Unknown Resource Reallocation in a Class of Multiagent Systems: A Distributed Approach With Formation Control Perspectives
提出一种分布式控制框架，使多智能体系统在总资源量未知的情况下，通过智能体间交互实现公平高效的资源重新分配，并保持资源总量不变。
This article deals with the challenge of a fair and efficient reallocation of resources in a multiagent system (MAS). A novel dynamic control framework is introduced in which the state of each agent evolves according to a distributed control law. This law ensures the allocation of resources according to predefined weights and promotes a fair distribution among the agents. The control strategy is thoroughly analyzed to show that resource allocation remains well-defined, avoids singularities, and ensures consistent performance. Moreover, it is proved that the total resource allocated remains constant, showing that the system can preserve a crucial invariant throughout its evolution. Theoretical results validate the approach and show its adaptability to changing agent states while maintaining the overall resource constraints. A key advantage of the proposed strategy is that the total amount of resources matching the initial distribution remains unknown to the agents. Instead, through emergent behavior resulting from interaction with other agents, each agent uses an appropriate amount of the resource based on its own requirements. This work contributes to an effective control strategy to achieve stable and fair resource allocation in multiagent networks.