Aplicación de un algoritmo de resiliencia para la identificación de agentes no cooperativos y la reestructuración de formaciones en sistemas multiagente heterogéneos

Abstract

This work presents a resilience algorithm applied to heterogeneous multi-agent systems that work with a distributed control architecture. The systems used are made up of virtual leaders that form a surface called a convex hull, within which agents with two different dynamics are distributed. The application of a consensus algorithm and a resilience algorithm allow us to have a system capable of using the information shared locally to guarantee the consensus of the multi-agent system, despite the loss or malfunction of one or more agents called non-cooperative agents. This is achieved by identifying the agents that do not fulfill the assigned tasks, isolating any information that they provide during the consensus, forcing the rest of the elements of the system to function without this information, obtaining a new distribution of the agents inside the convex envelope to complete the proposed trajectory.