ABSTRACT In the natural world, coordinated behaviors like group migration and collective defense exemplify the capacity of biological populations to process spatial information and adapt their actions correspondingly. Drawing inspiration from these natural phenomena, researchers have introduced diverse multi‐agent cooperative control strategies, such as fixed neighborhood algorithms, fixed number algorithms, and farthest neighborhood algorithms. Nevertheless, these approaches encounter obstacles including sluggish convergence rates, limited resilience to interference, and constrained interpretability. To overcome these limitations, this paper introduces an innovative distributed multi‐agent cooperative control algorithm leveraging an attention mechanism. The algorithm dynamically modulates the information perception weights of agents, facilitating effective coordination in intricate environments. Furthermore, it integrates a pioneering obstacle avoidance strategy, guaranteeing resilient performance in dynamic and densely populated environments. In comparison to existing approaches, the proposed algorithm demonstrates accelerated convergence, enhanced interference resistance, and improved scalability in formation control, all while preserving theoretical interpretability. This study offers a fresh perspective and innovative methodology for advancing cooperative control in multi‐agent systems, with wide‐ranging applications in fields like robotics and autonomous systems.
Liu et al. (Mon,) studied this question.