International Journal of Robust and Nonlinear Control Electrical and Electronic Engineering
Distributed Optimization of Discrete‐Time Strict‐Feedback Nonlinear Multi‐Agent Systems
Abstract
ABSTRACT This paper investigates the distributed optimization problem (DOP) of discrete‐time strict‐feedback nonlinear multi‐agent systems (NMASs) for the first time. To address this issue, a hierarchical distributed optimization control algorithm (DOCA) is proposed to address the DOP. Specifically, in the network layer, a distributed optimal signal generator (DOSG) is first designed to produce the optimal signals which converge to the global optimal solution and the DOP is transformed into a tracking problem for the optimal signals. Then, in the physical layer, the state‐feedback controllers are constructed for agents to ensure their outputs track the optimal signals. The generator relies on the gradient value measured from the real‐time outputs of agents, rather than the known gradient function of the local objective function. A residual propagation mechanism, combined with the variable substitution technique, is employed in the backstepping design process to address the non‐causality issue. It is proven that the output of each agent can converge to the optimal point, thereby solving the DOP. Finally, the rendezvous problem for a group of five single‐link manipulators is considered to validate the presented DOCA.