作者: Mohamed-Ali Belabbas
DOI:
关键词: Mathematical optimization 、 Stability (learning theory) 、 Topology (chemistry) 、 Information flow (information theory) 、 Mathematics 、 Autonomous agent 、 Control theory 、 Task (project management) 、 Decentralised system 、 Control (management) 、 Need to know
摘要: Given an ensemble of autonomous agents and a task to achieve cooperatively, how much do the need know about state in order it? We introduce new methods understand these aspects decentralized control. Precisely, we framework capture what with partial information can by cooperating illustrate its use deriving results global stabilization directed formations. This underscores differentiate knowledge agent has accomplish from current system. The control formations proven be more difficult than initially thought, as is exemplified lack result for n \geq 4 agents. established part I that space planar non-trivial topology. propose here extension notion stability which, because it acknowledges this topology, applied study formation then develop reduces question whether feedback stabilize system two sets functions intersect. apply = show not have enough implement locally stabilizing laws. Additionally, laws respect flow cannot target configuration without other, unwanted configurations.