Distributed optimization for nonlinear multi-agent systems with an upper-triangular structure

This paper proposes a new two-step distributed optimization framework with a low-gain parameter for a class of upper-triangular nonlinear multi-agent systems. Specifically, distributed optimal coordinators are first established to generate signals close to the optimal solution by using the real-time values of gradient functions. Then, reference-tracking controllers are designed to make system outputs converge to the generated signals by coordinators. Correspondingly, system outputs can converge to the optimal solution by choosing the appropriate low-gain parameter. Finally, simulation results illustrate the proposed control scheme.