This paper mainly studies quasi-bipartite synchronization (QBPS) of signed delayed neural networks (SDNNs) under impulsive effects, in which the nodes have cooperative as well as antagonistic interactions. It is assumed that disturbance occurs in the communication channels between some neighboring agents at impulsive occurring instants. Under the balanced network topology, some sufficient criteria to achieve QBPS of SDNNs are proposed by utilizing algebraic graph theory and extended Halanay differential inequality. Moreover, for the QBPS error of SDNNs, the upper bound of the final error state is also provided explicitly. Two numerical examples are presented to demonstrate the correctness of the theoretical results.

本文主要研究脉冲效应下带符号的延迟神经网络（SDNN）的准二分同步（QBPS），其中节点之间具有协同作用和拮抗作用。假定在冲动发生的瞬间在一些相邻代理之间的通信信道中发生扰动。在平衡网络拓扑下，利用代数图论和扩展的Halanay微分不等式，提出了一些实现SDNNs QBPS的充分标准。此外，对于SDNN的QBPS错误，还明确提供了最终错误状态的上限。给出两个数值例子，以证明理论结果的正确性。