MSSE Thesis Defense: Samuel Langlois, "Decentralized Multiagent Metareasoning"
Friday, April 9, 2021
MSSE Thesis Defense
Decentralized Multiagent Metareasoning Applications in Task Allocation and Path Finding
Dr. Jeffrey W. Herrmann, Advisor
Dr. Michael Otte, Committee member
Dr. Huan Xu, Committee member Thesis
Decentralized task allocation and path finding are two problems where no single fixed algorithm provides the best solution in all environments. Past research has considered metareasoning approaches to these problems that take in map, multiagent system, or communication information. None of these papers address the application of metareasoning about individual agent state features inside a multiagent system.
This thesis presents the application of a meta-level policy that is conducted offline using supervised learning through extreme gradient boosting. The multiagent system used here operates under full communication and the system uses an independent multiagent metareasoning structure.
This research compares the effects of metareasoning on the CBAA and DHBA multiagent task allocation algorithms to controlled, random, and necessity-based applications. We also compare the results of the metareasoning to three, fixed path-finding algorithms: LRA*, WHCA*, and CBS.
The results of this comparative research suggest that this metareasoning approach can reduce the communication and computational overhead without sacrificing performance.