Control of multi-agent network finds its application in numerous areas like energy systems, epidemics, healthcare, social networks, and traffic and transportation. The control of misbehaving nodes in the overall network is the research area actively being worked on, where the issue is feedback control cannot be applied to every node due to the obvious reasons of resource limitations. Researchers at the University of South Florida proposed a solution to this problem by using a limited number of nodes, which could guarantee the stability of the overall system. In the proposed system, the multi-agent network works in directed graph topologies, where the number of nodes on which the feedback control is already applied (driver nodes) is less than the total number of uncontrolled nodes or misbehaving nodes. The proportional-integral feedback controller is used in driver nodes to make the rest of the system stable with various intermediate steps that allow the system to gain stability over time. Also, it involves the use of an effective methodology for selecting the driver nodes that could irradicate the effects of misbehaving nodes in a graph-based topology.
Interaction of nodes in a multi-agent network environment, (driver (blue) nodes, misbehaving (red) nodes, floating (white) nodes, and a critical (yellow) node)