具有变时滞非自治递归神经网络有界性和全局指数稳定性  

Boundedness and global exponential stability for non-autonomous recurrent neural networks with time-varying delays

作　　者：李圣荣[1] 

机构地区：[1]咸阳职业技术学院,陕西咸阳712000

出　　处：《宝鸡文理学院学报（自然科学版）》2013年第2期15-20,共6页Journal of Baoji University of Arts and Sciences(Natural Science Edition)

摘　　要：目的研究变时滞的非自治递归神经网络的有界性和全局指数稳定性。方法利用广义的Halanay不等式和Lyapunov泛函方法。结果给出了保证变时滞非自治递归神经网络的全局指数稳定性的充分条件。结论文中无需考查模型平衡点的数目,也不要求激活函数可导或是单调,同时也摆脱时滞项可导的要求,所得结果具有一般特性和新颖性,改善了相关文献的理论结果。2个实例及其相应的数值仿真进一步验证了所得结果的有效性。Objective-To investigate the boundedness and global exponential stability for non-au- tonomous recurrent neural networks with time-varying delays. Methods-The generalized Halanay in-equality and Lyapunov functional method are applied. Results-New sufficient conditions are obtained ensuring the global exponential stability of the solution of the non-autonomous recurrent neural net- works with time-varying delays. Conclusion--The considered model have not been assumed any equi- libria and the activation functions are not supposed to be differentiable or nondecreasing. What is more, for the delay terms, they are gotten rid of the condition of differentiable. So, the results ob-tained in this paper are general, new and improve the previous works. Two illustrative examples and their numerical simulations are also given to demonstrate the effectiveness of the proposed results.

关 键 词：非自治递归神经网络 DINI导数 Lyapunov泛函 Halanay不等式 全局指数稳定 

分 类 号：O175[理学—数学]