Adaptive neural network event-triggered distributed time-varying optimisation consensus control for nonlinear multi-agent systems

This paper investigates the adaptive neural network (NN) event-triggered distributed optimisation consensus control problem for nonlinear multi-agent systems (MASs) under directed stochastically switching topologies. Firstly, a distributed estimator is formulated to estimate the global optimal trajectory, which is generated by minimising the sum of local cost functions. Then, L1 convergence conditions of the distributed estimator are established. Secondly, an NN state observer is established to estimate unmeasurable states of the agents. Based on the distributed estimator and NN state observer, an event-triggered distributed optimisation consensus control scheme is developed via backstepping control technology. The proposed control method has main features as follows: it can ensure that the closed-loop nonlinear MASs are stable and the tracking errors converge to a small neighbourhood of zero in the L1 sense. Finally, we apply the proposed control method to marine surface vehicles, the simulation experiment results and comparison results confirm the effectiveness.