Negative updating combined with opinion pooling in the best-of-n problem in swarm robotics

Lee, Chanelle; Lawry, Jonathan; Winfield, Alan F.

Abstract

© 2018, Springer Nature Switzerland AG. There is a need for effective collective decision making in decentralised multi-agent and robotic systems. This paper introduces a novel approach to the best-of-n decision problem with large n. It utilises negative feedback obtained from direct pairwise comparison of options and evidence preserving opinion pooling. We present agent-based simulation experiments that explore the effects of pool size and the number of options on the speed of consensus. Robotic simulation experiments are then used to investigate the potential of the approach as a method for solving the best-of-n decision problem in swarm robotic applications. Overall, the results suggest that the proposed approach is highly scalable with regards to n.

• Presentation Conference Type: Conference Paper (published)
• Conference Name: Eleventh International Conference on Swarm Intelligence (ANTS 2018)
• Start Date: Oct 29, 2018
• End Date: Oct 31, 2018
• Acceptance Date: Jun 14, 2018
• Online Publication Date: Oct 3, 2018
• Publication Date: Jan 1, 2018
• Deposit Date: Nov 26, 2018
• Publicly Available Date: Dec 18, 2019
• Journal: Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
• Print ISSN: 0302-9743
• Publisher: Springer Verlag
• Peer Reviewed: Peer Reviewed
• Volume: 11172 LNCS
• Pages: 97-108
• Series Title: Swarm Intelligence. ANTS 2018. Lecture Notes in Computer Science
• Series Number: 11172
• DOI: https://doi.org/10.1007/978-3-030-00533-7_8
• Keywords: best-of-n, opinion pooling, swarm robotics
• Public URL: https://uwe-repository.worktribe.com/output/858278
• Publisher URL: https://link.springer.com/chapter/10.1007/978-3-030-00533-7_8
• Additional Information: Title of Conference or Conference Proceedings : Proceedings of the Eleventh International Conference on Swarm Intelligence (ANTS 2018)
• Contract Date: Nov 26, 2018