摘要
通过鞍点定理和投影理论,提出了一个解二次极大极小问题的变时滞神经网络。利用泛函微分方程理论,给出了确保该变时滞神经网络全局指数稳定的充分条件。由于稳定性分析中不需要原极大极小问题的凸性,该网络可以用来求解一类非凸优化问题。仿真实例验证了理论的正确性和网络的性能。
In this paper, a neural network model with time -varying delays is proposed to solve a class of maximin problems by employing the saddle point theorem and projection theory. The sufficient conditions are derived to ensure the global exponential stability of the delayed neural network by the theory of functional differential equations. Since the stability of the neural network don＇ t require the convexity of the original maximin problem, the obtained results can be applied to solve a class of nonconvex optimization problem. Finally, numerical examples are presented to demonstrate the validity of the obtained results and the performance of this network.
出处
《云南师范大学学报（自然科学版）》
2009年第5期29-33,共5页
Journal of Yunnan Normal University:Natural Sciences Edition
基金
国家自然科学基金资助项目(60671063)
陕西省自然科学基础研究计划(2006A02)
关键词
时变时滞
二次极大极小
指数稳定性
Time -varying delays
Quadratic maximin
Global exponential stability
