Lipschiz-type的高阶非线性多智能体系统分布式优化算法研究

Distributed optimization for Lipschiz-type high-order nonlinear multi-agent systems

主要研究了具有Lipschiz-type非线性多智能体系统的分布式优化问题.在多智能体网络中,每个个体都拥有一个代价函数,整个多智能体网络的好坏由这些代价函数的和来进行评判.在整个过程中,每个目标仅知道局部的交互信息和其自身代价函数的梯度.为了实现协同优化的目标,提出了一个新的分布式优化算法,运用李雅普诺夫稳定性分析的方法可以证明该算法能够保证所有智能体实现协同优化.最后进行数值仿真,成功地验证了该算法的正确性和可行性.

In this paper,the distributed optimization problem of multi-agent system with Lipschiz-type nonlinear dynamics is studied. For a multi-agent system,each agent has a local performance function. The quality of the whole multi-agent network is judged by the sum of these local performance functions.In the whole process,each agent only knows local interaction information and the gradient of its own objective function.In order to achieve the goal of cooperative optimization,a new distributed optimization algorithm is proposed. Furthermore,by utilizing the Lyapunov stability and optimization analysis,the distributed cooperative optimization problems for Lipschiz-type nonlinear multi-agent systems are solved.Finally,a numerical simulation is shown to verify the effectiveness of the proposed algorithm.