Model-Free H∞ Optimal Hierarchical Control of Heterogeneous Multiagent Systems via Adaptive Dynamic Programming

This article investigates the $H_{\infty } $ optimal output-feedback control problem of heterogeneous multiagent systems. First, a hierarchical control scheme is designed to reduce the algorithm’s complexity and the expected global performance constraints can be ensured by designing compensation input indicator. Second, a relaxation parameter is introduced to derive the optimal solution under output feedback. Additionally, a policy iteration algorithm and vectorization method are presented to determine local and collaborative control gains. This relaxation parameter serves to ease the design conditions for performance indicator. In addition, adaptive dynamic programming (ADP) is introduced and reversible datasets are designed to obtain optimal parameters with unknown drift dynamics. This design achieves model-free control of optimal output feedback for heterogeneous multiagent systems. Finally, the effectiveness of the control schemes is validated using F-16 aircraft and 4-wheel autonomous vehicles as examples.