Adaptive distributed optimization for nonlinear stochastic multi-agent systems with event-triggered strategy
摘要
This article focuses on the distributed optimization problem for nonlinear stochastic multi-agent systems. The agents communicate over an undirected graph, and each agent is associated with a strongly convex local cost function whose gradient satisfies the Lipschitz condition. Firstly, two auxiliary variables are introduced to generate reference signals that facilitate the accurate estimation of the optimal state and its higher-order derivatives. Based on these variables, a novel distributed algorithm is developed using feedback control techniques. Secondly, to accelerate convergence to the optimal solution and reduce communication overhead, an event-triggered communication mechanism is integrated into the system. This mechanism governs the information exchange among agents according to a predefined triggering condition, such that each agent updates its state only when the condition is met. A gradient-based event-triggered algorithm is thus proposed to address the cooperative optimization task under stochastic disturbances. Moreover, the convergence properties of the proposed algorithm are analyzed using the framework of the It