H∞ Consensus Design and Optimization for Multi-Agent Systems with Time-Varying Delay

This paper takes multi-agent systems (MASs) under switching communication topologies as the research object, and studies the multi-agent systems consensus problem in view of communication delay in complex environment. Based on the average dwell time (ADT) method and Lyapunov function theory, sufficient conditions for the stability and robustness of the systems are given. The model uncertainty is estimated and compensated online in real time based on deep reinforcement learning (DRL) algorithms, which improves the robustness of the system against strong uncertainties and disturbances while ensuring the stability. Through comparative simulations, it is proved that the optimized control system has better control accuracy and dynamic performance, which can enhance the reliability and intelligence of multi-flight formation.