摘要
考虑了一类非线性极大极小问题,通过将其转化为等价非线性凸规划提出了求解它的一个神经网络模型,严格证明了新模型是Lyapunov稳定的,并且在有限时间内收敛到原问题的一个精确解。与已有模型相比,新模型结构简单,更适合硬件实现。数值实验表明,该模型不仅可行而且有效。
The paper considers the nonlinear minimax problem. We convert it into equivalent nonlinear convex promgramming problem. Then we propose a new neural network to solve it. It is shown that the proposed neural network is stable in the sense of Lyapunov and can converge to exact optimal solution of the original problem. Compared with the existing neural networks, the new model has simple structure and can be implemented in hardware easily. Simu- lation results demonstrate the effectiveness and characteristics of the proposed neural network.
出处
《陕西科技大学学报（自然科学版）》
2006年第4期90-93,101,共5页
Journal of Shaanxi University of Science & Technology
基金
国家自然科学基金(10571115)
关键词
非线性极大极小问题
非线性凸规划
神经网络
有限时间收敛
nonlinear minimax problem
nonlinear convex programming
neural network
finite-time convergence
