【作者】 佘秋野； 陈丽换； 张玉民； 郭雷；

【Author】 She Qiuye Chen Lihuan Zhang Yumin Guo Lei (Institute of Automation, Southeast University, Nanjing 210096, China) (College of Mechanical and Electrical Engineering, Jinling Institute of Technology, Nanjing 210001, China)

【机构】 东南大学自动化研究所； 金陵科技学院机电工程学院；

【摘要】 为了更加深入地了解神经网络的复杂性,研究了中立型Hopfield神经网络模型,对该模型的稳定性的判定进行了探讨．通过构造合适的Lyapunov泛函,借助线性矩阵不等式(LMI)的一些技巧以及Lyapunov-Krasovskii稳定性理论,给出了该神经网络模型稳定的充分条件,以及模型稳定时中立项所必须满足的必要条件．给出的结果改进了已有的结论和方法,在一定程度上降低了求解的难度,在实际应用方面更加简捷方便．仿真实例表明了算法的有效性,以及所存在的保守性.更多还原

【Abstract】 To understand the complexity of neural networks more thoroughly, the neutral Hopfield neural networks model is researched. A discussion for the judgement of its stability is proceeded. By constructing proper Lyapunov functional, a sufficient condition and a necessary condition which the neutral item must satisfy when the model is steady are given based on LMI technique and Lyapunov-Krasovskii stability theory. The given result has improved the conclusion and the method existed. It has lowered the difficulty of solution in certain degree and is more convenient in application as well. The simulation shows the efficiency and the conservative of the presented algorithm.更多还原