多重时滞离散神经网络的指数稳定性

Exponential stability of discrete-time neural networks with multiple delays

论文研究了带有多重时滞的连续神经网络的离散模型.离散神经网络模型对于连续模型的实际应用及其数值计算具有十分重要的意义,不仅如此,它自身也有独特的动力学性质与广泛的应用领域.论文通过不动点理论等分析方法,在没有假设激励函数有界性和可微性的情况下,给出了系统的不动点存在唯一性和全局指数稳定性的充分条件.

A discrete analogue of continuous-time neural networks with multiple delays was investigated. The discrete neural models was of great importance in application and numerical computation of continuous ones. Furthermore, it had its own merits and application areas. In this paper, by analytic techniques such as fixed point theorem, sufficient conditions guaranteeing the existence, uniqueness and global exponential stability of an equilibrium point were obtained, without assuming the boundedness and differentiability of activation functions.