In this paper, the finite-time asymmetric bipartite consensus problem of multi-agent systems with signed digraph is considered. Firstly, an asymmetric index is introduced to describe a desired output relationship of the agents in terms of quantity. Based on the information of the agent's communication and the index, a novel iterative learning control protocol is proposed. By establishing the input and output errors relationship of the agents along the iteration domain, a sufficient condition is derived and a defined leaderless tracking error is proved to be asymptotically convergence as iteration increases. The result shows that the proposed design can ensure the agents achieve the asymmetric bipartite consensus goal in the finite-time. Moreover, the proposed design is also extended to deal with the problem of the multi-agent systems with heterogeneous dynamics. Finally, numerical simulation examples verify the effectiveness of the proposed protocol.

中文翻译：

---

基于迭代学习控制的多智能体系统有限时间非对称二分共识

本文研究了带符号有向图的多智能体系统的有限时间非对称二部共识问题。首先，引入非对称指数来描述代理在数量方面的期望输出关系。基于代理的通信信息和索引，提出了一种新颖的迭代学习控制协议。通过建立代理沿迭代域的输入和输出误差关系，推导出一个充分条件，并证明定义的无领导跟踪误差随着迭代的增加渐近收敛。结果表明，所提出的设计可以确保代理在有限时间内实现非对称二方共识目标。而且，所提出的设计还扩展到处理具有异构动态的多代理系统的问题。最后，数值模拟实例验证了所提出协议的有效性。