Stability Conditions for Discrete Neural Networks in Partial Simultaneous Updating Mode
Stability Conditions for Discrete Neural Networks in Partial Simultaneous Updating Mode
复制标题
DOI：
10.1007/11427391_39
复制
发表时间：
2005-05
期刊：
影响因子：
--
通讯作者：
Runnian Ma;Shengrui Zhang;Sheping Lei
中科院分区：
文献类型：
--
作者：
Runnian Ma;Shengrui Zhang;Sheping Lei
离散Hopfield神经网络的稳定性分析不仅具有重要的理论意义，而且可以广泛应用于联想记忆、组合优化等领域。本文主要研究了非对称离散Hopfield神经网络在部分同时更新模式下的动态行为，并利用Lyapunov方法和一些分析技巧给出了一些新的简单的稳定性条件。给出了部分同时更新模式下网络收敛于稳定状态的几个新的充分条件。所得结果改进和推广了前人的相应结果。此外，我们还提供了一种分析和设计稳定的离散Hopfield神经网络的方法。
The stability analysis of discrete Hopfield neural networks not only has an important theoretical significance, but also can be widely used in the associative memory, combinatorial optimization, etc. The dynamic behavior of asymmetric discrete Hopfield neural network is mainly studied in partial simultaneous updating mode, and some new simple stability conditions of the networks are presented by using the Lyapunov method and some analysis techniques. Several new sufficient conditions for the networks in partial simultaneous updating mode converging towards a stable state are obtained. The results established here improve and extend the corresponding results given in the earlier references. Furthermore, we provide one method to analyze and design the stable discrete Hopfield neural networks.