc.typePoster

Files

Original bundle
Now showing 1 - 1 of 1
Thumbnail Image
Name:
2 Joshua Riley Yes.pdf
Size:
1.83 MB
Format:
Adobe Portable Document Format
Download

Collections

DSDS 21