具有通信时延和有向网络的多智能体分布式优化(英文)

Distributed optimization of multi-agent with communication delay and directed network

中文关键词  通信时延  多智能体系统  分布式优化  非平衡有向网络

英文关键词  communication delay  multi-agent systems  distributed optimization  unbalanced directed network

中文摘要

      本文对非平衡有向拓扑下一阶多智能体系统的分布式优化问题进行研究. 研究的智能体在与邻居的通信过程 中都有一个变化的时延. 本文的目标是找到使得目标函数f(x) =NΣi=1fi(xi)最小的智能体的状态. 提出了一种基于有向 图的拉普拉斯矩阵的零特征值对应的左特征向量和智能体的局部信息的控制器. 在这项研究中, 去掉了fi(xi)的梯度有 界要求, 并且不要求网络平衡. 在某些假设下, 所有智能体都达到相同的状态, 同时最小化目标函数NΣi=1fi(xi). 最后, 通 过数值模拟验证了本文的研究结果.

英文摘要

      This paper studies the distributed optimization problem for the first-order multi-agent systems in unbalanced directed topology. Each agent has a time delay in the communication process with its neighbors. Our goal is to find a state of the agent that minimizes the objective function f(x) =NΣi=1fi(xi). A controller based on the left eigenvector corresponding to the zero eigenvalue of L of the graph G and the local information of each agent is proposed. In this study, the bounded requirement of the gradient of fi(xi) is removed, and it does not require the network to be balanced. Under some assumptions, all agents reach the same state while minimizing the objective function NΣi=1fi(xi). Finally, a numerical simulation illustrates the results of this article.