摘要
考虑了一类线性变分不等式问题,提出了求解它的一个射影神经网络模型,构造恰当的能量函数,给出了确保新模型稳定和有限时间收敛的3个充分条件,并在适当的条件下,证明了该模型的指数稳定性。新模型结构简单,易于硬件实现,可用来求解非单调的问题。数值试验表明该网络不仅有效而且可行。
This paper considers a class of linear variational inequalities, and presents a projected neural network to solve it. Three sufficient conditions are provided to ensure that the proposed neural network is stable in the sense of lyapunov and has a finite - time convergence by introducing the energy functions.Meanwhile, the global exponential stability of the new model is also shown under some mild conditions. The new model has a lower complexity for parallel implementation,and can be applied to solve some nonmonotone problems.The feasibility and efficiency are confirmed by numerical experiments.
出处
《济南大学学报（自然科学版）》
CAS
2007年第1期82-85,共4页
Journal of University of Jinan(Science and Technology)
基金
陕西省自然科学基础研究计划基金(2006A02)
国家自然科学基金(60671063)
关键词
线性变分不等式
神经网络
指数稳定性
有限时间
收敛性
linear variational inequality
neural network
exponential stability
finite- time convergence
作者简介
王婧（1982-），女，宁夏银川人，硕士生。
